def calculate_list_of_independent_viper_run_indices_used_for_outlier_elimination(no_of_viper_runs_analyzed_together,
no_of_viper_runs_analyzed_together_from_user_options, masterdir, rviper_iter, criterion_name):
from utilities import combinations_of_n_taken_by_k
# generate all possible combinations of (no_of_viper_runs_analyzed_together - 1) taken (3 - 1) at a time
import itertools
number_of_additional_combinations_for_this_viper_iteration = combinations_of_n_taken_by_k(no_of_viper_runs_analyzed_together - 1,
no_of_viper_runs_analyzed_together_from_user_options - 1)
criterion_measure = [0.0] * number_of_additional_combinations_for_this_viper_iteration
all_n_minus_1_combinations_taken_k_minus_1_at_a_time = list(itertools.combinations(range(no_of_viper_runs_analyzed_together - 1),
no_of_viper_runs_analyzed_together_from_user_options - 1))
no_of_processors = mpi_comm_size(MPI_COMM_WORLD)
my_rank = mpi_comm_rank(MPI_COMM_WORLD)
for idx, tuple_of_projection_indices in enumerate(all_n_minus_1_combinations_taken_k_minus_1_at_a_time):
if (my_rank == idx % no_of_processors):
list_of_viper_run_indices = list(tuple_of_projection_indices) + [no_of_viper_runs_analyzed_together - 1]
criterion_measure[idx] = measure_for_outlier_criterion(criterion_name, masterdir, rviper_iter, list_of_viper_run_indices)
plot_errors_between_any_number_of_projections(masterdir, rviper_iter, list_of_viper_run_indices, criterion_measure[idx])
criterion_measure = mpi_reduce(criterion_measure, number_of_additional_combinations_for_this_viper_iteration, MPI_FLOAT, MPI_SUM, 0, MPI_COMM_WORLD)
if (my_rank == 0):
index_of_sorted_criterion_measure_list = [i[0] for i in sorted(enumerate(criterion_measure), reverse=False, key=lambda x: x[1])]
list_of_viper_run_indices_for_the_current_rrr_viper_iteration = list(all_n_minus_1_combinations_taken_k_minus_1_at_a_time[index_of_sorted_criterion_measure_list[0]]) + \
[no_of_viper_runs_analyzed_together - 1]
mainoutputdir = masterdir + DIR_DELIM + NAME_OF_MAIN_DIR + ("%03d" + DIR_DELIM) % (rviper_iter)
if criterion_measure[index_of_sorted_criterion_measure_list[0]] == TRIPLET_WITH_ANGLE_ERROR_LESS_THAN_THRESHOLD_HAS_BEEN_FOUND:
list_of_viper_run_indices_for_the_current_rrr_viper_iteration.insert(0,MUST_END_PROGRAM_THIS_ITERATION)
else:
list_of_viper_run_indices_for_the_current_rrr_viper_iteration.insert(0,DUMMY_INDEX_USED_AS_BUFFER)
if criterion_name == "80th percentile":
pass_criterion = criterion_measure[index_of_sorted_criterion_measure_list[0]] < PERCENT_THRESHOLD_Y
elif criterion_name == "fastest increase in the last quartile":
pass_criterion = criterion_measure[index_of_sorted_criterion_measure_list[-1]] > PERCENT_THRESHOLD_Y
else:
pass_criterion = False
if not pass_criterion:
list_of_viper_run_indices_for_the_current_rrr_viper_iteration = [EMPTY_VIPER_RUN_INDICES_LIST]
import json; f = open(mainoutputdir + "list_of_viper_runs_included_in_outlier_elimination.json", 'w')
json.dump(list_of_viper_run_indices_for_the_current_rrr_viper_iteration[1:],f); f.close()
mpi_barrier(MPI_COMM_WORLD)
return list_of_viper_run_indices_for_the_current_rrr_viper_iteration
mpi_barrier(MPI_COMM_WORLD)
return [EMPTY_VIPER_RUN_INDICES_LIST]
def comp_rep(refrings, data, itout, modout, vol, group, nima, nx, myid,
main_node, outdir):
import os
from fundamentals import rot_shift2D
from utilities import get_params_proj, params_3D_2D
from mpi import mpi_reduce, MPI_COMM_WORLD, MPI_FLOAT, MPI_SUM
avg = [EMData() for i in xrange(len(refrings))]
avg_csum = [0.0 for i in xrange(len(refrings))]
for i in xrange(len(refrings)):
avg[i] = EMData()
avg[i].set_size(nx, nx)
phi = refrings[i].get_attr("phi")
theta = refrings[i].get_attr("theta")
t = Transform({
"type": "spider",
"phi": phi,
"theta": theta,
"psi": 0.0
})
avg[i].set_attr("xform.projection", t)
for im in xrange(nima):
iref = data[im].get_attr("assign")
gim = data[im].get_attr("group")
if gim == group:
[phi, theta, psi, s2x, s2y] = get_params_proj(data[im])
[alpha, sx, sy, mirror] = params_3D_2D(phi, theta, psi, s2x, s2y)
temp = rot_shift2D(data[im], alpha, sx, sy, mirror, 1.0)
avg[iref] = avg[iref] + temp
avg_csum[iref] = avg_csum[iref] + 1
from utilities import reduce_EMData_to_root
for i in xrange(len(refrings)):
reduce_EMData_to_root(avg[i], myid, main_node)
avg_sum = mpi_reduce(avg_csum[i], 1, MPI_FLOAT, MPI_SUM, 0,
MPI_COMM_WORLD)
outfile_repro = os.path.join(outdir,
"repro_%s%s.hdf" % (itout, modout))
if myid == 0:
outfile = os.path.join(outdir,
"compare_repro_%s%s.hdf" % (itout, modout))
avg[i].write_image(outfile, -1)
t = avg[i].get_attr("xform.projection")
proj = vol.project("pawel", t)
proj.set_attr("xform.projection", t)
proj.set_attr("Raw_im_count", float(avg_sum))
proj.write_image(outfile, -1)
proj.write_image(outfile_repro, -1)
return outfile_repro
def comp_rep(refrings, data, itout, modout, vol, group, nima, nx, myid, main_node, outdir):
import os
from fundamentals import rot_shift2D
from utilities import get_params_proj, params_3D_2D
from mpi import mpi_reduce, MPI_COMM_WORLD, MPI_FLOAT, MPI_SUM
avg = [EMData() for i in xrange(len(refrings))]
avg_csum = [0.0 for i in xrange(len(refrings))]
for i in xrange(len(refrings)):
avg[i] = EMData()
avg[i].set_size(nx,nx)
phi = refrings[i].get_attr("phi")
theta = refrings[i].get_attr("theta")
t = Transform({"type":"spider","phi":phi,"theta":theta,"psi":0.0})
avg[i].set_attr("xform.projection",t)
for im in xrange(nima):
iref = data[im].get_attr("assign")
gim = data[im].get_attr("group")
if gim == group:
[phi, theta, psi, s2x, s2y] = get_params_proj(data[im])
[alpha, sx,sy,mirror] = params_3D_2D(phi,theta,psi,s2x,s2y)
temp = rot_shift2D(data[im],alpha, sx, sy, mirror, 1.0)
avg[iref] = avg[iref] + temp
avg_csum[iref] = avg_csum[iref] + 1
from utilities import reduce_EMData_to_root
for i in xrange(len(refrings)):
reduce_EMData_to_root(avg[i], myid, main_node)
avg_sum = mpi_reduce(avg_csum[i],1,MPI_FLOAT,MPI_SUM,0,MPI_COMM_WORLD)
outfile_repro = os.path.join(outdir, "repro_%s%s.hdf"%(itout,modout))
if myid ==0:
outfile = os.path.join(outdir, "compare_repro_%s%s.hdf"%(itout,modout))
avg[i].write_image(outfile,-1)
t = avg[i].get_attr("xform.projection")
proj = vol.project("pawel",t)
proj.set_attr("xform.projection",t)
proj.set_attr("Raw_im_count", float(avg_sum))
proj.write_image(outfile,-1)
proj.write_image(outfile_repro,-1)
return outfile_repro
def cml_find_structure2(Prj, Ori, Rot, outdir, outname, maxit, first_zero, flag_weights, myid, main_node, number_of_proc):
from projection import cml_export_progress, cml_disc, cml_export_txtagls
import time, sys
from random import shuffle,random
from mpi import MPI_FLOAT, MPI_INT, MPI_SUM, MPI_COMM_WORLD
from mpi import mpi_reduce, mpi_bcast, mpi_barrier
# global vars
global g_i_prj, g_n_prj, g_n_anglst, g_anglst, g_d_psi, g_debug, g_n_lines, g_seq
# list of free orientation
ocp = [-1] * g_n_anglst
if first_zero:
listprj = range(1, g_n_prj)
ocp[0] = 0
else: listprj = range(g_n_prj)
# to stop when the solution oscillates
period_disc = [0, 0, 0]
period_ct = 0
period_th = 2
#if not flag_weights: weights = [1.0] * g_n_lines
# iteration loop
for ite in xrange(maxit):
#print ">>>>>>>>>>>>>>>>>>>>>>>>>>>>>> ite = ", ite, " myid = ", myid
t_start = time.time()
# loop over i prj
change = False
tlistprj = listprj[:]
shuffle(tlistprj)
nnn = len(tlistprj)
tlistprj = mpi_bcast(tlistprj, nnn, MPI_INT, main_node, MPI_COMM_WORLD)
tlistprj = map(int, tlistprj)
"""
if(ite>1 and ite%5 == 0 and ite<140):
if(myid == main_node):
for i in xrange(0,len(tlistprj),5):
ind = 4*i
Ori[ind] = 360.*random()
Ori[ind+1] = 180.*random()
Ori[ind+2] = 360.*random()
Ori[ind+3] = -1
for i in xrange(len(tlistprj)):
ind = 4*i
Ori[ind+3] = float(Ori[ind+3])
nnn = len(Ori)
Ori = mpi_bcast(Ori, nnn, MPI_FLOAT, main_node, MPI_COMM_WORLD)
Ori = map(float, Ori)
for i in xrange(len(tlistprj)):
ind = 4*i
Ori[ind+3] = int(Ori[ind+3])
"""
for iprj in tlistprj:
#print "********************************** iprj = ", iprj, g_n_anglst
# Store current the current orientation
ind = 4*iprj
store_phi = Ori[ind]
store_theta = Ori[ind+1]
store_psi = Ori[ind+2]
cur_agl = Ori[ind+3]
if cur_agl != -1: ocp[cur_agl] = -1
# prepare active index of cml for weighting in order to earn time later
iw = [0] * (g_n_prj - 1)
c = 0
ct = 0
for i in xrange(g_n_prj):
for j in xrange(i+1, g_n_prj):
if i == iprj or j == iprj:
iw[ct] = c
ct += 1
c += 1
# loop over all angles
best_disc_list = [0]*g_n_anglst
best_psi_list = [0]*g_n_anglst
for iagl in xrange(myid, g_n_anglst, number_of_proc):
# if orientation is free
if ocp[iagl] == -1:
# assign new orientation
Ori[ind] = g_anglst[iagl][0]
Ori[ind+1] = g_anglst[iagl][1]
Rot = Util.cml_update_rot(Rot, iprj, Ori[ind], Ori[ind+1], 0.0)
# weights
if flag_weights:
cml = Util.cml_line_in3d(Ori, g_seq, g_n_prj, g_n_lines)
weights = Util.cml_weights(cml)
mw = max(weights)
for i in xrange(g_n_lines): weights[i] = mw - weights[i]
sw = sum(weights)
if sw == 0:
weights = [6.28 / float(g_n_lines)] * g_n_lines
else:
for i in xrange(g_n_lines):
weights[i] /= sw
weights[i] *= weights[i]
# spin all psi
com = Util.cml_line_insino(Rot, iprj, g_n_prj)
if flag_weights:
res = Util.cml_spin_psi(Prj, com, weights, iprj, iw, g_n_psi, g_d_psi, g_n_prj)
else:
res = Util.cml_spin_psi_now(Prj, com, iprj, iw, g_n_psi, g_d_psi, g_n_prj)
# select the best
best_disc_list[iagl] = res[0]
best_psi_list[iagl] = res[1]
if g_debug: cml_export_progress(outdir, ite, iprj, iagl, res[1], res[0], 'progress')
else:
if g_debug: cml_export_progress(outdir, ite, iprj, iagl, -1, -1, 'progress')
best_disc_list = mpi_reduce(best_disc_list, g_n_anglst, MPI_FLOAT, MPI_SUM, main_node, MPI_COMM_WORLD)
best_psi_list = mpi_reduce(best_psi_list, g_n_anglst, MPI_FLOAT, MPI_SUM, main_node, MPI_COMM_WORLD)
best_psi = -1
best_iagl = -1
if myid == main_node:
best_disc = 1.0e20
for iagl in xrange(g_n_anglst):
if best_disc_list[iagl] > 0.0 and best_disc_list[iagl] < best_disc:
best_disc = best_disc_list[iagl]
best_psi = best_psi_list[iagl]
best_iagl = iagl
best_psi = mpi_bcast(best_psi, 1, MPI_FLOAT, main_node, MPI_COMM_WORLD)
best_iagl = mpi_bcast(best_iagl, 1, MPI_INT, main_node, MPI_COMM_WORLD)
best_psi = float(best_psi[0])
best_iagl = int(best_iagl[0])
#print "xxxxx myid = ", myid, " best_psi = ", best_psi, " best_ialg = ", best_iagl
# if change, assign
if best_iagl != cur_agl:
ocp[best_iagl] = iprj
Ori[ind] = g_anglst[best_iagl][0] # phi
Ori[ind+1] = g_anglst[best_iagl][1] # theta
Ori[ind+2] = best_psi * g_d_psi # psi
Ori[ind+3] = best_iagl # index
change = True
else:
if cur_agl != -1: ocp[cur_agl] = iprj
Ori[ind] = store_phi
Ori[ind+1] = store_theta
Ori[ind+2] = store_psi
Ori[ind+3] = cur_agl
Rot = Util.cml_update_rot(Rot, iprj, Ori[ind], Ori[ind+1], Ori[ind+2])
if g_debug: cml_export_progress(outdir, ite, iprj, best_iagl, best_psi * g_d_psi, best_disc, 'choose')
# if one change, compute new full disc
disc = cml_disc(Prj, Ori, Rot, flag_weights)
# display in the progress file
if myid == main_node:
cml_export_txtagls(outdir, outname, Ori, disc, 'Ite: %03i' % (ite + 1))
if not change: break
# to stop when the solution oscillates
period_disc.pop(0)
period_disc.append(disc)
if period_disc[0] == period_disc[2]:
period_ct += 1
if period_ct >= period_th and min(period_disc) == disc and myid == main_node:
angfile = open(outdir + '/' + outname, 'a')
angfile.write('\nSTOP SOLUTION UNSTABLE\n')
angfile.write('Discrepancy period: %s\n' % period_disc)
angfile.close()
break
else:
period_ct = 0
mpi_barrier(MPI_COMM_WORLD)
return Ori, disc, ite
# set initial guess for the value of the grid
psi[:, :] = 1.0
do_force(forf, i1, i2, j1, j2)
#set boundary conditions
bc(psi, i1, i2, j1, j2)
new_psi[:, :] = psi[:, :]
iout = vals.steps / 100
if (iout == 0):
iout = 1
iw = 0
r1 = range(1, (i2 - i1) + 2)
r2 = range(1, (j2 - j1) + 2)
ttot = 0
for i in range(0, vals.steps):
do_transfer(psi, i1, i2, j1, j2)
diff = do_jacobi(psi, new_psi, i1, i2, j1, j2)
diff = mpi.mpi_reduce(diff, 1, mpi.MPI_DOUBLE, mpi.MPI_SUM, 0,
mpi.MPI_COMM_WORLD)
if (myid == 0):
if ((i + 1) % iout) == 0:
print(i + 1, diff[0])
t2 = walltime()
if (myid == 0):
print("total time=", t2 - t1, " time spent in do_jacobi=", ttot)
mpi.mpi_finalize()
# if is acting as the executable call main
#if __name__ == '__main__':
# main()
def ali3d_MPI(stack, ref_vol, outdir, maskfile = None, ir = 1, ou = -1, rs = 1,
xr = "4 2 2 1", yr = "-1", ts = "1 1 0.5 0.25", delta = "10 6 4 4", an = "-1",
center = 0, maxit = 5, term = 95, CTF = False, fourvar = False, snr = 1.0, ref_a = "S", sym = "c1",
sort=True, cutoff=999.99, pix_cutoff="0", two_tail=False, model_jump="1 1 1 1 1", restart=False, save_half=False,
protos=None, oplane=None, lmask=-1, ilmask=-1, findseam=False, vertstep=None, hpars="-1", hsearch="73.0 170.0",
full_output = False, compare_repro = False, compare_ref_free = "-1", ref_free_cutoff= "-1 -1 -1 -1",
wcmask = None, debug = False, recon_pad = 4):
from alignment import Numrinit, prepare_refrings
from utilities import model_circle, get_image, drop_image, get_input_from_string
from utilities import bcast_list_to_all, bcast_number_to_all, reduce_EMData_to_root, bcast_EMData_to_all
from utilities import send_attr_dict
from utilities import get_params_proj, file_type
from fundamentals import rot_avg_image
import os
import types
from utilities import print_begin_msg, print_end_msg, print_msg
from mpi import mpi_bcast, mpi_comm_size, mpi_comm_rank, MPI_FLOAT, MPI_COMM_WORLD, mpi_barrier, mpi_reduce
from mpi import mpi_reduce, MPI_INT, MPI_SUM, mpi_finalize
from filter import filt_ctf
from projection import prep_vol, prgs
from statistics import hist_list, varf3d_MPI, fsc_mask
from numpy import array, bincount, array2string, ones
number_of_proc = mpi_comm_size(MPI_COMM_WORLD)
myid = mpi_comm_rank(MPI_COMM_WORLD)
main_node = 0
if myid == main_node:
if os.path.exists(outdir): ERROR('Output directory exists, please change the name and restart the program', "ali3d_MPI", 1)
os.mkdir(outdir)
mpi_barrier(MPI_COMM_WORLD)
if debug:
from time import sleep
while not os.path.exists(outdir):
print "Node ",myid," waiting..."
sleep(5)
info_file = os.path.join(outdir, "progress%04d"%myid)
finfo = open(info_file, 'w')
else:
finfo = None
mjump = get_input_from_string(model_jump)
xrng = get_input_from_string(xr)
if yr == "-1": yrng = xrng
else : yrng = get_input_from_string(yr)
step = get_input_from_string(ts)
delta = get_input_from_string(delta)
ref_free_cutoff = get_input_from_string(ref_free_cutoff)
pix_cutoff = get_input_from_string(pix_cutoff)
lstp = min(len(xrng), len(yrng), len(step), len(delta))
if an == "-1":
an = [-1] * lstp
else:
an = get_input_from_string(an)
# make sure pix_cutoff is set for all iterations
if len(pix_cutoff)<lstp:
for i in xrange(len(pix_cutoff),lstp):
pix_cutoff.append(pix_cutoff[-1])
# don't waste time on sub-pixel alignment for low-resolution ang incr
for i in range(len(step)):
if (delta[i] > 4 or delta[i] == -1) and step[i] < 1:
step[i] = 1
first_ring = int(ir)
rstep = int(rs)
last_ring = int(ou)
max_iter = int(maxit)
center = int(center)
nrefs = EMUtil.get_image_count( ref_vol )
nmasks = 0
if maskfile:
# read number of masks within each maskfile (mc)
nmasks = EMUtil.get_image_count( maskfile )
# open masks within maskfile (mc)
maskF = EMData.read_images(maskfile, xrange(nmasks))
vol = EMData.read_images(ref_vol, xrange(nrefs))
nx = vol[0].get_xsize()
## make sure box sizes are the same
if myid == main_node:
im=EMData.read_images(stack,[0])
bx = im[0].get_xsize()
if bx!=nx:
print_msg("Error: Stack box size (%i) differs from initial model (%i)\n"%(bx,nx))
sys.exit()
del im,bx
# for helical processing:
helicalrecon = False
if protos is not None or hpars != "-1" or findseam is True:
helicalrecon = True
# if no out-of-plane param set, use 5 degrees
if oplane is None:
oplane=5.0
if protos is not None:
proto = get_input_from_string(protos)
if len(proto) != nrefs:
print_msg("Error: insufficient protofilament numbers supplied")
sys.exit()
if hpars != "-1":
hpars = get_input_from_string(hpars)
if len(hpars) != 2*nrefs:
print_msg("Error: insufficient helical parameters supplied")
sys.exit()
## create helical parameter file for helical reconstruction
if helicalrecon is True and myid == main_node:
from hfunctions import createHpar
# create initial helical parameter files
dp=[0]*nrefs
dphi=[0]*nrefs
vdp=[0]*nrefs
vdphi=[0]*nrefs
for iref in xrange(nrefs):
hpar = os.path.join(outdir,"hpar%02d.spi"%(iref))
params = False
if hpars != "-1":
# if helical parameters explicitly given, set twist & rise
params = [float(hpars[iref*2]),float(hpars[(iref*2)+1])]
dp[iref],dphi[iref],vdp[iref],vdphi[iref] = createHpar(hpar,proto[iref],params,vertstep)
# get values for helical search parameters
hsearch = get_input_from_string(hsearch)
if len(hsearch) != 2:
print_msg("Error: specify outer and inner radii for helical search")
sys.exit()
if last_ring < 0 or last_ring > int(nx/2)-2 : last_ring = int(nx/2) - 2
if myid == main_node:
# import user_functions
# user_func = user_functions.factory[user_func_name]
print_begin_msg("ali3d_MPI")
print_msg("Input stack : %s\n"%(stack))
print_msg("Reference volume : %s\n"%(ref_vol))
print_msg("Output directory : %s\n"%(outdir))
if nmasks > 0:
print_msg("Maskfile (number of masks) : %s (%i)\n"%(maskfile,nmasks))
print_msg("Inner radius : %i\n"%(first_ring))
print_msg("Outer radius : %i\n"%(last_ring))
print_msg("Ring step : %i\n"%(rstep))
print_msg("X search range : %s\n"%(xrng))
print_msg("Y search range : %s\n"%(yrng))
print_msg("Translational step : %s\n"%(step))
print_msg("Angular step : %s\n"%(delta))
print_msg("Angular search range : %s\n"%(an))
print_msg("Maximum iteration : %i\n"%(max_iter))
print_msg("Center type : %i\n"%(center))
print_msg("CTF correction : %s\n"%(CTF))
print_msg("Signal-to-Noise Ratio : %f\n"%(snr))
print_msg("Reference projection method : %s\n"%(ref_a))
print_msg("Symmetry group : %s\n"%(sym))
print_msg("Fourier padding for 3D : %i\n"%(recon_pad))
print_msg("Number of reference models : %i\n"%(nrefs))
print_msg("Sort images between models : %s\n"%(sort))
print_msg("Allow images to jump : %s\n"%(mjump))
print_msg("CC cutoff standard dev : %f\n"%(cutoff))
print_msg("Two tail cutoff : %s\n"%(two_tail))
print_msg("Termination pix error : %f\n"%(term))
print_msg("Pixel error cutoff : %s\n"%(pix_cutoff))
print_msg("Restart : %s\n"%(restart))
print_msg("Full output : %s\n"%(full_output))
print_msg("Compare reprojections : %s\n"%(compare_repro))
print_msg("Compare ref free class avgs : %s\n"%(compare_ref_free))
print_msg("Use cutoff from ref free : %s\n"%(ref_free_cutoff))
if protos:
print_msg("Protofilament numbers : %s\n"%(proto))
print_msg("Using helical search range : %s\n"%hsearch)
if findseam is True:
print_msg("Using seam-based reconstruction\n")
if hpars != "-1":
print_msg("Using hpars : %s\n"%hpars)
if vertstep != None:
print_msg("Using vertical step : %.2f\n"%vertstep)
if save_half is True:
print_msg("Saving even/odd halves\n")
for i in xrange(100) : print_msg("*")
print_msg("\n\n")
if maskfile:
if type(maskfile) is types.StringType: mask3D = get_image(maskfile)
else: mask3D = maskfile
else: mask3D = model_circle(last_ring, nx, nx, nx)
numr = Numrinit(first_ring, last_ring, rstep, "F")
mask2D = model_circle(last_ring,nx,nx) - model_circle(first_ring,nx,nx)
fscmask = model_circle(last_ring,nx,nx,nx)
if CTF:
from filter import filt_ctf
from reconstruction_rjh import rec3D_MPI_noCTF
if myid == main_node:
active = EMUtil.get_all_attributes(stack, 'active')
list_of_particles = []
for im in xrange(len(active)):
if active[im]: list_of_particles.append(im)
del active
nima = len(list_of_particles)
else:
nima = 0
total_nima = bcast_number_to_all(nima, source_node = main_node)
if myid != main_node:
list_of_particles = [-1]*total_nima
list_of_particles = bcast_list_to_all(list_of_particles, source_node = main_node)
image_start, image_end = MPI_start_end(total_nima, number_of_proc, myid)
# create a list of images for each node
list_of_particles = list_of_particles[image_start: image_end]
nima = len(list_of_particles)
if debug:
finfo.write("image_start, image_end: %d %d\n" %(image_start, image_end))
finfo.flush()
data = EMData.read_images(stack, list_of_particles)
t_zero = Transform({"type":"spider","phi":0,"theta":0,"psi":0,"tx":0,"ty":0})
transmulti = [[t_zero for i in xrange(nrefs)] for j in xrange(nima)]
for iref,im in ((iref,im) for iref in xrange(nrefs) for im in xrange(nima)):
if nrefs == 1:
transmulti[im][iref] = data[im].get_attr("xform.projection")
else:
# if multi models, keep track of eulers for all models
try:
transmulti[im][iref] = data[im].get_attr("eulers_txty.%i"%iref)
except:
data[im].set_attr("eulers_txty.%i"%iref,t_zero)
scoremulti = [[0.0 for i in xrange(nrefs)] for j in xrange(nima)]
pixelmulti = [[0.0 for i in xrange(nrefs)] for j in xrange(nima)]
ref_res = [0.0 for x in xrange(nrefs)]
apix = data[0].get_attr('apix_x')
# for oplane parameter, create cylindrical mask
if oplane is not None and myid == main_node:
from hfunctions import createCylMask
cmaskf=os.path.join(outdir, "mask3D_cyl.mrc")
mask3D = createCylMask(data,ou,lmask,ilmask,cmaskf)
# if finding seam of helix, create wedge masks
if findseam is True:
wedgemask=[]
for pf in xrange(nrefs):
wedgemask.append(EMData())
# wedgemask option
if wcmask is not None:
wcmask = get_input_from_string(wcmask)
if len(wcmask) != 3:
print_msg("Error: wcmask option requires 3 values: x y radius")
sys.exit()
# determine if particles have helix info:
try:
data[0].get_attr('h_angle')
original_data = []
boxmask = True
from hfunctions import createBoxMask
except:
boxmask = False
# prepare particles
for im in xrange(nima):
data[im].set_attr('ID', list_of_particles[im])
data[im].set_attr('pix_score', int(0))
if CTF:
# only phaseflip particles, not full CTF correction
ctf_params = data[im].get_attr("ctf")
st = Util.infomask(data[im], mask2D, False)
data[im] -= st[0]
data[im] = filt_ctf(data[im], ctf_params, sign = -1, binary=1)
data[im].set_attr('ctf_applied', 1)
# for window mask:
if boxmask is True:
h_angle = data[im].get_attr("h_angle")
original_data.append(data[im].copy())
bmask = createBoxMask(nx,apix,ou,lmask,h_angle)
data[im]*=bmask
del bmask
if debug:
finfo.write( '%d loaded \n' % nima )
finfo.flush()
if myid == main_node:
# initialize data for the reference preparation function
ref_data = [ mask3D, max(center,0), None, None, None, None ]
# for method -1, switch off centering in user function
from time import time
# this is needed for gathering of pixel errors
disps = []
recvcount = []
disps_score = []
recvcount_score = []
for im in xrange(number_of_proc):
if( im == main_node ):
disps.append(0)
disps_score.append(0)
else:
disps.append(disps[im-1] + recvcount[im-1])
disps_score.append(disps_score[im-1] + recvcount_score[im-1])
ib, ie = MPI_start_end(total_nima, number_of_proc, im)
recvcount.append( ie - ib )
recvcount_score.append((ie-ib)*nrefs)
pixer = [0.0]*nima
cs = [0.0]*3
total_iter = 0
volodd = EMData.read_images(ref_vol, xrange(nrefs))
voleve = EMData.read_images(ref_vol, xrange(nrefs))
if restart:
# recreate initial volumes from alignments stored in header
itout = "000_00"
for iref in xrange(nrefs):
if(nrefs == 1):
modout = ""
else:
modout = "_model_%02d"%(iref)
if(sort):
group = iref
for im in xrange(nima):
imgroup = data[im].get_attr('group')
if imgroup == iref:
data[im].set_attr('xform.projection',transmulti[im][iref])
else:
group = int(999)
for im in xrange(nima):
data[im].set_attr('xform.projection',transmulti[im][iref])
fscfile = os.path.join(outdir, "fsc_%s%s"%(itout,modout))
vol[iref], fscc, volodd[iref], voleve[iref] = rec3D_MPI_noCTF(data, sym, fscmask, fscfile, myid, main_node, index = group, npad = recon_pad)
if myid == main_node:
if helicalrecon:
from hfunctions import processHelicalVol
vstep=None
if vertstep is not None:
vstep=(vdp[iref],vdphi[iref])
print_msg("Old rise and twist for model %i : %8.3f, %8.3f\n"%(iref,dp[iref],dphi[iref]))
hvals=processHelicalVol(vol[iref],voleve[iref],volodd[iref],iref,outdir,itout,
dp[iref],dphi[iref],apix,hsearch,findseam,vstep,wcmask)
(vol[iref],voleve[iref],volodd[iref],dp[iref],dphi[iref],vdp[iref],vdphi[iref])=hvals
print_msg("New rise and twist for model %i : %8.3f, %8.3f\n"%(iref,dp[iref],dphi[iref]))
# get new FSC from symmetrized half volumes
fscc = fsc_mask( volodd[iref], voleve[iref], mask3D, rstep, fscfile)
else:
vol[iref].write_image(os.path.join(outdir, "vol_%s.hdf"%itout),-1)
if save_half is True:
volodd[iref].write_image(os.path.join(outdir, "volodd_%s.hdf"%itout),-1)
voleve[iref].write_image(os.path.join(outdir, "voleve_%s.hdf"%itout),-1)
if nmasks > 1:
# Read mask for multiplying
ref_data[0] = maskF[iref]
ref_data[2] = vol[iref]
ref_data[3] = fscc
# call user-supplied function to prepare reference image, i.e., center and filter it
vol[iref], cs,fl = ref_ali3d(ref_data)
vol[iref].write_image(os.path.join(outdir, "volf_%s.hdf"%(itout)),-1)
if (apix == 1):
res_msg = "Models filtered at spatial frequency of:\t"
res = fl
else:
res_msg = "Models filtered at resolution of: \t"
res = apix / fl
ares = array2string(array(res), precision = 2)
print_msg("%s%s\n\n"%(res_msg,ares))
bcast_EMData_to_all(vol[iref], myid, main_node)
# write out headers, under MPI writing has to be done sequentially
mpi_barrier(MPI_COMM_WORLD)
# projection matching
for N_step in xrange(lstp):
terminate = 0
Iter = -1
while(Iter < max_iter-1 and terminate == 0):
Iter += 1
total_iter += 1
itout = "%03g_%02d" %(delta[N_step], Iter)
if myid == main_node:
print_msg("ITERATION #%3d, inner iteration #%3d\nDelta = %4.1f, an = %5.2f, xrange = %5.2f, yrange = %5.2f, step = %5.2f\n\n"%(N_step, Iter, delta[N_step], an[N_step], xrng[N_step],yrng[N_step],step[N_step]))
for iref in xrange(nrefs):
if myid == main_node: start_time = time()
volft,kb = prep_vol( vol[iref] )
## constrain projections to out of plane parameter
theta1 = None
theta2 = None
if oplane is not None:
theta1 = 90-oplane
theta2 = 90+oplane
refrings = prepare_refrings( volft, kb, nx, delta[N_step], ref_a, sym, numr, MPI=True, phiEqpsi = "Minus", initial_theta=theta1, delta_theta=theta2)
del volft,kb
if myid== main_node:
print_msg( "Time to prepare projections for model %i: %s\n" % (iref, legibleTime(time()-start_time)) )
start_time = time()
for im in xrange( nima ):
data[im].set_attr("xform.projection", transmulti[im][iref])
if an[N_step] == -1:
t1, peak, pixer[im] = proj_ali_incore(data[im],refrings,numr,xrng[N_step],yrng[N_step],step[N_step],finfo)
else:
t1, peak, pixer[im] = proj_ali_incore_local(data[im],refrings,numr,xrng[N_step],yrng[N_step],step[N_step],an[N_step],finfo)
#data[im].set_attr("xform.projection"%iref, t1)
if nrefs > 1: data[im].set_attr("eulers_txty.%i"%iref,t1)
scoremulti[im][iref] = peak
from pixel_error import max_3D_pixel_error
# t1 is the current param, t2 is old
t2 = transmulti[im][iref]
pixelmulti[im][iref] = max_3D_pixel_error(t1,t2,numr[-3])
transmulti[im][iref] = t1
if myid == main_node:
print_msg("Time of alignment for model %i: %s\n"%(iref, legibleTime(time()-start_time)))
start_time = time()
# gather scoring data from all processors
from mpi import mpi_gatherv
scoremultisend = sum(scoremulti,[])
pixelmultisend = sum(pixelmulti,[])
tmp = mpi_gatherv(scoremultisend,len(scoremultisend),MPI_FLOAT, recvcount_score, disps_score, MPI_FLOAT, main_node,MPI_COMM_WORLD)
tmp1 = mpi_gatherv(pixelmultisend,len(pixelmultisend),MPI_FLOAT, recvcount_score, disps_score, MPI_FLOAT, main_node,MPI_COMM_WORLD)
tmp = mpi_bcast(tmp,(total_nima * nrefs), MPI_FLOAT,0, MPI_COMM_WORLD)
tmp1 = mpi_bcast(tmp1,(total_nima * nrefs), MPI_FLOAT,0, MPI_COMM_WORLD)
tmp = map(float,tmp)
tmp1 = map(float,tmp1)
score = array(tmp).reshape(-1,nrefs)
pixelerror = array(tmp1).reshape(-1,nrefs)
score_local = array(scoremulti)
mean_score = score.mean(axis=0)
std_score = score.std(axis=0)
cut = mean_score - (cutoff * std_score)
cut2 = mean_score + (cutoff * std_score)
res_max = score_local.argmax(axis=1)
minus_cc = [0.0 for x in xrange(nrefs)]
minus_pix = [0.0 for x in xrange(nrefs)]
minus_ref = [0.0 for x in xrange(nrefs)]
#output pixel errors
if(myid == main_node):
from statistics import hist_list
lhist = 20
pixmin = pixelerror.min(axis=1)
region, histo = hist_list(pixmin, lhist)
if(region[0] < 0.0): region[0] = 0.0
print_msg("Histogram of pixel errors\n ERROR number of particles\n")
for lhx in xrange(lhist):
print_msg(" %10.3f %7d\n"%(region[lhx], histo[lhx]))
# Terminate if 95% within 1 pixel error
im = 0
for lhx in xrange(lhist):
if(region[lhx] > 1.0): break
im += histo[lhx]
print_msg( "Percent of particles with pixel error < 1: %f\n\n"% (im/float(total_nima)*100))
term_cond = float(term)/100
if(im/float(total_nima) > term_cond):
terminate = 1
print_msg("Terminating internal loop\n")
del region, histo
terminate = mpi_bcast(terminate, 1, MPI_INT, 0, MPI_COMM_WORLD)
terminate = int(terminate[0])
for im in xrange(nima):
if(sort==False):
data[im].set_attr('group',999)
elif (mjump[N_step]==1):
data[im].set_attr('group',int(res_max[im]))
pix_run = data[im].get_attr('pix_score')
if (pix_cutoff[N_step]==1 and (terminate==1 or Iter == max_iter-1)):
if (pixelmulti[im][int(res_max[im])] > 1):
data[im].set_attr('pix_score',int(777))
if (score_local[im][int(res_max[im])]<cut[int(res_max[im])]) or (two_tail and score_local[im][int(res_max[im])]>cut2[int(res_max[im])]):
data[im].set_attr('group',int(888))
minus_cc[int(res_max[im])] = minus_cc[int(res_max[im])] + 1
if(pix_run == 777):
data[im].set_attr('group',int(777))
minus_pix[int(res_max[im])] = minus_pix[int(res_max[im])] + 1
if (compare_ref_free != "-1") and (ref_free_cutoff[N_step] != -1) and (total_iter > 1):
id = data[im].get_attr('ID')
if id in rejects:
data[im].set_attr('group',int(666))
minus_ref[int(res_max[im])] = minus_ref[int(res_max[im])] + 1
minus_cc_tot = mpi_reduce(minus_cc,nrefs,MPI_FLOAT,MPI_SUM,0,MPI_COMM_WORLD)
minus_pix_tot = mpi_reduce(minus_pix,nrefs,MPI_FLOAT,MPI_SUM,0,MPI_COMM_WORLD)
minus_ref_tot = mpi_reduce(minus_ref,nrefs,MPI_FLOAT,MPI_SUM,0,MPI_COMM_WORLD)
if (myid == main_node):
if(sort):
tot_max = score.argmax(axis=1)
res = bincount(tot_max)
else:
res = ones(nrefs) * total_nima
print_msg("Particle distribution: \t\t%s\n"%(res*1.0))
afcut1 = res - minus_cc_tot
afcut2 = afcut1 - minus_pix_tot
afcut3 = afcut2 - minus_ref_tot
print_msg("Particle distribution after cc cutoff:\t\t%s\n"%(afcut1))
print_msg("Particle distribution after pix cutoff:\t\t%s\n"%(afcut2))
print_msg("Particle distribution after ref cutoff:\t\t%s\n\n"%(afcut3))
res = [0.0 for i in xrange(nrefs)]
for iref in xrange(nrefs):
if(center == -1):
from utilities import estimate_3D_center_MPI, rotate_3D_shift
dummy=EMData()
cs[0], cs[1], cs[2], dummy, dummy = estimate_3D_center_MPI(data, total_nima, myid, number_of_proc, main_node)
cs = mpi_bcast(cs, 3, MPI_FLOAT, main_node, MPI_COMM_WORLD)
cs = [-float(cs[0]), -float(cs[1]), -float(cs[2])]
rotate_3D_shift(data, cs)
if(sort):
group = iref
for im in xrange(nima):
imgroup = data[im].get_attr('group')
if imgroup == iref:
data[im].set_attr('xform.projection',transmulti[im][iref])
else:
group = int(999)
for im in xrange(nima):
data[im].set_attr('xform.projection',transmulti[im][iref])
if(nrefs == 1):
modout = ""
else:
modout = "_model_%02d"%(iref)
fscfile = os.path.join(outdir, "fsc_%s%s"%(itout,modout))
vol[iref], fscc, volodd[iref], voleve[iref] = rec3D_MPI_noCTF(data, sym, fscmask, fscfile, myid, main_node, index=group, npad=recon_pad)
if myid == main_node:
print_msg("3D reconstruction time for model %i: %s\n"%(iref, legibleTime(time()-start_time)))
start_time = time()
# Compute Fourier variance
if fourvar:
outvar = os.path.join(outdir, "volVar_%s.hdf"%(itout))
ssnr_file = os.path.join(outdir, "ssnr_%s"%(itout))
varf = varf3d_MPI(data, ssnr_text_file=ssnr_file, mask2D=None, reference_structure=vol[iref], ou=last_ring, rw=1.0, npad=1, CTF=None, sign=1, sym=sym, myid=myid)
if myid == main_node:
print_msg("Time to calculate 3D Fourier variance for model %i: %s\n"%(iref, legibleTime(time()-start_time)))
start_time = time()
varf = 1.0/varf
varf.write_image(outvar,-1)
else: varf = None
if myid == main_node:
if helicalrecon:
from hfunctions import processHelicalVol
vstep=None
if vertstep is not None:
vstep=(vdp[iref],vdphi[iref])
print_msg("Old rise and twist for model %i : %8.3f, %8.3f\n"%(iref,dp[iref],dphi[iref]))
hvals=processHelicalVol(vol[iref],voleve[iref],volodd[iref],iref,outdir,itout,
dp[iref],dphi[iref],apix,hsearch,findseam,vstep,wcmask)
(vol[iref],voleve[iref],volodd[iref],dp[iref],dphi[iref],vdp[iref],vdphi[iref])=hvals
print_msg("New rise and twist for model %i : %8.3f, %8.3f\n"%(iref,dp[iref],dphi[iref]))
# get new FSC from symmetrized half volumes
fscc = fsc_mask( volodd[iref], voleve[iref], mask3D, rstep, fscfile)
print_msg("Time to search and apply helical symmetry for model %i: %s\n\n"%(iref, legibleTime(time()-start_time)))
start_time = time()
else:
vol[iref].write_image(os.path.join(outdir, "vol_%s.hdf"%(itout)),-1)
if save_half is True:
volodd[iref].write_image(os.path.join(outdir, "volodd_%s.hdf"%(itout)),-1)
voleve[iref].write_image(os.path.join(outdir, "voleve_%s.hdf"%(itout)),-1)
if nmasks > 1:
# Read mask for multiplying
ref_data[0] = maskF[iref]
ref_data[2] = vol[iref]
ref_data[3] = fscc
ref_data[4] = varf
# call user-supplied function to prepare reference image, i.e., center and filter it
vol[iref], cs,fl = ref_ali3d(ref_data)
vol[iref].write_image(os.path.join(outdir, "volf_%s.hdf"%(itout)),-1)
if (apix == 1):
res_msg = "Models filtered at spatial frequency of:\t"
res[iref] = fl
else:
res_msg = "Models filtered at resolution of: \t"
res[iref] = apix / fl
del varf
bcast_EMData_to_all(vol[iref], myid, main_node)
if compare_ref_free != "-1": compare_repro = True
if compare_repro:
outfile_repro = comp_rep(refrings, data, itout, modout, vol[iref], group, nima, nx, myid, main_node, outdir)
mpi_barrier(MPI_COMM_WORLD)
if compare_ref_free != "-1":
ref_free_output = os.path.join(outdir,"ref_free_%s%s"%(itout,modout))
rejects = compare(compare_ref_free, outfile_repro,ref_free_output,yrng[N_step], xrng[N_step], rstep,nx,apix,ref_free_cutoff[N_step], number_of_proc, myid, main_node)
# retrieve alignment params from all processors
par_str = ['xform.projection','ID','group']
if nrefs > 1:
for iref in xrange(nrefs):
par_str.append('eulers_txty.%i'%iref)
if myid == main_node:
from utilities import recv_attr_dict
recv_attr_dict(main_node, stack, data, par_str, image_start, image_end, number_of_proc)
else: send_attr_dict(main_node, data, par_str, image_start, image_end)
if myid == main_node:
ares = array2string(array(res), precision = 2)
print_msg("%s%s\n\n"%(res_msg,ares))
dummy = EMData()
if full_output:
nimat = EMUtil.get_image_count(stack)
output_file = os.path.join(outdir, "paramout_%s"%itout)
foutput = open(output_file, 'w')
for im in xrange(nimat):
# save the parameters for each of the models
outstring = ""
dummy.read_image(stack,im,True)
param3d = dummy.get_attr('xform.projection')
g = dummy.get_attr("group")
# retrieve alignments in EMAN-format
pE = param3d.get_params('eman')
outstring += "%f\t%f\t%f\t%f\t%f\t%i\n" %(pE["az"], pE["alt"], pE["phi"], pE["tx"], pE["ty"],g)
foutput.write(outstring)
foutput.close()
del dummy
mpi_barrier(MPI_COMM_WORLD)
# mpi_finalize()
if myid == main_node: print_end_msg("ali3d_MPI")
def main():
progname = os.path.basename(sys.argv[0])
usage = progname + """ input_micrograph_list_file input_micrograph_pattern input_coordinates_pattern output_directory --coordinates_format --box_size=box_size --invert --import_ctf=ctf_file --limit_ctf --resample_ratio=resample_ratio --defocus_error=defocus_error --astigmatism_error=astigmatism_error
Window particles from micrographs in input list file. The coordinates of the particles should be given as input.
Please specify name pattern of input micrographs and coordinates files with a wild card (*). Use the wild card to indicate the place of micrograph ID (e.g. serial number, time stamp, and etc).
The name patterns must be enclosed by single quotes (') or double quotes ("). (Note: sxgui.py automatically adds single quotes (')).
BDB files can not be selected as input micrographs.
sxwindow.py mic_list.txt ./mic*.hdf info/mic*_info.json particles --coordinates_format=eman2 --box_size=64 --invert --import_ctf=outdir_cter/partres/partres.txt
If micrograph list file name is not provided, all files matched with the micrograph name pattern will be processed.
sxwindow.py ./mic*.hdf info/mic*_info.json particles --coordinates_format=eman2 --box_size=64 --invert --import_ctf=outdir_cter/partres/partres.txt
"""
parser = OptionParser(usage, version=SPARXVERSION)
parser.add_option("--coordinates_format", type="string", default="eman1", help="format of input coordinates files: 'sparx', 'eman1', 'eman2', or 'spider'. the coordinates of sparx, eman2, and spider format is particle center. the coordinates of eman1 format is particle box conner associated with the original box size. (default eman1)")
parser.add_option("--box_size", type="int", default=256, help="x and y dimension of square area to be windowed (in pixels): pixel size after resampling is assumed when resample_ratio < 1.0 (default 256)")
parser.add_option("--invert", action="store_true", default=False, help="invert image contrast: recommended for cryo data (default False)")
parser.add_option("--import_ctf", type="string", default="", help="file name of sxcter output: normally partres.txt (default none)")
parser.add_option("--limit_ctf", action="store_true", default=False, help="filter micrographs based on the CTF limit: this option requires --import_ctf. (default False)")
parser.add_option("--resample_ratio", type="float", default=1.0, help="ratio of new to old image size (or old to new pixel size) for resampling: Valid range is 0.0 < resample_ratio <= 1.0. (default 1.0)")
parser.add_option("--defocus_error", type="float", default=1000000.0, help="defocus errror limit: exclude micrographs whose relative defocus error as estimated by sxcter is larger than defocus_error percent. the error is computed as (std dev defocus)/defocus*100%. (default 1000000.0)" )
parser.add_option("--astigmatism_error", type="float", default=360.0, help="astigmatism error limit: Set to zero astigmatism for micrographs whose astigmatism angular error as estimated by sxcter is larger than astigmatism_error degrees. (default 360.0)")
### detect if program is running under MPI
RUNNING_UNDER_MPI = "OMPI_COMM_WORLD_SIZE" in os.environ
main_node = 0
if RUNNING_UNDER_MPI:
from mpi import mpi_init
from mpi import MPI_COMM_WORLD, mpi_comm_rank, mpi_comm_size, mpi_barrier, mpi_reduce, MPI_INT, MPI_SUM
mpi_init(0, [])
myid = mpi_comm_rank(MPI_COMM_WORLD)
number_of_processes = mpi_comm_size(MPI_COMM_WORLD)
else:
number_of_processes = 1
myid = 0
(options, args) = parser.parse_args(sys.argv[1:])
mic_list_file_path = None
mic_pattern = None
coords_pattern = None
error_status = None
while True:
if len(args) < 3 or len(args) > 4:
error_status = ("Please check usage for number of arguments.\n Usage: " + usage + "\n" + "Please run %s -h for help." % (progname), getframeinfo(currentframe()))
break
if len(args) == 3:
mic_pattern = args[0]
coords_pattern = args[1]
out_dir = args[2]
else: # assert(len(args) == 4)
mic_list_file_path = args[0]
mic_pattern = args[1]
coords_pattern = args[2]
out_dir = args[3]
if mic_list_file_path != None:
if os.path.splitext(mic_list_file_path)[1] != ".txt":
error_status = ("Extension of input micrograph list file must be \".txt\". Please check input_micrograph_list_file argument. Run %s -h for help." % (progname), getframeinfo(currentframe()))
break
if mic_pattern[:len("bdb:")].lower() == "bdb":
error_status = ("BDB file can not be selected as input micrographs. Please convert the format, and restart the program. Run %s -h for help." % (progname), getframeinfo(currentframe()))
break
if mic_pattern.find("*") == -1:
error_status = ("Input micrograph file name pattern must contain wild card (*). Please check input_micrograph_pattern argument. Run %s -h for help." % (progname), getframeinfo(currentframe()))
break
if coords_pattern.find("*") == -1:
error_status = ("Input coordinates file name pattern must contain wild card (*). Please check input_coordinates_pattern argument. Run %s -h for help." % (progname), getframeinfo(currentframe()))
break
if myid == main_node:
if os.path.exists(out_dir):
error_status = ("Output directory exists. Please change the name and restart the program.", getframeinfo(currentframe()))
break
break
if_error_then_all_processes_exit_program(error_status)
# Check invalid conditions of options
check_options(options, progname)
mic_name_list = None
error_status = None
if myid == main_node:
if mic_list_file_path != None:
print("Loading micrograph list from %s file ..." % (mic_list_file_path))
mic_name_list = read_text_file(mic_list_file_path)
if len(mic_name_list) == 0:
print("Directory of first micrograph entry is " % (os.path.dirname(mic_name_list[0])))
else: # assert (mic_list_file_path == None)
print("Generating micrograph list in %s directory..." % (os.path.dirname(mic_pattern)))
mic_name_list = glob.glob(mic_pattern)
if len(mic_name_list) == 0:
error_status = ("No micrograph file is found. Please check input_micrograph_pattern and/or input_micrograph_list_file argument. Run %s -h for help." % (progname), getframeinfo(currentframe()))
else:
print("Found %d microgarphs" % len(mic_name_list))
if_error_then_all_processes_exit_program(error_status)
if RUNNING_UNDER_MPI:
mic_name_list = wrap_mpi_bcast(mic_name_list, main_node)
coords_name_list = None
error_status = None
if myid == main_node:
coords_name_list = glob.glob(coords_pattern)
if len(coords_name_list) == 0:
error_status = ("No coordinates file is found. Please check input_coordinates_pattern argument. Run %s -h for help." % (progname), getframeinfo(currentframe()))
if_error_then_all_processes_exit_program(error_status)
if RUNNING_UNDER_MPI:
coords_name_list = wrap_mpi_bcast(coords_name_list, main_node)
##################################################################################################################################################################################################################
##################################################################################################################################################################################################################
##################################################################################################################################################################################################################
# all processes must have access to indices
if options.import_ctf:
i_enum = -1
i_enum += 1; idx_cter_def = i_enum # defocus [um]; index must be same as ctf object format
i_enum += 1; idx_cter_cs = i_enum # Cs [mm]; index must be same as ctf object format
i_enum += 1; idx_cter_vol = i_enum # voltage[kV]; index must be same as ctf object format
i_enum += 1; idx_cter_apix = i_enum # pixel size [A]; index must be same as ctf object format
i_enum += 1; idx_cter_bfactor = i_enum # B-factor [A^2]; index must be same as ctf object format
i_enum += 1; idx_cter_ac = i_enum # amplitude contrast [%]; index must be same as ctf object format
i_enum += 1; idx_cter_astig_amp = i_enum # astigmatism amplitude [um]; index must be same as ctf object format
i_enum += 1; idx_cter_astig_ang = i_enum # astigmatism angle [degree]; index must be same as ctf object format
i_enum += 1; idx_cter_sd_def = i_enum # std dev of defocus [um]
i_enum += 1; idx_cter_sd_astig_amp = i_enum # std dev of ast amp [A]
i_enum += 1; idx_cter_sd_astig_ang = i_enum # std dev of ast angle [degree]
i_enum += 1; idx_cter_cv_def = i_enum # coefficient of variation of defocus [%]
i_enum += 1; idx_cter_cv_astig_amp = i_enum # coefficient of variation of ast amp [%]
i_enum += 1; idx_cter_spectra_diff = i_enum # average of differences between with- and without-astig. experimental 1D spectra at extrema
i_enum += 1; idx_cter_error_def = i_enum # frequency at which signal drops by 50% due to estimated error of defocus alone [1/A]
i_enum += 1; idx_cter_error_astig = i_enum # frequency at which signal drops by 50% due to estimated error of defocus and astigmatism [1/A]
i_enum += 1; idx_cter_error_ctf = i_enum # limit frequency by CTF error [1/A]
i_enum += 1; idx_cter_mic_name = i_enum # micrograph name
i_enum += 1; n_idx_cter = i_enum
# Prepare loop variables
mic_basename_pattern = os.path.basename(mic_pattern) # file pattern without path
mic_baseroot_pattern = os.path.splitext(mic_basename_pattern)[0] # file pattern without path and extension
coords_format = options.coordinates_format.lower()
box_size = options.box_size
box_half = box_size // 2
mask2d = model_circle(box_size//2, box_size, box_size) # Create circular 2D mask to Util.infomask of particle images
resample_ratio = options.resample_ratio
n_mic_process = 0
n_mic_reject_no_coords = 0
n_mic_reject_no_cter_entry = 0
n_global_coords_detect = 0
n_global_coords_process = 0
n_global_coords_reject_out_of_boundary = 0
serial_id_list = []
error_status = None
## not a real while, an if with the opportunity to use break when errors need to be reported
while myid == main_node:
#
# NOTE: 2016/05/24 Toshio Moriya
# Now, ignores the path in mic_pattern and entries of mic_name_list to create serial ID
# Only the basename (file name) in micrograph path must be match
#
# Create list of micrograph serial ID
# Break micrograph name pattern into prefix and suffix to find the head index of the micrograph serial id
#
mic_basename_tokens = mic_basename_pattern.split('*')
# assert (len(mic_basename_tokens) == 2)
serial_id_head_index = len(mic_basename_tokens[0])
# Loop through micrograph names
for mic_name in mic_name_list:
# Find the tail index of the serial id and extract serial id from the micrograph name
mic_basename = os.path.basename(mic_name)
serial_id_tail_index = mic_basename.index(mic_basename_tokens[1])
serial_id = mic_basename[serial_id_head_index:serial_id_tail_index]
serial_id_list.append(serial_id)
# assert (len(serial_id_list) == len(mic_name))
del mic_name_list # Do not need this anymore
# Load CTFs if necessary
if options.import_ctf:
ctf_list = read_text_row(options.import_ctf)
# print("Detected CTF entries : %6d ..." % (len(ctf_list)))
if len(ctf_list) == 0:
error_status = ("No CTF entry is found in %s. Please check --import_ctf option. Run %s -h for help." % (options.import_ctf, progname), getframeinfo(currentframe()))
break
if (len(ctf_list[0]) != n_idx_cter):
error_status = ("Number of columns (%d) must be %d in %s. The format might be old. Please run sxcter.py again." % (len(ctf_list[0]), n_idx_cter, options.import_ctf), getframeinfo(currentframe()))
break
ctf_dict={}
n_reject_defocus_error = 0
ctf_error_limit = [options.defocus_error/100.0, options.astigmatism_error]
for ctf_params in ctf_list:
assert(len(ctf_params) == n_idx_cter)
# mic_baseroot is name of micrograph minus the path and extension
mic_baseroot = os.path.splitext(os.path.basename(ctf_params[idx_cter_mic_name]))[0]
if(ctf_params[idx_cter_sd_def] / ctf_params[idx_cter_def] > ctf_error_limit[0]):
print("Defocus error %f exceeds the threshold. Micrograph %s is rejected." % (ctf_params[idx_cter_sd_def] / ctf_params[idx_cter_def], mic_baseroot))
n_reject_defocus_error += 1
else:
if(ctf_params[idx_cter_sd_astig_ang] > ctf_error_limit[1]):
ctf_params[idx_cter_astig_amp] = 0.0
ctf_params[idx_cter_astig_ang] = 0.0
ctf_dict[mic_baseroot] = ctf_params
del ctf_list # Do not need this anymore
break
if_error_then_all_processes_exit_program(error_status)
if options.import_ctf:
if options.limit_ctf:
cutoff_histogram = [] #@ming compute the histogram for micrographs cut of by ctf_params limit.
##################################################################################################################################################################################################################
##################################################################################################################################################################################################################
##################################################################################################################################################################################################################
restricted_serial_id_list = []
if myid == main_node:
# Loop over serial IDs of micrographs
for serial_id in serial_id_list:
# mic_baseroot is name of micrograph minus the path and extension
mic_baseroot = mic_baseroot_pattern.replace("*", serial_id)
mic_name = mic_pattern.replace("*", serial_id)
coords_name = coords_pattern.replace("*", serial_id)
########### # CHECKS: BEGIN
if coords_name not in coords_name_list:
print(" Cannot read %s. Skipping %s ..." % (coords_name, mic_baseroot))
n_mic_reject_no_coords += 1
continue
# IF mic is in CTER results
if options.import_ctf:
if mic_baseroot not in ctf_dict:
print(" Is not listed in CTER results. Skipping %s ..." % (mic_baseroot))
n_mic_reject_no_cter_entry += 1
continue
else:
ctf_params = ctf_dict[mic_baseroot]
# CHECKS: END
n_mic_process += 1
restricted_serial_id_list.append(serial_id)
# restricted_serial_id_list = restricted_serial_id_list[:128] ## for testing against the nonMPI version
if myid != main_node:
if options.import_ctf:
ctf_dict = None
error_status = None
if len(restricted_serial_id_list) < number_of_processes:
error_status = ('Number of processes (%d) supplied by --np in mpirun cannot be greater than %d (number of micrographs that satisfy all criteria to be processed) ' % (number_of_processes, len(restricted_serial_id_list)), getframeinfo(currentframe()))
if_error_then_all_processes_exit_program(error_status)
## keep a copy of the original output directory where the final bdb will be created
original_out_dir = out_dir
if RUNNING_UNDER_MPI:
mpi_barrier(MPI_COMM_WORLD)
restricted_serial_id_list = wrap_mpi_bcast(restricted_serial_id_list, main_node)
mic_start, mic_end = MPI_start_end(len(restricted_serial_id_list), number_of_processes, myid)
restricted_serial_id_list_not_sliced = restricted_serial_id_list
restricted_serial_id_list = restricted_serial_id_list[mic_start:mic_end]
if options.import_ctf:
ctf_dict = wrap_mpi_bcast(ctf_dict, main_node)
# generate subdirectories of out_dir, one for each process
out_dir = os.path.join(out_dir,"%03d"%myid)
if myid == main_node:
print("Micrographs processed by main process (including percent complete):")
len_processed_by_main_node_divided_by_100 = len(restricted_serial_id_list)/100.0
##################################################################################################################################################################################################################
##################################################################################################################################################################################################################
##################################################################################################################################################################################################################
##### Starting main parallel execution
for my_idx, serial_id in enumerate(restricted_serial_id_list):
mic_baseroot = mic_baseroot_pattern.replace("*", serial_id)
mic_name = mic_pattern.replace("*", serial_id)
coords_name = coords_pattern.replace("*", serial_id)
if myid == main_node:
print(mic_name, " ---> % 2.2f%%"%(my_idx/len_processed_by_main_node_divided_by_100))
mic_img = get_im(mic_name)
# Read coordinates according to the specified format and
# make the coordinates the center of particle image
if coords_format == "sparx":
coords_list = read_text_row(coords_name)
elif coords_format == "eman1":
coords_list = read_text_row(coords_name)
for i in xrange(len(coords_list)):
coords_list[i] = [(coords_list[i][0] + coords_list[i][2] // 2), (coords_list[i][1] + coords_list[i][3] // 2)]
elif coords_format == "eman2":
coords_list = js_open_dict(coords_name)["boxes"]
for i in xrange(len(coords_list)):
coords_list[i] = [coords_list[i][0], coords_list[i][1]]
elif coords_format == "spider":
coords_list = read_text_row(coords_name)
for i in xrange(len(coords_list)):
coords_list[i] = [coords_list[i][2], coords_list[i][3]]
# else: assert (False) # Unreachable code
# Calculate the new pixel size
if options.import_ctf:
ctf_params = ctf_dict[mic_baseroot]
pixel_size_origin = ctf_params[idx_cter_apix]
if resample_ratio < 1.0:
# assert (resample_ratio > 0.0)
new_pixel_size = pixel_size_origin / resample_ratio
print("Resample micrograph to pixel size %6.4f and window segments from resampled micrograph." % new_pixel_size)
else:
# assert (resample_ratio == 1.0)
new_pixel_size = pixel_size_origin
# Set ctf along with new pixel size in resampled micrograph
ctf_params[idx_cter_apix] = new_pixel_size
else:
# assert (not options.import_ctf)
if resample_ratio < 1.0:
# assert (resample_ratio > 0.0)
print("Resample micrograph with ratio %6.4f and window segments from resampled micrograph." % resample_ratio)
# else:
# assert (resample_ratio == 1.0)
# Apply filters to micrograph
fftip(mic_img)
if options.limit_ctf:
# assert (options.import_ctf)
# Cut off frequency components higher than CTF limit
q1, q2 = ctflimit(box_size, ctf_params[idx_cter_def], ctf_params[idx_cter_cs], ctf_params[idx_cter_vol], new_pixel_size)
# This is absolute frequency of CTF limit in scale of original micrograph
if resample_ratio < 1.0:
# assert (resample_ratio > 0.0)
q1 = resample_ratio * q1 / float(box_size) # q1 = (pixel_size_origin / new_pixel_size) * q1/float(box_size)
else:
# assert (resample_ratio == 1.0) -> pixel_size_origin == new_pixel_size -> pixel_size_origin / new_pixel_size == 1.0
q1 = q1 / float(box_size)
if q1 < 0.5:
mic_img = filt_tanl(mic_img, q1, 0.01)
cutoff_histogram.append(q1)
# Cut off frequency components lower than the box size can express
mic_img = fft(filt_gaussh(mic_img, resample_ratio / box_size))
# Resample micrograph, map coordinates, and window segments from resampled micrograph using new coordinates
# after resampling by resample_ratio, new pixel size will be pixel_size/resample_ratio = new_pixel_size
# NOTE: 2015/04/13 Toshio Moriya
# resample() efficiently takes care of the case resample_ratio = 1.0 but
# it does not set apix_*. Even though it sets apix_* when resample_ratio < 1.0 ...
mic_img = resample(mic_img, resample_ratio)
if options.invert:
mic_stats = Util.infomask(mic_img, None, True) # mic_stat[0:mean, 1:SD, 2:min, 3:max]
Util.mul_scalar(mic_img, -1.0)
mic_img += 2 * mic_stats[0]
if options.import_ctf:
from utilities import generate_ctf
ctf_obj = generate_ctf(ctf_params) # indexes 0 to 7 (idx_cter_def to idx_cter_astig_ang) must be same in cter format & ctf object format.
# Prepare loop variables
nx = mic_img.get_xsize()
ny = mic_img.get_ysize()
x0 = nx//2
y0 = ny//2
n_coords_reject_out_of_boundary = 0
local_stack_name = "bdb:%s#" % out_dir + mic_baseroot + '_ptcls'
local_particle_id = 0 # can be different from coordinates_id
# Loop over coordinates
for coords_id in xrange(len(coords_list)):
x = int(coords_list[coords_id][0])
y = int(coords_list[coords_id][1])
if resample_ratio < 1.0:
# assert (resample_ratio > 0.0)
x = int(x * resample_ratio)
y = int(y * resample_ratio)
# else:
# assert(resample_ratio == 1.0)
if( (0 <= x - box_half) and ( x + box_half <= nx ) and (0 <= y - box_half) and ( y + box_half <= ny ) ):
particle_img = Util.window(mic_img, box_size, box_size, 1, x-x0, y-y0)
else:
print("In %s, coordinates ID = %04d (x = %4d, y = %4d, box_size = %4d) is out of micrograph bound, skipping ..." % (mic_baseroot, coords_id, x, y, box_size))
n_coords_reject_out_of_boundary += 1
continue
particle_img = ramp(particle_img)
particle_stats = Util.infomask(particle_img, mask2d, False) # particle_stats[0:mean, 1:SD, 2:min, 3:max]
particle_img -= particle_stats[0]
particle_img /= particle_stats[1]
# NOTE: 2015/04/09 Toshio Moriya
# ptcl_source_image might be redundant information ...
# Consider re-organizing header entries...
particle_img.set_attr("ptcl_source_image", mic_name)
particle_img.set_attr("ptcl_source_coord_id", coords_id)
particle_img.set_attr("ptcl_source_coord", [int(coords_list[coords_id][0]), int(coords_list[coords_id][1])])
particle_img.set_attr("resample_ratio", resample_ratio)
# NOTE: 2015/04/13 Toshio Moriya
# apix_* attributes are updated by resample() only when resample_ratio != 1.0
# Let's make sure header info is consistent by setting apix_* = 1.0
# regardless of options, so it is not passed down the processing line
particle_img.set_attr("apix_x", 1.0)
particle_img.set_attr("apix_y", 1.0)
particle_img.set_attr("apix_z", 1.0)
if options.import_ctf:
particle_img.set_attr("ctf",ctf_obj)
particle_img.set_attr("ctf_applied", 0)
particle_img.set_attr("pixel_size_origin", pixel_size_origin)
# particle_img.set_attr("apix_x", new_pixel_size)
# particle_img.set_attr("apix_y", new_pixel_size)
# particle_img.set_attr("apix_z", new_pixel_size)
# NOTE: 2015/04/13 Toshio Moriya
# Pawel Comment: Micrograph is not supposed to have CTF header info.
# So, let's assume it does not exist & ignore its presence.
# Note that resample() "correctly" updates pixel size of CTF header info if it exists
# elif (particle_img.has_ctff()):
# assert(not options.import_ctf)
# ctf_origin = particle_img.get_attr("ctf_obj")
# pixel_size_origin = round(ctf_origin.apix, 5) # Because SXCTER ouputs up to 5 digits
# particle_img.set_attr("apix_x",pixel_size_origin)
# particle_img.set_attr("apix_y",pixel_size_origin)
# particle_img.set_attr("apix_z",pixel_size_origin)
# print("local_stack_name, local_particle_id", local_stack_name, local_particle_id)
particle_img.write_image(local_stack_name, local_particle_id)
local_particle_id += 1
n_global_coords_detect += len(coords_list)
n_global_coords_process += local_particle_id
n_global_coords_reject_out_of_boundary += n_coords_reject_out_of_boundary
# # MRK_DEBUG: Toshio Moriya 2016/05/03
# # Following codes are for debugging bdb. Delete in future
# result = db_check_dict(local_stack_name)
# print('# MRK_DEBUG: result = db_check_dict(local_stack_name): %s' % (result))
# result = db_list_dicts('bdb:%s' % out_dir)
# print('# MRK_DEBUG: result = db_list_dicts(out_dir): %s' % (result))
# result = db_get_image_info(local_stack_name)
# print('# MRK_DEBUG: result = db_get_image_info(local_stack_name)', result)
# Release the data base of local stack from this process
# so that the subprocess can access to the data base
db_close_dict(local_stack_name)
# # MRK_DEBUG: Toshio Moriya 2016/05/03
# # Following codes are for debugging bdb. Delete in future
# cmd_line = "e2iminfo.py %s" % (local_stack_name)
# print('# MRK_DEBUG: Executing the command: %s' % (cmd_line))
# cmdexecute(cmd_line)
# # MRK_DEBUG: Toshio Moriya 2016/05/03
# # Following codes are for debugging bdb. Delete in future
# cmd_line = "e2iminfo.py bdb:%s#data" % (out_dir)
# print('# MRK_DEBUG: Executing the command: %s' % (cmd_line))
# cmdexecute(cmd_line)
if RUNNING_UNDER_MPI:
if options.import_ctf:
if options.limit_ctf:
cutoff_histogram = wrap_mpi_gatherv(cutoff_histogram, main_node)
if myid == main_node:
if options.limit_ctf:
# Print out the summary of CTF-limit filtering
print(" ")
print("Global summary of CTF-limit filtering (--limit_ctf) ...")
print("Percentage of filtered micrographs: %8.2f\n" % (len(cutoff_histogram) * 100.0 / len(restricted_serial_id_list_not_sliced)))
n_bins = 10
if len(cutoff_histogram) >= n_bins:
from statistics import hist_list
cutoff_region, cutoff_counts = hist_list(cutoff_histogram, n_bins)
print(" Histogram of cut-off frequency")
print(" cut-off counts")
for bin_id in xrange(n_bins):
print(" %14.7f %7d" % (cutoff_region[bin_id], cutoff_counts[bin_id]))
else:
print("The number of filtered micrographs (%d) is less than the number of bins (%d). No histogram is produced." % (len(cutoff_histogram), n_bins))
n_mic_process = mpi_reduce(n_mic_process, 1, MPI_INT, MPI_SUM, main_node, MPI_COMM_WORLD)
n_mic_reject_no_coords = mpi_reduce(n_mic_reject_no_coords, 1, MPI_INT, MPI_SUM, main_node, MPI_COMM_WORLD)
n_mic_reject_no_cter_entry = mpi_reduce(n_mic_reject_no_cter_entry, 1, MPI_INT, MPI_SUM, main_node, MPI_COMM_WORLD)
n_global_coords_detect = mpi_reduce(n_global_coords_detect, 1, MPI_INT, MPI_SUM, main_node, MPI_COMM_WORLD)
n_global_coords_process = mpi_reduce(n_global_coords_process, 1, MPI_INT, MPI_SUM, main_node, MPI_COMM_WORLD)
n_global_coords_reject_out_of_boundary = mpi_reduce(n_global_coords_reject_out_of_boundary, 1, MPI_INT, MPI_SUM, main_node, MPI_COMM_WORLD)
# Print out the summary of all micrographs
if main_node == myid:
print(" ")
print("Global summary of micrographs ...")
print("Detected : %6d" % (len(restricted_serial_id_list_not_sliced)))
print("Processed : %6d" % (n_mic_process))
print("Rejected by no coordinates file : %6d" % (n_mic_reject_no_coords))
print("Rejected by no CTER entry : %6d" % (n_mic_reject_no_cter_entry))
print(" ")
print("Global summary of coordinates ...")
print("Detected : %6d" % (n_global_coords_detect))
print("Processed : %6d" % (n_global_coords_process))
print("Rejected by out of boundary : %6d" % (n_global_coords_reject_out_of_boundary))
# print(" ")
# print("DONE!!!")
mpi_barrier(MPI_COMM_WORLD)
if main_node == myid:
import time
time.sleep(1)
print("\n Creating bdb:%s/data\n"%original_out_dir)
for proc_i in range(number_of_processes):
mic_start, mic_end = MPI_start_end(len(restricted_serial_id_list_not_sliced), number_of_processes, proc_i)
for serial_id in restricted_serial_id_list_not_sliced[mic_start:mic_end]:
e2bdb_command = "e2bdb.py "
mic_baseroot = mic_baseroot_pattern.replace("*", serial_id)
if RUNNING_UNDER_MPI:
e2bdb_command += "bdb:" + os.path.join(original_out_dir,"%03d/"%proc_i) + mic_baseroot + "_ptcls "
else:
e2bdb_command += "bdb:" + os.path.join(original_out_dir, mic_baseroot + "_ptcls ")
e2bdb_command += " --appendvstack=bdb:%s/data 1>/dev/null"%original_out_dir
cmdexecute(e2bdb_command, printing_on_success = False)
print("Done!\n")
if RUNNING_UNDER_MPI:
mpi_barrier(MPI_COMM_WORLD)
from mpi import mpi_finalize
mpi_finalize()
sys.stdout.flush()
sys.exit(0)
def shiftali_MPI(stack,
maskfile=None,
maxit=100,
CTF=False,
snr=1.0,
Fourvar=False,
search_rng=-1,
oneDx=False,
search_rng_y=-1):
number_of_proc = mpi.mpi_comm_size(mpi.MPI_COMM_WORLD)
myid = mpi.mpi_comm_rank(mpi.MPI_COMM_WORLD)
main_node = 0
ftp = file_type(stack)
if myid == main_node:
print_begin_msg("shiftali_MPI")
max_iter = int(maxit)
if myid == main_node:
if ftp == "bdb":
from EMAN2db import db_open_dict
dummy = db_open_dict(stack, True)
nima = EMUtil.get_image_count(stack)
else:
nima = 0
nima = bcast_number_to_all(nima, source_node=main_node)
list_of_particles = list(range(nima))
image_start, image_end = MPI_start_end(nima, number_of_proc, myid)
list_of_particles = list_of_particles[image_start:image_end]
# read nx and ctf_app (if CTF) and broadcast to all nodes
if myid == main_node:
ima = EMData()
ima.read_image(stack, list_of_particles[0], True)
nx = ima.get_xsize()
ny = ima.get_ysize()
if CTF: ctf_app = ima.get_attr_default('ctf_applied', 2)
del ima
else:
nx = 0
ny = 0
if CTF: ctf_app = 0
nx = bcast_number_to_all(nx, source_node=main_node)
ny = bcast_number_to_all(ny, source_node=main_node)
if CTF:
ctf_app = bcast_number_to_all(ctf_app, source_node=main_node)
if ctf_app > 0:
ERROR("data cannot be ctf-applied", myid=myid)
if maskfile == None:
mrad = min(nx, ny)
mask = model_circle(mrad // 2 - 2, nx, ny)
else:
mask = get_im(maskfile)
if CTF:
from sp_filter import filt_ctf
from sp_morphology import ctf_img
ctf_abs_sum = EMData(nx, ny, 1, False)
ctf_2_sum = EMData(nx, ny, 1, False)
else:
ctf_2_sum = None
from sp_global_def import CACHE_DISABLE
if CACHE_DISABLE:
data = EMData.read_images(stack, list_of_particles)
else:
for i in range(number_of_proc):
if myid == i:
data = EMData.read_images(stack, list_of_particles)
if ftp == "bdb": mpi.mpi_barrier(mpi.MPI_COMM_WORLD)
for im in range(len(data)):
data[im].set_attr('ID', list_of_particles[im])
st = Util.infomask(data[im], mask, False)
data[im] -= st[0]
if CTF:
ctf_params = data[im].get_attr("ctf")
ctfimg = ctf_img(nx, ctf_params, ny=ny)
Util.add_img2(ctf_2_sum, ctfimg)
Util.add_img_abs(ctf_abs_sum, ctfimg)
if CTF:
reduce_EMData_to_root(ctf_2_sum, myid, main_node)
reduce_EMData_to_root(ctf_abs_sum, myid, main_node)
else:
ctf_2_sum = None
if CTF:
if myid != main_node:
del ctf_2_sum
del ctf_abs_sum
else:
temp = EMData(nx, ny, 1, False)
for i in range(0, nx, 2):
for j in range(ny):
temp.set_value_at(i, j, snr)
Util.add_img(ctf_2_sum, temp)
del temp
total_iter = 0
# apply initial xform.align2d parameters stored in header
init_params = []
for im in range(len(data)):
t = data[im].get_attr('xform.align2d')
init_params.append(t)
p = t.get_params("2d")
data[im] = rot_shift2D(data[im],
p['alpha'],
sx=p['tx'],
sy=p['ty'],
mirror=p['mirror'],
scale=p['scale'])
# fourier transform all images, and apply ctf if CTF
for im in range(len(data)):
if CTF:
ctf_params = data[im].get_attr("ctf")
data[im] = filt_ctf(fft(data[im]), ctf_params)
else:
data[im] = fft(data[im])
sx_sum = 0
sy_sum = 0
sx_sum_total = 0
sy_sum_total = 0
shift_x = [0.0] * len(data)
shift_y = [0.0] * len(data)
ishift_x = [0.0] * len(data)
ishift_y = [0.0] * len(data)
for Iter in range(max_iter):
if myid == main_node:
start_time = time()
print_msg("Iteration #%4d\n" % (total_iter))
total_iter += 1
avg = EMData(nx, ny, 1, False)
for im in data:
Util.add_img(avg, im)
reduce_EMData_to_root(avg, myid, main_node)
if myid == main_node:
if CTF:
tavg = Util.divn_filter(avg, ctf_2_sum)
else:
tavg = Util.mult_scalar(avg, 1.0 / float(nima))
else:
tavg = EMData(nx, ny, 1, False)
if Fourvar:
bcast_EMData_to_all(tavg, myid, main_node)
vav, rvar = varf2d_MPI(myid, data, tavg, mask, "a", CTF)
if myid == main_node:
if Fourvar:
tavg = fft(Util.divn_img(fft(tavg), vav))
vav_r = Util.pack_complex_to_real(vav)
# normalize and mask tavg in real space
tavg = fft(tavg)
stat = Util.infomask(tavg, mask, False)
tavg -= stat[0]
Util.mul_img(tavg, mask)
# For testing purposes: shift tavg to some random place and see if the centering is still correct
#tavg = rot_shift3D(tavg,sx=3,sy=-4)
tavg = fft(tavg)
if Fourvar: del vav
bcast_EMData_to_all(tavg, myid, main_node)
sx_sum = 0
sy_sum = 0
if search_rng > 0: nwx = 2 * search_rng + 1
else: nwx = nx
if search_rng_y > 0: nwy = 2 * search_rng_y + 1
else: nwy = ny
not_zero = 0
for im in range(len(data)):
if oneDx:
ctx = Util.window(ccf(data[im], tavg), nwx, 1)
p1 = peak_search(ctx)
p1_x = -int(p1[0][3])
ishift_x[im] = p1_x
sx_sum += p1_x
else:
p1 = peak_search(Util.window(ccf(data[im], tavg), nwx, nwy))
p1_x = -int(p1[0][4])
p1_y = -int(p1[0][5])
ishift_x[im] = p1_x
ishift_y[im] = p1_y
sx_sum += p1_x
sy_sum += p1_y
if not_zero == 0:
if (not (ishift_x[im] == 0.0)) or (not (ishift_y[im] == 0.0)):
not_zero = 1
sx_sum = mpi.mpi_reduce(sx_sum, 1, mpi.MPI_INT, mpi.MPI_SUM, main_node,
mpi.MPI_COMM_WORLD)
if not oneDx:
sy_sum = mpi.mpi_reduce(sy_sum, 1, mpi.MPI_INT, mpi.MPI_SUM,
main_node, mpi.MPI_COMM_WORLD)
if myid == main_node:
sx_sum_total = int(sx_sum[0])
if not oneDx:
sy_sum_total = int(sy_sum[0])
else:
sx_sum_total = 0
sy_sum_total = 0
sx_sum_total = bcast_number_to_all(sx_sum_total, source_node=main_node)
if not oneDx:
sy_sum_total = bcast_number_to_all(sy_sum_total,
source_node=main_node)
sx_ave = round(float(sx_sum_total) / nima)
sy_ave = round(float(sy_sum_total) / nima)
for im in range(len(data)):
p1_x = ishift_x[im] - sx_ave
p1_y = ishift_y[im] - sy_ave
params2 = {
"filter_type": Processor.fourier_filter_types.SHIFT,
"x_shift": p1_x,
"y_shift": p1_y,
"z_shift": 0.0
}
data[im] = Processor.EMFourierFilter(data[im], params2)
shift_x[im] += p1_x
shift_y[im] += p1_y
# stop if all shifts are zero
not_zero = mpi.mpi_reduce(not_zero, 1, mpi.MPI_INT, mpi.MPI_SUM,
main_node, mpi.MPI_COMM_WORLD)
if myid == main_node:
not_zero_all = int(not_zero[0])
else:
not_zero_all = 0
not_zero_all = bcast_number_to_all(not_zero_all, source_node=main_node)
if myid == main_node:
print_msg("Time of iteration = %12.2f\n" % (time() - start_time))
start_time = time()
if not_zero_all == 0: break
#for im in xrange(len(data)): data[im] = fft(data[im]) This should not be required as only header information is used
# combine shifts found with the original parameters
for im in range(len(data)):
t0 = init_params[im]
t1 = Transform()
t1.set_params({
"type": "2D",
"alpha": 0,
"scale": t0.get_scale(),
"mirror": 0,
"tx": shift_x[im],
"ty": shift_y[im]
})
# combine t0 and t1
tt = t1 * t0
data[im].set_attr("xform.align2d", tt)
# write out headers and STOP, under MPI writing has to be done sequentially
mpi.mpi_barrier(mpi.MPI_COMM_WORLD)
par_str = ["xform.align2d", "ID"]
if myid == main_node:
from sp_utilities import file_type
if (file_type(stack) == "bdb"):
from sp_utilities import recv_attr_dict_bdb
recv_attr_dict_bdb(main_node, stack, data, par_str, image_start,
image_end, number_of_proc)
else:
from sp_utilities import recv_attr_dict
recv_attr_dict(main_node, stack, data, par_str, image_start,
image_end, number_of_proc)
else:
send_attr_dict(main_node, data, par_str, image_start, image_end)
if myid == main_node: print_end_msg("shiftali_MPI")
parent = mpi.mpi_comm_get_parent() parentSize = mpi.mpi_comm_size(parent) print "parentSize", parentSize tod = stamp() s = sys.argv[1] + "%2.2d" % myid print "hello from python worker", myid, " writing to ", s x = array([5, 3, 4, 2], 'i') print "starting bcast" buffer = mpi.mpi_bcast(x, 4, mpi.MPI_INT, 0, parent) out = open(s, "w") out.write(str(buffer)) out.write(tod + "\n") out.close() print myid, " got ", buffer junk = mpi.mpi_scatter(x, 1, mpi.MPI_INT, 1, mpi.MPI_INT, 0, parent) print myid, " got scatter ", junk back = mpi.mpi_recv(1, mpi.MPI_INT, 0, 1234, parent) back[0] = back[0] + 1 mpi.mpi_send(back, 1, mpi.MPI_INT, 0, 5678, parent) dummy = myid final = mpi.mpi_reduce(dummy, 1, mpi.MPI_INT, mpi.MPI_SUM, 0, parent) sleep(10) mpi.mpi_comm_free(parent) mpi.mpi_finalize()
mpi_root = 0
#each processor will get count elements from the root
count = 4
# in python we do not need to preallocate the array myray
# we do need to assign a dummy value to the send_ray
send_ray = zeros(0, "i")
if myid == mpi_root:
size = count * numnodes
send_ray = zeros(size, "i")
for i in range(0, size):
send_ray[i] = i
#send different data to each processor
myray = mpi.mpi_scatter(send_ray, count, mpi.MPI_INT, count, mpi.MPI_INT,
mpi_root, mpi.MPI_COMM_WORLD)
#each processor does a local sum
total = 0
for i in range(0, count):
total = total + myray[i]
print "myid=", myid, "total=", total
#reduce back to the root and print
back_ray = mpi.mpi_reduce(total, 1, mpi.MPI_INT, mpi.MPI_SUM, mpi_root,
mpi.MPI_COMM_WORLD)
if myid == mpi_root:
print "results from all processors=", back_ray
mpi.mpi_finalize()
def shiftali_MPI(stack, maskfile=None, maxit=100, CTF=False, snr=1.0, Fourvar=False, search_rng=-1, oneDx=False, search_rng_y=-1):
from applications import MPI_start_end
from utilities import model_circle, model_blank, get_image, peak_search, get_im
from utilities import reduce_EMData_to_root, bcast_EMData_to_all, send_attr_dict, file_type, bcast_number_to_all, bcast_list_to_all
from statistics import varf2d_MPI
from fundamentals import fft, ccf, rot_shift3D, rot_shift2D
from utilities import get_params2D, set_params2D
from utilities import print_msg, print_begin_msg, print_end_msg
import os
import sys
from mpi import mpi_init, mpi_comm_size, mpi_comm_rank, MPI_COMM_WORLD
from mpi import mpi_reduce, mpi_bcast, mpi_barrier, mpi_gatherv
from mpi import MPI_SUM, MPI_FLOAT, MPI_INT
from EMAN2 import Processor
from time import time
number_of_proc = mpi_comm_size(MPI_COMM_WORLD)
myid = mpi_comm_rank(MPI_COMM_WORLD)
main_node = 0
ftp = file_type(stack)
if myid == main_node:
print_begin_msg("shiftali_MPI")
max_iter=int(maxit)
if myid == main_node:
if ftp == "bdb":
from EMAN2db import db_open_dict
dummy = db_open_dict(stack, True)
nima = EMUtil.get_image_count(stack)
else:
nima = 0
nima = bcast_number_to_all(nima, source_node = main_node)
list_of_particles = range(nima)
image_start, image_end = MPI_start_end(nima, number_of_proc, myid)
list_of_particles = list_of_particles[image_start: image_end]
# read nx and ctf_app (if CTF) and broadcast to all nodes
if myid == main_node:
ima = EMData()
ima.read_image(stack, list_of_particles[0], True)
nx = ima.get_xsize()
ny = ima.get_ysize()
if CTF: ctf_app = ima.get_attr_default('ctf_applied', 2)
del ima
else:
nx = 0
ny = 0
if CTF: ctf_app = 0
nx = bcast_number_to_all(nx, source_node = main_node)
ny = bcast_number_to_all(ny, source_node = main_node)
if CTF:
ctf_app = bcast_number_to_all(ctf_app, source_node = main_node)
if ctf_app > 0: ERROR("data cannot be ctf-applied", "shiftali_MPI", 1, myid)
if maskfile == None:
mrad = min(nx, ny)
mask = model_circle(mrad//2-2, nx, ny)
else:
mask = get_im(maskfile)
if CTF:
from filter import filt_ctf
from morphology import ctf_img
ctf_abs_sum = EMData(nx, ny, 1, False)
ctf_2_sum = EMData(nx, ny, 1, False)
else:
ctf_2_sum = None
from global_def import CACHE_DISABLE
if CACHE_DISABLE:
data = EMData.read_images(stack, list_of_particles)
else:
for i in xrange(number_of_proc):
if myid == i:
data = EMData.read_images(stack, list_of_particles)
if ftp == "bdb": mpi_barrier(MPI_COMM_WORLD)
for im in xrange(len(data)):
data[im].set_attr('ID', list_of_particles[im])
st = Util.infomask(data[im], mask, False)
data[im] -= st[0]
if CTF:
ctf_params = data[im].get_attr("ctf")
ctfimg = ctf_img(nx, ctf_params, ny=ny)
Util.add_img2(ctf_2_sum, ctfimg)
Util.add_img_abs(ctf_abs_sum, ctfimg)
if CTF:
reduce_EMData_to_root(ctf_2_sum, myid, main_node)
reduce_EMData_to_root(ctf_abs_sum, myid, main_node)
else: ctf_2_sum = None
if CTF:
if myid != main_node:
del ctf_2_sum
del ctf_abs_sum
else:
temp = EMData(nx, ny, 1, False)
for i in xrange(0,nx,2):
for j in xrange(ny):
temp.set_value_at(i,j,snr)
Util.add_img(ctf_2_sum, temp)
del temp
total_iter = 0
# apply initial xform.align2d parameters stored in header
init_params = []
for im in xrange(len(data)):
t = data[im].get_attr('xform.align2d')
init_params.append(t)
p = t.get_params("2d")
data[im] = rot_shift2D(data[im], p['alpha'], sx=p['tx'], sy=p['ty'], mirror=p['mirror'], scale=p['scale'])
# fourier transform all images, and apply ctf if CTF
for im in xrange(len(data)):
if CTF:
ctf_params = data[im].get_attr("ctf")
data[im] = filt_ctf(fft(data[im]), ctf_params)
else:
data[im] = fft(data[im])
sx_sum=0
sy_sum=0
sx_sum_total=0
sy_sum_total=0
shift_x = [0.0]*len(data)
shift_y = [0.0]*len(data)
ishift_x = [0.0]*len(data)
ishift_y = [0.0]*len(data)
for Iter in xrange(max_iter):
if myid == main_node:
start_time = time()
print_msg("Iteration #%4d\n"%(total_iter))
total_iter += 1
avg = EMData(nx, ny, 1, False)
for im in data: Util.add_img(avg, im)
reduce_EMData_to_root(avg, myid, main_node)
if myid == main_node:
if CTF:
tavg = Util.divn_filter(avg, ctf_2_sum)
else: tavg = Util.mult_scalar(avg, 1.0/float(nima))
else:
tavg = EMData(nx, ny, 1, False)
if Fourvar:
bcast_EMData_to_all(tavg, myid, main_node)
vav, rvar = varf2d_MPI(myid, data, tavg, mask, "a", CTF)
if myid == main_node:
if Fourvar:
tavg = fft(Util.divn_img(fft(tavg), vav))
vav_r = Util.pack_complex_to_real(vav)
# normalize and mask tavg in real space
tavg = fft(tavg)
stat = Util.infomask( tavg, mask, False )
tavg -= stat[0]
Util.mul_img(tavg, mask)
# For testing purposes: shift tavg to some random place and see if the centering is still correct
#tavg = rot_shift3D(tavg,sx=3,sy=-4)
tavg = fft(tavg)
if Fourvar: del vav
bcast_EMData_to_all(tavg, myid, main_node)
sx_sum=0
sy_sum=0
if search_rng > 0: nwx = 2*search_rng+1
else: nwx = nx
if search_rng_y > 0: nwy = 2*search_rng_y+1
else: nwy = ny
not_zero = 0
for im in xrange(len(data)):
if oneDx:
ctx = Util.window(ccf(data[im],tavg),nwx,1)
p1 = peak_search(ctx)
p1_x = -int(p1[0][3])
ishift_x[im] = p1_x
sx_sum += p1_x
else:
p1 = peak_search(Util.window(ccf(data[im],tavg), nwx,nwy))
p1_x = -int(p1[0][4])
p1_y = -int(p1[0][5])
ishift_x[im] = p1_x
ishift_y[im] = p1_y
sx_sum += p1_x
sy_sum += p1_y
if not_zero == 0:
if (not(ishift_x[im] == 0.0)) or (not(ishift_y[im] == 0.0)):
not_zero = 1
sx_sum = mpi_reduce(sx_sum, 1, MPI_INT, MPI_SUM, main_node, MPI_COMM_WORLD)
if not oneDx:
sy_sum = mpi_reduce(sy_sum, 1, MPI_INT, MPI_SUM, main_node, MPI_COMM_WORLD)
if myid == main_node:
sx_sum_total = int(sx_sum[0])
if not oneDx:
sy_sum_total = int(sy_sum[0])
else:
sx_sum_total = 0
sy_sum_total = 0
sx_sum_total = bcast_number_to_all(sx_sum_total, source_node = main_node)
if not oneDx:
sy_sum_total = bcast_number_to_all(sy_sum_total, source_node = main_node)
sx_ave = round(float(sx_sum_total)/nima)
sy_ave = round(float(sy_sum_total)/nima)
for im in xrange(len(data)):
p1_x = ishift_x[im] - sx_ave
p1_y = ishift_y[im] - sy_ave
params2 = {"filter_type" : Processor.fourier_filter_types.SHIFT, "x_shift" : p1_x, "y_shift" : p1_y, "z_shift" : 0.0}
data[im] = Processor.EMFourierFilter(data[im], params2)
shift_x[im] += p1_x
shift_y[im] += p1_y
# stop if all shifts are zero
not_zero = mpi_reduce(not_zero, 1, MPI_INT, MPI_SUM, main_node, MPI_COMM_WORLD)
if myid == main_node:
not_zero_all = int(not_zero[0])
else:
not_zero_all = 0
not_zero_all = bcast_number_to_all(not_zero_all, source_node = main_node)
if myid == main_node:
print_msg("Time of iteration = %12.2f\n"%(time()-start_time))
start_time = time()
if not_zero_all == 0: break
#for im in xrange(len(data)): data[im] = fft(data[im]) This should not be required as only header information is used
# combine shifts found with the original parameters
for im in xrange(len(data)):
t0 = init_params[im]
t1 = Transform()
t1.set_params({"type":"2D","alpha":0,"scale":t0.get_scale(),"mirror":0,"tx":shift_x[im],"ty":shift_y[im]})
# combine t0 and t1
tt = t1*t0
data[im].set_attr("xform.align2d", tt)
# write out headers and STOP, under MPI writing has to be done sequentially
mpi_barrier(MPI_COMM_WORLD)
par_str = ["xform.align2d", "ID"]
if myid == main_node:
from utilities import file_type
if(file_type(stack) == "bdb"):
from utilities import recv_attr_dict_bdb
recv_attr_dict_bdb(main_node, stack, data, par_str, image_start, image_end, number_of_proc)
else:
from utilities import recv_attr_dict
recv_attr_dict(main_node, stack, data, par_str, image_start, image_end, number_of_proc)
else: send_attr_dict(main_node, data, par_str, image_start, image_end)
if myid == main_node: print_end_msg("shiftali_MPI")
def helicalshiftali_MPI(stack, maskfile=None, maxit=100, CTF=False, snr=1.0, Fourvar=False, search_rng=-1):
from applications import MPI_start_end
from utilities import model_circle, model_blank, get_image, peak_search, get_im, pad
from utilities import reduce_EMData_to_root, bcast_EMData_to_all, send_attr_dict, file_type, bcast_number_to_all, bcast_list_to_all
from statistics import varf2d_MPI
from fundamentals import fft, ccf, rot_shift3D, rot_shift2D, fshift
from utilities import get_params2D, set_params2D, chunks_distribution
from utilities import print_msg, print_begin_msg, print_end_msg
import os
import sys
from mpi import mpi_init, mpi_comm_size, mpi_comm_rank, MPI_COMM_WORLD
from mpi import mpi_reduce, mpi_bcast, mpi_barrier, mpi_gatherv
from mpi import MPI_SUM, MPI_FLOAT, MPI_INT
from time import time
from pixel_error import ordersegments
from math import sqrt, atan2, tan, pi
nproc = mpi_comm_size(MPI_COMM_WORLD)
myid = mpi_comm_rank(MPI_COMM_WORLD)
main_node = 0
ftp = file_type(stack)
if myid == main_node:
print_begin_msg("helical-shiftali_MPI")
max_iter=int(maxit)
if( myid == main_node):
infils = EMUtil.get_all_attributes(stack, "filament")
ptlcoords = EMUtil.get_all_attributes(stack, 'ptcl_source_coord')
filaments = ordersegments(infils, ptlcoords)
total_nfils = len(filaments)
inidl = [0]*total_nfils
for i in xrange(total_nfils): inidl[i] = len(filaments[i])
linidl = sum(inidl)
nima = linidl
tfilaments = []
for i in xrange(total_nfils): tfilaments += filaments[i]
del filaments
else:
total_nfils = 0
linidl = 0
total_nfils = bcast_number_to_all(total_nfils, source_node = main_node)
if myid != main_node:
inidl = [-1]*total_nfils
inidl = bcast_list_to_all(inidl, myid, source_node = main_node)
linidl = bcast_number_to_all(linidl, source_node = main_node)
if myid != main_node:
tfilaments = [-1]*linidl
tfilaments = bcast_list_to_all(tfilaments, myid, source_node = main_node)
filaments = []
iendi = 0
for i in xrange(total_nfils):
isti = iendi
iendi = isti+inidl[i]
filaments.append(tfilaments[isti:iendi])
del tfilaments,inidl
if myid == main_node:
print_msg( "total number of filaments: %d"%total_nfils)
if total_nfils< nproc:
ERROR('number of CPUs (%i) is larger than the number of filaments (%i), please reduce the number of CPUs used'%(nproc, total_nfils), "ehelix_MPI", 1,myid)
# balanced load
temp = chunks_distribution([[len(filaments[i]), i] for i in xrange(len(filaments))], nproc)[myid:myid+1][0]
filaments = [filaments[temp[i][1]] for i in xrange(len(temp))]
nfils = len(filaments)
#filaments = [[0,1]]
#print "filaments",filaments
list_of_particles = []
indcs = []
k = 0
for i in xrange(nfils):
list_of_particles += filaments[i]
k1 = k+len(filaments[i])
indcs.append([k,k1])
k = k1
data = EMData.read_images(stack, list_of_particles)
ldata = len(data)
print "ldata=", ldata
nx = data[0].get_xsize()
ny = data[0].get_ysize()
if maskfile == None:
mrad = min(nx, ny)//2-2
mask = pad( model_blank(2*mrad+1, ny, 1, 1.0), nx, ny, 1, 0.0)
else:
mask = get_im(maskfile)
# apply initial xform.align2d parameters stored in header
init_params = []
for im in xrange(ldata):
t = data[im].get_attr('xform.align2d')
init_params.append(t)
p = t.get_params("2d")
data[im] = rot_shift2D(data[im], p['alpha'], p['tx'], p['ty'], p['mirror'], p['scale'])
if CTF:
from filter import filt_ctf
from morphology import ctf_img
ctf_abs_sum = EMData(nx, ny, 1, False)
ctf_2_sum = EMData(nx, ny, 1, False)
else:
ctf_2_sum = None
ctf_abs_sum = None
from utilities import info
for im in xrange(ldata):
data[im].set_attr('ID', list_of_particles[im])
st = Util.infomask(data[im], mask, False)
data[im] -= st[0]
if CTF:
ctf_params = data[im].get_attr("ctf")
qctf = data[im].get_attr("ctf_applied")
if qctf == 0:
data[im] = filt_ctf(fft(data[im]), ctf_params)
data[im].set_attr('ctf_applied', 1)
elif qctf != 1:
ERROR('Incorrectly set qctf flag', "helicalshiftali_MPI", 1,myid)
ctfimg = ctf_img(nx, ctf_params, ny=ny)
Util.add_img2(ctf_2_sum, ctfimg)
Util.add_img_abs(ctf_abs_sum, ctfimg)
else: data[im] = fft(data[im])
del list_of_particles
if CTF:
reduce_EMData_to_root(ctf_2_sum, myid, main_node)
reduce_EMData_to_root(ctf_abs_sum, myid, main_node)
if CTF:
if myid != main_node:
del ctf_2_sum
del ctf_abs_sum
else:
temp = EMData(nx, ny, 1, False)
tsnr = 1./snr
for i in xrange(0,nx+2,2):
for j in xrange(ny):
temp.set_value_at(i,j,tsnr)
temp.set_value_at(i+1,j,0.0)
#info(ctf_2_sum)
Util.add_img(ctf_2_sum, temp)
#info(ctf_2_sum)
del temp
total_iter = 0
shift_x = [0.0]*ldata
for Iter in xrange(max_iter):
if myid == main_node:
start_time = time()
print_msg("Iteration #%4d\n"%(total_iter))
total_iter += 1
avg = EMData(nx, ny, 1, False)
for im in xrange(ldata):
Util.add_img(avg, fshift(data[im], shift_x[im]))
reduce_EMData_to_root(avg, myid, main_node)
if myid == main_node:
if CTF: tavg = Util.divn_filter(avg, ctf_2_sum)
else: tavg = Util.mult_scalar(avg, 1.0/float(nima))
else:
tavg = model_blank(nx,ny)
if Fourvar:
bcast_EMData_to_all(tavg, myid, main_node)
vav, rvar = varf2d_MPI(myid, data, tavg, mask, "a", CTF)
if myid == main_node:
if Fourvar:
tavg = fft(Util.divn_img(fft(tavg), vav))
vav_r = Util.pack_complex_to_real(vav)
# normalize and mask tavg in real space
tavg = fft(tavg)
stat = Util.infomask( tavg, mask, False )
tavg -= stat[0]
Util.mul_img(tavg, mask)
tavg.write_image("tavg.hdf",Iter)
# For testing purposes: shift tavg to some random place and see if the centering is still correct
#tavg = rot_shift3D(tavg,sx=3,sy=-4)
if Fourvar: del vav
bcast_EMData_to_all(tavg, myid, main_node)
tavg = fft(tavg)
sx_sum = 0.0
nxc = nx//2
for ifil in xrange(nfils):
"""
# Calculate filament average
avg = EMData(nx, ny, 1, False)
filnima = 0
for im in xrange(indcs[ifil][0], indcs[ifil][1]):
Util.add_img(avg, data[im])
filnima += 1
tavg = Util.mult_scalar(avg, 1.0/float(filnima))
"""
# Calculate 1D ccf between each segment and filament average
nsegms = indcs[ifil][1]-indcs[ifil][0]
ctx = [None]*nsegms
pcoords = [None]*nsegms
for im in xrange(indcs[ifil][0], indcs[ifil][1]):
ctx[im-indcs[ifil][0]] = Util.window(ccf(tavg, data[im]), nx, 1)
pcoords[im-indcs[ifil][0]] = data[im].get_attr('ptcl_source_coord')
#ctx[im-indcs[ifil][0]].write_image("ctx.hdf",im-indcs[ifil][0])
#print " CTX ",myid,im,Util.infomask(ctx[im-indcs[ifil][0]], None, True)
# search for best x-shift
cents = nsegms//2
dst = sqrt(max((pcoords[cents][0] - pcoords[0][0])**2 + (pcoords[cents][1] - pcoords[0][1])**2, (pcoords[cents][0] - pcoords[-1][0])**2 + (pcoords[cents][1] - pcoords[-1][1])**2))
maxincline = atan2(ny//2-2-float(search_rng),dst)
kang = int(dst*tan(maxincline)+0.5)
#print " settings ",nsegms,cents,dst,search_rng,maxincline,kang
# ## C code for alignment. @ming
results = [0.0]*3;
results = Util.helixshiftali(ctx, pcoords, nsegms, maxincline, kang, search_rng,nxc)
sib = int(results[0])
bang = results[1]
qm = results[2]
#print qm, sib, bang
# qm = -1.e23
#
# for six in xrange(-search_rng, search_rng+1,1):
# q0 = ctx[cents].get_value_at(six+nxc)
# for incline in xrange(kang+1):
# qt = q0
# qu = q0
# if(kang>0): tang = tan(maxincline/kang*incline)
# else: tang = 0.0
# for kim in xrange(cents+1,nsegms):
# dst = sqrt((pcoords[cents][0] - pcoords[kim][0])**2 + (pcoords[cents][1] - pcoords[kim][1])**2)
# xl = dst*tang+six+nxc
# ixl = int(xl)
# dxl = xl - ixl
# #print " A ", ifil,six,incline,kim,xl,ixl,dxl
# qt += (1.0-dxl)*ctx[kim].get_value_at(ixl) + dxl*ctx[kim].get_value_at(ixl+1)
# xl = -dst*tang+six+nxc
# ixl = int(xl)
# dxl = xl - ixl
# qu += (1.0-dxl)*ctx[kim].get_value_at(ixl) + dxl*ctx[kim].get_value_at(ixl+1)
# for kim in xrange(cents):
# dst = sqrt((pcoords[cents][0] - pcoords[kim][0])**2 + (pcoords[cents][1] - pcoords[kim][1])**2)
# xl = -dst*tang+six+nxc
# ixl = int(xl)
# dxl = xl - ixl
# qt += (1.0-dxl)*ctx[kim].get_value_at(ixl) + dxl*ctx[kim].get_value_at(ixl+1)
# xl = dst*tang+six+nxc
# ixl = int(xl)
# dxl = xl - ixl
# qu += (1.0-dxl)*ctx[kim].get_value_at(ixl) + dxl*ctx[kim].get_value_at(ixl+1)
# if( qt > qm ):
# qm = qt
# sib = six
# bang = tang
# if( qu > qm ):
# qm = qu
# sib = six
# bang = -tang
#if incline == 0: print "incline = 0 ",six,tang,qt,qu
#print qm,six,sib,bang
#print " got results ",indcs[ifil][0], indcs[ifil][1], ifil,myid,qm,sib,tang,bang,len(ctx),Util.infomask(ctx[0], None, True)
for im in xrange(indcs[ifil][0], indcs[ifil][1]):
kim = im-indcs[ifil][0]
dst = sqrt((pcoords[cents][0] - pcoords[kim][0])**2 + (pcoords[cents][1] - pcoords[kim][1])**2)
if(kim < cents): xl = -dst*bang+sib
else: xl = dst*bang+sib
shift_x[im] = xl
# Average shift
sx_sum += shift_x[indcs[ifil][0]+cents]
# #print myid,sx_sum,total_nfils
sx_sum = mpi_reduce(sx_sum, 1, MPI_FLOAT, MPI_SUM, main_node, MPI_COMM_WORLD)
if myid == main_node:
sx_sum = float(sx_sum[0])/total_nfils
print_msg("Average shift %6.2f\n"%(sx_sum))
else:
sx_sum = 0.0
sx_sum = 0.0
sx_sum = bcast_number_to_all(sx_sum, source_node = main_node)
for im in xrange(ldata):
shift_x[im] -= sx_sum
#print " %3d %6.3f"%(im,shift_x[im])
#exit()
# combine shifts found with the original parameters
for im in xrange(ldata):
t1 = Transform()
##import random
##shix=random.randint(-10, 10)
##t1.set_params({"type":"2D","tx":shix})
t1.set_params({"type":"2D","tx":shift_x[im]})
# combine t0 and t1
tt = t1*init_params[im]
data[im].set_attr("xform.align2d", tt)
# write out headers and STOP, under MPI writing has to be done sequentially
mpi_barrier(MPI_COMM_WORLD)
par_str = ["xform.align2d", "ID"]
if myid == main_node:
from utilities import file_type
if(file_type(stack) == "bdb"):
from utilities import recv_attr_dict_bdb
recv_attr_dict_bdb(main_node, stack, data, par_str, 0, ldata, nproc)
else:
from utilities import recv_attr_dict
recv_attr_dict(main_node, stack, data, par_str, 0, ldata, nproc)
else: send_attr_dict(main_node, data, par_str, 0, ldata)
if myid == main_node: print_end_msg("helical-shiftali_MPI")
print "head did bcast" ##### scatter #### scat=array([10,20,30],"i") junk=mpi.mpi_scatter(scat,1,mpi.MPI_INT,1,mpi.MPI_INT,mpi.MPI_ROOT,newcom1) ##### send/recv #### for i in range(0,copies): k=(i+1)*100 mpi.mpi_send(k,1,mpi.MPI_INT,i,1234,newcom1) back=mpi.mpi_recv(1,mpi.MPI_INT,i,5678,newcom1) print "from ",i,back ##### reduce #### dummy=1000 final=mpi.mpi_reduce(dummy,1,mpi.MPI_INT,mpi.MPI_SUM,mpi.MPI_ROOT,newcom1) sleep(5) print "the final answer is=",final toRun=getcwd()+"/worker" print mpi.mpi_get_processor_name(),"starting",toRun newcom2=mpi.mpi_comm_spawn(toRun,"from_C_",copies,mpi.MPI_INFO_NULL,0,mpi.MPI_COMM_WORLD) errors=mpi.mpi_array_of_errcodes() print "errors=",errors newcom2Size=mpi.mpi_comm_size(newcom2) print "newcom2Size",newcom2Size sleep(15)
def mref_ali2d_MPI(stack,
refim,
outdir,
maskfile=None,
ir=1,
ou=-1,
rs=1,
xrng=0,
yrng=0,
step=1,
center=1,
maxit=10,
CTF=False,
snr=1.0,
user_func_name="ref_ali2d",
rand_seed=1000):
# 2D multi-reference alignment using rotational ccf in polar coordinates and quadratic interpolation
from sp_utilities import model_circle, combine_params2, inverse_transform2, drop_image, get_image, get_im
from sp_utilities import reduce_EMData_to_root, bcast_EMData_to_all, bcast_number_to_all
from sp_utilities import send_attr_dict
from sp_utilities import center_2D
from sp_statistics import fsc_mask
from sp_alignment import Numrinit, ringwe, search_range
from sp_fundamentals import rot_shift2D, fshift
from sp_utilities import get_params2D, set_params2D
from random import seed, randint
from sp_morphology import ctf_2
from sp_filter import filt_btwl, filt_params
from numpy import reshape, shape
from sp_utilities import print_msg, print_begin_msg, print_end_msg
import os
import sys
import shutil
from sp_applications import MPI_start_end
from mpi import mpi_comm_size, mpi_comm_rank, MPI_COMM_WORLD
from mpi import mpi_reduce, mpi_bcast, mpi_barrier, mpi_recv, mpi_send
from mpi import MPI_SUM, MPI_FLOAT, MPI_INT
number_of_proc = mpi_comm_size(MPI_COMM_WORLD)
myid = mpi_comm_rank(MPI_COMM_WORLD)
main_node = 0
# create the output directory, if it does not exist
if os.path.exists(outdir):
ERROR(
'Output directory exists, please change the name and restart the program',
"mref_ali2d_MPI ", 1, myid)
mpi_barrier(MPI_COMM_WORLD)
import sp_global_def
if myid == main_node:
os.mkdir(outdir)
sp_global_def.LOGFILE = os.path.join(outdir, sp_global_def.LOGFILE)
print_begin_msg("mref_ali2d_MPI")
nima = EMUtil.get_image_count(stack)
image_start, image_end = MPI_start_end(nima, number_of_proc, myid)
nima = EMUtil.get_image_count(stack)
ima = EMData()
ima.read_image(stack, image_start)
first_ring = int(ir)
last_ring = int(ou)
rstep = int(rs)
max_iter = int(maxit)
if max_iter == 0:
max_iter = 10
auto_stop = True
else:
auto_stop = False
if myid == main_node:
print_msg("Input stack : %s\n" % (stack))
print_msg("Reference stack : %s\n" % (refim))
print_msg("Output directory : %s\n" % (outdir))
print_msg("Maskfile : %s\n" % (maskfile))
print_msg("Inner radius : %i\n" % (first_ring))
nx = ima.get_xsize()
# default value for the last ring
if last_ring == -1: last_ring = nx / 2 - 2
if myid == main_node:
print_msg("Outer radius : %i\n" % (last_ring))
print_msg("Ring step : %i\n" % (rstep))
print_msg("X search range : %f\n" % (xrng))
print_msg("Y search range : %f\n" % (yrng))
print_msg("Translational step : %f\n" % (step))
print_msg("Center type : %i\n" % (center))
print_msg("Maximum iteration : %i\n" % (max_iter))
print_msg("CTF correction : %s\n" % (CTF))
print_msg("Signal-to-Noise Ratio : %f\n" % (snr))
print_msg("Random seed : %i\n\n" % (rand_seed))
print_msg("User function : %s\n" % (user_func_name))
import sp_user_functions
user_func = sp_user_functions.factory[user_func_name]
if maskfile:
import types
if type(maskfile) is bytes: mask = get_image(maskfile)
else: mask = maskfile
else: mask = model_circle(last_ring, nx, nx)
# references, do them on all processors...
refi = []
numref = EMUtil.get_image_count(refim)
# IMAGES ARE SQUARES! center is in SPIDER convention
cnx = nx / 2 + 1
cny = cnx
mode = "F"
#precalculate rings
numr = Numrinit(first_ring, last_ring, rstep, mode)
wr = ringwe(numr, mode)
# prepare reference images on all nodes
ima.to_zero()
for j in range(numref):
# even, odd, numer of even, number of images. After frc, totav
refi.append([get_im(refim, j), ima.copy(), 0])
# for each node read its share of data
data = EMData.read_images(stack, list(range(image_start, image_end)))
for im in range(image_start, image_end):
data[im - image_start].set_attr('ID', im)
if myid == main_node: seed(rand_seed)
a0 = -1.0
again = True
Iter = 0
ref_data = [mask, center, None, None]
while Iter < max_iter and again:
ringref = []
mashi = cnx - last_ring - 2
for j in range(numref):
refi[j][0].process_inplace("normalize.mask", {
"mask": mask,
"no_sigma": 1
}) # normalize reference images to N(0,1)
cimage = Util.Polar2Dm(refi[j][0], cnx, cny, numr, mode)
Util.Frngs(cimage, numr)
Util.Applyws(cimage, numr, wr)
ringref.append(cimage)
# zero refi
refi[j][0].to_zero()
refi[j][1].to_zero()
refi[j][2] = 0
assign = [[] for i in range(numref)]
# begin MPI section
for im in range(image_start, image_end):
alpha, sx, sy, mirror, scale = get_params2D(data[im - image_start])
# Why inverse? 07/11/2015 PAP
alphai, sxi, syi, scalei = inverse_transform2(alpha, sx, sy)
# normalize
data[im - image_start].process_inplace("normalize.mask", {
"mask": mask,
"no_sigma": 0
}) # subtract average under the mask
# If shifts are outside of the permissible range, reset them
if (abs(sxi) > mashi or abs(syi) > mashi):
sxi = 0.0
syi = 0.0
set_params2D(data[im - image_start], [0.0, 0.0, 0.0, 0, 1.0])
ny = nx
txrng = search_range(nx, last_ring, sxi, xrng, "mref_ali2d_MPI")
txrng = [txrng[1], txrng[0]]
tyrng = search_range(ny, last_ring, syi, yrng, "mref_ali2d_MPI")
tyrng = [tyrng[1], tyrng[0]]
# align current image to the reference
[angt, sxst, syst, mirrort, xiref,
peakt] = Util.multiref_polar_ali_2d(data[im - image_start],
ringref, txrng, tyrng, step,
mode, numr, cnx + sxi,
cny + syi)
iref = int(xiref)
# combine parameters and set them to the header, ignore previous angle and mirror
[alphan, sxn, syn,
mn] = combine_params2(0.0, -sxi, -syi, 0, angt, sxst, syst,
(int)(mirrort))
set_params2D(data[im - image_start],
[alphan, sxn, syn, int(mn), scale])
data[im - image_start].set_attr('assign', iref)
# apply current parameters and add to the average
temp = rot_shift2D(data[im - image_start], alphan, sxn, syn, mn)
it = im % 2
Util.add_img(refi[iref][it], temp)
assign[iref].append(im)
#assign[im] = iref
refi[iref][2] += 1.0
del ringref
# end MPI section, bring partial things together, calculate new reference images, broadcast them back
for j in range(numref):
reduce_EMData_to_root(refi[j][0], myid, main_node)
reduce_EMData_to_root(refi[j][1], myid, main_node)
refi[j][2] = mpi_reduce(refi[j][2], 1, MPI_FLOAT, MPI_SUM,
main_node, MPI_COMM_WORLD)
if (myid == main_node): refi[j][2] = int(refi[j][2][0])
# gather assignements
for j in range(numref):
if myid == main_node:
for n in range(number_of_proc):
if n != main_node:
import sp_global_def
ln = mpi_recv(1, MPI_INT, n,
sp_global_def.SPARX_MPI_TAG_UNIVERSAL,
MPI_COMM_WORLD)
lis = mpi_recv(ln[0], MPI_INT, n,
sp_global_def.SPARX_MPI_TAG_UNIVERSAL,
MPI_COMM_WORLD)
for l in range(ln[0]):
assign[j].append(int(lis[l]))
else:
import sp_global_def
mpi_send(len(assign[j]), 1, MPI_INT, main_node,
sp_global_def.SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
mpi_send(assign[j], len(assign[j]), MPI_INT, main_node,
sp_global_def.SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
if myid == main_node:
# replace the name of the stack with reference with the current one
refim = os.path.join(outdir, "aqm%03d.hdf" % Iter)
a1 = 0.0
ave_fsc = []
for j in range(numref):
if refi[j][2] < 4:
#ERROR("One of the references vanished","mref_ali2d_MPI",1)
# if vanished, put a random image (only from main node!) there
assign[j] = []
assign[j].append(
randint(image_start, image_end - 1) - image_start)
refi[j][0] = data[assign[j][0]].copy()
#print 'ERROR', j
else:
#frsc = fsc_mask(refi[j][0], refi[j][1], mask, 1.0, os.path.join(outdir,"drm%03d%04d"%(Iter, j)))
from sp_statistics import fsc
frsc = fsc(
refi[j][0], refi[j][1], 1.0,
os.path.join(outdir, "drm%03d%04d.txt" % (Iter, j)))
Util.add_img(refi[j][0], refi[j][1])
Util.mul_scalar(refi[j][0], 1.0 / float(refi[j][2]))
if ave_fsc == []:
for i in range(len(frsc[1])):
ave_fsc.append(frsc[1][i])
c_fsc = 1
else:
for i in range(len(frsc[1])):
ave_fsc[i] += frsc[1][i]
c_fsc += 1
#print 'OK', j, len(frsc[1]), frsc[1][0:5], ave_fsc[0:5]
#print 'sum', sum(ave_fsc)
if sum(ave_fsc) != 0:
for i in range(len(ave_fsc)):
ave_fsc[i] /= float(c_fsc)
frsc[1][i] = ave_fsc[i]
for j in range(numref):
ref_data[2] = refi[j][0]
ref_data[3] = frsc
refi[j][0], cs = user_func(ref_data)
# write the current average
TMP = []
for i_tmp in range(len(assign[j])):
TMP.append(float(assign[j][i_tmp]))
TMP.sort()
refi[j][0].set_attr_dict({'ave_n': refi[j][2], 'members': TMP})
del TMP
refi[j][0].process_inplace("normalize.mask", {
"mask": mask,
"no_sigma": 1
})
refi[j][0].write_image(refim, j)
Iter += 1
msg = "ITERATION #%3d %d\n\n" % (Iter, again)
print_msg(msg)
for j in range(numref):
msg = " group #%3d number of particles = %7d\n" % (
j, refi[j][2])
print_msg(msg)
Iter = bcast_number_to_all(Iter, main_node) # need to tell all
if again:
for j in range(numref):
bcast_EMData_to_all(refi[j][0], myid, main_node)
# clean up
del assign
# write out headers and STOP, under MPI writing has to be done sequentially (time-consumming)
mpi_barrier(MPI_COMM_WORLD)
if CTF and data_had_ctf == 0:
for im in range(len(data)):
data[im].set_attr('ctf_applied', 0)
par_str = ['xform.align2d', 'assign', 'ID']
if myid == main_node:
from sp_utilities import file_type
if (file_type(stack) == "bdb"):
from sp_utilities import recv_attr_dict_bdb
recv_attr_dict_bdb(main_node, stack, data, par_str, image_start,
image_end, number_of_proc)
else:
from sp_utilities import recv_attr_dict
recv_attr_dict(main_node, stack, data, par_str, image_start,
image_end, number_of_proc)
else:
send_attr_dict(main_node, data, par_str, image_start, image_end)
if myid == main_node:
print_end_msg("mref_ali2d_MPI")
def main():
def params_3D_2D_NEW(phi, theta, psi, s2x, s2y, mirror):
if mirror:
m = 1
alpha, sx, sy, scalen = compose_transform2(0, s2x, s2y, 1.0, 540.0-psi, 0, 0, 1.0)
else:
m = 0
alpha, sx, sy, scalen = compose_transform2(0, s2x, s2y, 1.0, 360.0-psi, 0, 0, 1.0)
return alpha, sx, sy, m
progname = os.path.basename(sys.argv[0])
usage = progname + " prj_stack --ave2D= --var2D= --ave3D= --var3D= --img_per_grp= --fl=0.2 --aa=0.1 --sym=symmetry --CTF"
parser = OptionParser(usage, version=SPARXVERSION)
parser.add_option("--ave2D", type="string" , default=False, help="write to the disk a stack of 2D averages")
parser.add_option("--var2D", type="string" , default=False, help="write to the disk a stack of 2D variances")
parser.add_option("--ave3D", type="string" , default=False, help="write to the disk reconstructed 3D average")
parser.add_option("--var3D", type="string" , default=False, help="compute 3D variability (time consuming!)")
parser.add_option("--img_per_grp", type="int" , default=10 , help="number of neighbouring projections")
parser.add_option("--no_norm", action="store_true", default=False, help="do not use normalization")
parser.add_option("--radiusvar", type="int" , default=-1 , help="radius for 3D var" )
parser.add_option("--npad", type="int" , default=2 , help="number of time to pad the original images")
parser.add_option("--sym" , type="string" , default="c1" , help="symmetry")
parser.add_option("--fl", type="float" , default=0.0 , help="stop-band frequency (Default - no filtration)")
parser.add_option("--aa", type="float" , default=0.0 , help="fall off of the filter (Default - no filtration)")
parser.add_option("--CTF", action="store_true", default=False, help="use CFT correction")
parser.add_option("--VERBOSE", action="store_true", default=False, help="Long output for debugging")
#parser.add_option("--MPI" , action="store_true", default=False, help="use MPI version")
#parser.add_option("--radiuspca", type="int" , default=-1 , help="radius for PCA" )
#parser.add_option("--iter", type="int" , default=40 , help="maximum number of iterations (stop criterion of reconstruction process)" )
#parser.add_option("--abs", type="float" , default=0.0 , help="minimum average absolute change of voxels' values (stop criterion of reconstruction process)" )
#parser.add_option("--squ", type="float" , default=0.0 , help="minimum average squared change of voxels' values (stop criterion of reconstruction process)" )
parser.add_option("--VAR" , action="store_true", default=False, help="stack on input consists of 2D variances (Default False)")
parser.add_option("--decimate", type="float", default=1.0, help="image decimate rate, a number large than 1. default is 1")
parser.add_option("--window", type="int", default=0, help="reduce images to a small image size without changing pixel_size. Default value is zero.")
#parser.add_option("--SND", action="store_true", default=False, help="compute squared normalized differences (Default False)")
parser.add_option("--nvec", type="int" , default=0 , help="number of eigenvectors, default = 0 meaning no PCA calculated")
parser.add_option("--symmetrize", action="store_true", default=False, help="Prepare input stack for handling symmetry (Default False)")
(options,args) = parser.parse_args()
#####
from mpi import mpi_init, mpi_comm_rank, mpi_comm_size, mpi_recv, MPI_COMM_WORLD, MPI_TAG_UB
from mpi import mpi_barrier, mpi_reduce, mpi_bcast, mpi_send, MPI_FLOAT, MPI_SUM, MPI_INT, MPI_MAX
from applications import MPI_start_end
from reconstruction import recons3d_em, recons3d_em_MPI
from reconstruction import recons3d_4nn_MPI, recons3d_4nn_ctf_MPI
from utilities import print_begin_msg, print_end_msg, print_msg
from utilities import read_text_row, get_image, get_im
from utilities import bcast_EMData_to_all, bcast_number_to_all
from utilities import get_symt
# This is code for handling symmetries by the above program. To be incorporated. PAP 01/27/2015
from EMAN2db import db_open_dict
if options.symmetrize :
try:
sys.argv = mpi_init(len(sys.argv), sys.argv)
try:
number_of_proc = mpi_comm_size(MPI_COMM_WORLD)
if( number_of_proc > 1 ):
ERROR("Cannot use more than one CPU for symmetry prepration","sx3dvariability",1)
except:
pass
except:
pass
# Input
#instack = "Clean_NORM_CTF_start_wparams.hdf"
#instack = "bdb:data"
instack = args[0]
sym = options.sym
if( sym == "c1" ):
ERROR("Thre is no need to symmetrize stack for C1 symmetry","sx3dvariability",1)
if(instack[:4] !="bdb:"):
stack = "bdb:data"
delete_bdb(stack)
cmdexecute("sxcpy.py "+instack+" "+stack)
else:
stack = instack
qt = EMUtil.get_all_attributes(stack,'xform.projection')
na = len(qt)
ts = get_symt(sym)
ks = len(ts)
angsa = [None]*na
for k in xrange(ks):
delete_bdb("bdb:Q%1d"%k)
cmdexecute("e2bdb.py "+stack+" --makevstack=bdb:Q%1d"%k)
DB = db_open_dict("bdb:Q%1d"%k)
for i in xrange(na):
ut = qt[i]*ts[k]
DB.set_attr(i, "xform.projection", ut)
#bt = ut.get_params("spider")
#angsa[i] = [round(bt["phi"],3)%360.0, round(bt["theta"],3)%360.0, bt["psi"], -bt["tx"], -bt["ty"]]
#write_text_row(angsa, 'ptsma%1d.txt'%k)
#cmdexecute("e2bdb.py "+stack+" --makevstack=bdb:Q%1d"%k)
#cmdexecute("sxheader.py bdb:Q%1d --params=xform.projection --import=ptsma%1d.txt"%(k,k))
DB.close()
delete_bdb("bdb:sdata")
cmdexecute("e2bdb.py . --makevstack=bdb:sdata --filt=Q")
#cmdexecute("ls EMAN2DB/sdata*")
a = get_im("bdb:sdata")
a.set_attr("variabilitysymmetry",sym)
a.write_image("bdb:sdata")
else:
sys.argv = mpi_init(len(sys.argv), sys.argv)
myid = mpi_comm_rank(MPI_COMM_WORLD)
number_of_proc = mpi_comm_size(MPI_COMM_WORLD)
main_node = 0
if len(args) == 1:
stack = args[0]
else:
print( "usage: " + usage)
print( "Please run '" + progname + " -h' for detailed options")
return 1
t0 = time()
# obsolete flags
options.MPI = True
options.nvec = 0
options.radiuspca = -1
options.iter = 40
options.abs = 0.0
options.squ = 0.0
if options.fl > 0.0 and options.aa == 0.0:
ERROR("Fall off has to be given for the low-pass filter", "sx3dvariability", 1, myid)
if options.VAR and options.SND:
ERROR("Only one of var and SND can be set!", "sx3dvariability", myid)
exit()
if options.VAR and (options.ave2D or options.ave3D or options.var2D):
ERROR("When VAR is set, the program cannot output ave2D, ave3D or var2D", "sx3dvariability", 1, myid)
exit()
#if options.SND and (options.ave2D or options.ave3D):
# ERROR("When SND is set, the program cannot output ave2D or ave3D", "sx3dvariability", 1, myid)
# exit()
if options.nvec > 0 :
ERROR("PCA option not implemented", "sx3dvariability", 1, myid)
exit()
if options.nvec > 0 and options.ave3D == None:
ERROR("When doing PCA analysis, one must set ave3D", "sx3dvariability", myid=myid)
exit()
import string
options.sym = options.sym.lower()
if global_def.CACHE_DISABLE:
from utilities import disable_bdb_cache
disable_bdb_cache()
global_def.BATCH = True
if myid == main_node:
print_begin_msg("sx3dvariability")
print_msg("%-70s: %s\n"%("Input stack", stack))
img_per_grp = options.img_per_grp
nvec = options.nvec
radiuspca = options.radiuspca
symbaselen = 0
if myid == main_node:
nima = EMUtil.get_image_count(stack)
img = get_image(stack)
nx = img.get_xsize()
ny = img.get_ysize()
if options.sym != "c1" :
imgdata = get_im(stack)
try:
i = imgdata.get_attr("variabilitysymmetry")
if(i != options.sym):
ERROR("The symmetry provided does not agree with the symmetry of the input stack", "sx3dvariability", myid=myid)
except:
ERROR("Input stack is not prepared for symmetry, please follow instructions", "sx3dvariability", myid=myid)
from utilities import get_symt
i = len(get_symt(options.sym))
if((nima/i)*i != nima):
ERROR("The length of the input stack is incorrect for symmetry processing", "sx3dvariability", myid=myid)
symbaselen = nima/i
else: symbaselen = nima
else:
nima = 0
nx = 0
ny = 0
nima = bcast_number_to_all(nima)
nx = bcast_number_to_all(nx)
ny = bcast_number_to_all(ny)
Tracker ={}
Tracker["nx"] =nx
Tracker["ny"] =ny
Tracker["total_stack"]=nima
if options.decimate==1.:
if options.window !=0:
nx = options.window
ny = options.window
else:
if options.window ==0:
nx = int(nx/options.decimate)
ny = int(ny/options.decimate)
else:
nx = int(options.window/options.decimate)
ny = nx
symbaselen = bcast_number_to_all(symbaselen)
if radiuspca == -1: radiuspca = nx/2-2
if myid == main_node:
print_msg("%-70s: %d\n"%("Number of projection", nima))
img_begin, img_end = MPI_start_end(nima, number_of_proc, myid)
"""
if options.SND:
from projection import prep_vol, prgs
from statistics import im_diff
from utilities import get_im, model_circle, get_params_proj, set_params_proj
from utilities import get_ctf, generate_ctf
from filter import filt_ctf
imgdata = EMData.read_images(stack, range(img_begin, img_end))
if options.CTF:
vol = recons3d_4nn_ctf_MPI(myid, imgdata, 1.0, symmetry=options.sym, npad=options.npad, xysize=-1, zsize=-1)
else:
vol = recons3d_4nn_MPI(myid, imgdata, symmetry=options.sym, npad=options.npad, xysize=-1, zsize=-1)
bcast_EMData_to_all(vol, myid)
volft, kb = prep_vol(vol)
mask = model_circle(nx/2-2, nx, ny)
varList = []
for i in xrange(img_begin, img_end):
phi, theta, psi, s2x, s2y = get_params_proj(imgdata[i-img_begin])
ref_prj = prgs(volft, kb, [phi, theta, psi, -s2x, -s2y])
if options.CTF:
ctf_params = get_ctf(imgdata[i-img_begin])
ref_prj = filt_ctf(ref_prj, generate_ctf(ctf_params))
diff, A, B = im_diff(ref_prj, imgdata[i-img_begin], mask)
diff2 = diff*diff
set_params_proj(diff2, [phi, theta, psi, s2x, s2y])
varList.append(diff2)
mpi_barrier(MPI_COMM_WORLD)
"""
if options.VAR:
#varList = EMData.read_images(stack, range(img_begin, img_end))
varList = []
this_image = EMData()
for index_of_particle in xrange(img_begin,img_end):
this_image.read_image(stack,index_of_particle)
varList.append(image_decimate_window_xform_ctf(img,options.decimate,options.window,options.CTF))
else:
from utilities import bcast_number_to_all, bcast_list_to_all, send_EMData, recv_EMData
from utilities import set_params_proj, get_params_proj, params_3D_2D, get_params2D, set_params2D, compose_transform2
from utilities import model_blank, nearest_proj, model_circle
from applications import pca
from statistics import avgvar, avgvar_ctf, ccc
from filter import filt_tanl
from morphology import threshold, square_root
from projection import project, prep_vol, prgs
from sets import Set
if myid == main_node:
t1 = time()
proj_angles = []
aveList = []
tab = EMUtil.get_all_attributes(stack, 'xform.projection')
for i in xrange(nima):
t = tab[i].get_params('spider')
phi = t['phi']
theta = t['theta']
psi = t['psi']
x = theta
if x > 90.0: x = 180.0 - x
x = x*10000+psi
proj_angles.append([x, t['phi'], t['theta'], t['psi'], i])
t2 = time()
print_msg("%-70s: %d\n"%("Number of neighboring projections", img_per_grp))
print_msg("...... Finding neighboring projections\n")
if options.VERBOSE:
print "Number of images per group: ", img_per_grp
print "Now grouping projections"
proj_angles.sort()
proj_angles_list = [0.0]*(nima*4)
if myid == main_node:
for i in xrange(nima):
proj_angles_list[i*4] = proj_angles[i][1]
proj_angles_list[i*4+1] = proj_angles[i][2]
proj_angles_list[i*4+2] = proj_angles[i][3]
proj_angles_list[i*4+3] = proj_angles[i][4]
proj_angles_list = bcast_list_to_all(proj_angles_list, myid, main_node)
proj_angles = []
for i in xrange(nima):
proj_angles.append([proj_angles_list[i*4], proj_angles_list[i*4+1], proj_angles_list[i*4+2], int(proj_angles_list[i*4+3])])
del proj_angles_list
proj_list, mirror_list = nearest_proj(proj_angles, img_per_grp, range(img_begin, img_end))
all_proj = Set()
for im in proj_list:
for jm in im:
all_proj.add(proj_angles[jm][3])
all_proj = list(all_proj)
if options.VERBOSE:
print "On node %2d, number of images needed to be read = %5d"%(myid, len(all_proj))
index = {}
for i in xrange(len(all_proj)): index[all_proj[i]] = i
mpi_barrier(MPI_COMM_WORLD)
if myid == main_node:
print_msg("%-70s: %.2f\n"%("Finding neighboring projections lasted [s]", time()-t2))
print_msg("%-70s: %d\n"%("Number of groups processed on the main node", len(proj_list)))
if options.VERBOSE:
print "Grouping projections took: ", (time()-t2)/60 , "[min]"
print "Number of groups on main node: ", len(proj_list)
mpi_barrier(MPI_COMM_WORLD)
if myid == main_node:
print_msg("...... calculating the stack of 2D variances \n")
if options.VERBOSE:
print "Now calculating the stack of 2D variances"
proj_params = [0.0]*(nima*5)
aveList = []
varList = []
if nvec > 0:
eigList = [[] for i in xrange(nvec)]
if options.VERBOSE: print "Begin to read images on processor %d"%(myid)
ttt = time()
#imgdata = EMData.read_images(stack, all_proj)
img = EMData()
imgdata = []
for index_of_proj in xrange(len(all_proj)):
img.read_image(stack, all_proj[index_of_proj])
dmg = image_decimate_window_xform_ctf(img,options.decimate,options.window,options.CTF)
#print dmg.get_xsize(), "init"
imgdata.append(dmg)
if options.VERBOSE:
print "Reading images on processor %d done, time = %.2f"%(myid, time()-ttt)
print "On processor %d, we got %d images"%(myid, len(imgdata))
mpi_barrier(MPI_COMM_WORLD)
'''
imgdata2 = EMData.read_images(stack, range(img_begin, img_end))
if options.fl > 0.0:
for k in xrange(len(imgdata2)):
imgdata2[k] = filt_tanl(imgdata2[k], options.fl, options.aa)
if options.CTF:
vol = recons3d_4nn_ctf_MPI(myid, imgdata2, 1.0, symmetry=options.sym, npad=options.npad, xysize=-1, zsize=-1)
else:
vol = recons3d_4nn_MPI(myid, imgdata2, symmetry=options.sym, npad=options.npad, xysize=-1, zsize=-1)
if myid == main_node:
vol.write_image("vol_ctf.hdf")
print_msg("Writing to the disk volume reconstructed from averages as : %s\n"%("vol_ctf.hdf"))
del vol, imgdata2
mpi_barrier(MPI_COMM_WORLD)
'''
from applications import prepare_2d_forPCA
from utilities import model_blank
for i in xrange(len(proj_list)):
ki = proj_angles[proj_list[i][0]][3]
if ki >= symbaselen: continue
mi = index[ki]
phiM, thetaM, psiM, s2xM, s2yM = get_params_proj(imgdata[mi])
grp_imgdata = []
for j in xrange(img_per_grp):
mj = index[proj_angles[proj_list[i][j]][3]]
phi, theta, psi, s2x, s2y = get_params_proj(imgdata[mj])
alpha, sx, sy, mirror = params_3D_2D_NEW(phi, theta, psi, s2x, s2y, mirror_list[i][j])
if thetaM <= 90:
if mirror == 0: alpha, sx, sy, scale = compose_transform2(alpha, sx, sy, 1.0, phiM-phi, 0.0, 0.0, 1.0)
else: alpha, sx, sy, scale = compose_transform2(alpha, sx, sy, 1.0, 180-(phiM-phi), 0.0, 0.0, 1.0)
else:
if mirror == 0: alpha, sx, sy, scale = compose_transform2(alpha, sx, sy, 1.0, -(phiM-phi), 0.0, 0.0, 1.0)
else: alpha, sx, sy, scale = compose_transform2(alpha, sx, sy, 1.0, -(180-(phiM-phi)), 0.0, 0.0, 1.0)
set_params2D(imgdata[mj], [alpha, sx, sy, mirror, 1.0])
grp_imgdata.append(imgdata[mj])
#print grp_imgdata[j].get_xsize(), imgdata[mj].get_xsize()
if not options.no_norm:
#print grp_imgdata[j].get_xsize()
mask = model_circle(nx/2-2, nx, nx)
for k in xrange(img_per_grp):
ave, std, minn, maxx = Util.infomask(grp_imgdata[k], mask, False)
grp_imgdata[k] -= ave
grp_imgdata[k] /= std
del mask
if options.fl > 0.0:
from filter import filt_ctf, filt_table
from fundamentals import fft, window2d
nx2 = 2*nx
ny2 = 2*ny
if options.CTF:
from utilities import pad
for k in xrange(img_per_grp):
grp_imgdata[k] = window2d(fft( filt_tanl( filt_ctf(fft(pad(grp_imgdata[k], nx2, ny2, 1,0.0)), grp_imgdata[k].get_attr("ctf"), binary=1), options.fl, options.aa) ),nx,ny)
#grp_imgdata[k] = window2d(fft( filt_table( filt_tanl( filt_ctf(fft(pad(grp_imgdata[k], nx2, ny2, 1,0.0)), grp_imgdata[k].get_attr("ctf"), binary=1), options.fl, options.aa), fifi) ),nx,ny)
#grp_imgdata[k] = filt_tanl(grp_imgdata[k], options.fl, options.aa)
else:
for k in xrange(img_per_grp):
grp_imgdata[k] = filt_tanl( grp_imgdata[k], options.fl, options.aa)
#grp_imgdata[k] = window2d(fft( filt_table( filt_tanl( filt_ctf(fft(pad(grp_imgdata[k], nx2, ny2, 1,0.0)), grp_imgdata[k].get_attr("ctf"), binary=1), options.fl, options.aa), fifi) ),nx,ny)
#grp_imgdata[k] = filt_tanl(grp_imgdata[k], options.fl, options.aa)
else:
from utilities import pad, read_text_file
from filter import filt_ctf, filt_table
from fundamentals import fft, window2d
nx2 = 2*nx
ny2 = 2*ny
if options.CTF:
from utilities import pad
for k in xrange(img_per_grp):
grp_imgdata[k] = window2d( fft( filt_ctf(fft(pad(grp_imgdata[k], nx2, ny2, 1,0.0)), grp_imgdata[k].get_attr("ctf"), binary=1) ) , nx,ny)
#grp_imgdata[k] = window2d(fft( filt_table( filt_tanl( filt_ctf(fft(pad(grp_imgdata[k], nx2, ny2, 1,0.0)), grp_imgdata[k].get_attr("ctf"), binary=1), options.fl, options.aa), fifi) ),nx,ny)
#grp_imgdata[k] = filt_tanl(grp_imgdata[k], options.fl, options.aa)
'''
if i < 10 and myid == main_node:
for k in xrange(10):
grp_imgdata[k].write_image("grp%03d.hdf"%i, k)
'''
"""
if myid == main_node and i==0:
for pp in xrange(len(grp_imgdata)):
grp_imgdata[pp].write_image("pp.hdf", pp)
"""
ave, grp_imgdata = prepare_2d_forPCA(grp_imgdata)
"""
if myid == main_node and i==0:
for pp in xrange(len(grp_imgdata)):
grp_imgdata[pp].write_image("qq.hdf", pp)
"""
var = model_blank(nx,ny)
for q in grp_imgdata: Util.add_img2( var, q )
Util.mul_scalar( var, 1.0/(len(grp_imgdata)-1))
# Switch to std dev
var = square_root(threshold(var))
#if options.CTF: ave, var = avgvar_ctf(grp_imgdata, mode="a")
#else: ave, var = avgvar(grp_imgdata, mode="a")
"""
if myid == main_node:
ave.write_image("avgv.hdf",i)
var.write_image("varv.hdf",i)
"""
set_params_proj(ave, [phiM, thetaM, 0.0, 0.0, 0.0])
set_params_proj(var, [phiM, thetaM, 0.0, 0.0, 0.0])
aveList.append(ave)
varList.append(var)
if options.VERBOSE:
print "%5.2f%% done on processor %d"%(i*100.0/len(proj_list), myid)
if nvec > 0:
eig = pca(input_stacks=grp_imgdata, subavg="", mask_radius=radiuspca, nvec=nvec, incore=True, shuffle=False, genbuf=True)
for k in xrange(nvec):
set_params_proj(eig[k], [phiM, thetaM, 0.0, 0.0, 0.0])
eigList[k].append(eig[k])
"""
if myid == 0 and i == 0:
for k in xrange(nvec):
eig[k].write_image("eig.hdf", k)
"""
del imgdata
# To this point, all averages, variances, and eigenvectors are computed
if options.ave2D:
from fundamentals import fpol
if myid == main_node:
km = 0
for i in xrange(number_of_proc):
if i == main_node :
for im in xrange(len(aveList)):
aveList[im].write_image(options.ave2D, km)
km += 1
else:
nl = mpi_recv(1, MPI_INT, i, MPI_TAG_UB, MPI_COMM_WORLD)
nl = int(nl[0])
for im in xrange(nl):
ave = recv_EMData(i, im+i+70000)
"""
nm = mpi_recv(1, MPI_INT, i, MPI_TAG_UB, MPI_COMM_WORLD)
nm = int(nm[0])
members = mpi_recv(nm, MPI_INT, i, MPI_TAG_UB, MPI_COMM_WORLD)
ave.set_attr('members', map(int, members))
members = mpi_recv(nm, MPI_FLOAT, i, MPI_TAG_UB, MPI_COMM_WORLD)
ave.set_attr('pix_err', map(float, members))
members = mpi_recv(3, MPI_FLOAT, i, MPI_TAG_UB, MPI_COMM_WORLD)
ave.set_attr('refprojdir', map(float, members))
"""
tmpvol=fpol(ave, Tracker["nx"],Tracker["nx"],Tracker["nx"])
tmpvol.write_image(options.ave2D, km)
km += 1
else:
mpi_send(len(aveList), 1, MPI_INT, main_node, MPI_TAG_UB, MPI_COMM_WORLD)
for im in xrange(len(aveList)):
send_EMData(aveList[im], main_node,im+myid+70000)
"""
members = aveList[im].get_attr('members')
mpi_send(len(members), 1, MPI_INT, main_node, MPI_TAG_UB, MPI_COMM_WORLD)
mpi_send(members, len(members), MPI_INT, main_node, MPI_TAG_UB, MPI_COMM_WORLD)
members = aveList[im].get_attr('pix_err')
mpi_send(members, len(members), MPI_FLOAT, main_node, MPI_TAG_UB, MPI_COMM_WORLD)
try:
members = aveList[im].get_attr('refprojdir')
mpi_send(members, 3, MPI_FLOAT, main_node, MPI_TAG_UB, MPI_COMM_WORLD)
except:
mpi_send([-999.0,-999.0,-999.0], 3, MPI_FLOAT, main_node, MPI_TAG_UB, MPI_COMM_WORLD)
"""
if options.ave3D:
from fundamentals import fpol
if options.VERBOSE:
print "Reconstructing 3D average volume"
ave3D = recons3d_4nn_MPI(myid, aveList, symmetry=options.sym, npad=options.npad)
bcast_EMData_to_all(ave3D, myid)
if myid == main_node:
ave3D=fpol(ave3D,Tracker["nx"],Tracker["nx"],Tracker["nx"])
ave3D.write_image(options.ave3D)
print_msg("%-70s: %s\n"%("Writing to the disk volume reconstructed from averages as", options.ave3D))
del ave, var, proj_list, stack, phi, theta, psi, s2x, s2y, alpha, sx, sy, mirror, aveList
if nvec > 0:
for k in xrange(nvec):
if options.VERBOSE:
print "Reconstruction eigenvolumes", k
cont = True
ITER = 0
mask2d = model_circle(radiuspca, nx, nx)
while cont:
#print "On node %d, iteration %d"%(myid, ITER)
eig3D = recons3d_4nn_MPI(myid, eigList[k], symmetry=options.sym, npad=options.npad)
bcast_EMData_to_all(eig3D, myid, main_node)
if options.fl > 0.0:
eig3D = filt_tanl(eig3D, options.fl, options.aa)
if myid == main_node:
eig3D.write_image("eig3d_%03d.hdf"%k, ITER)
Util.mul_img( eig3D, model_circle(radiuspca, nx, nx, nx) )
eig3Df, kb = prep_vol(eig3D)
del eig3D
cont = False
icont = 0
for l in xrange(len(eigList[k])):
phi, theta, psi, s2x, s2y = get_params_proj(eigList[k][l])
proj = prgs(eig3Df, kb, [phi, theta, psi, s2x, s2y])
cl = ccc(proj, eigList[k][l], mask2d)
if cl < 0.0:
icont += 1
cont = True
eigList[k][l] *= -1.0
u = int(cont)
u = mpi_reduce([u], 1, MPI_INT, MPI_MAX, main_node, MPI_COMM_WORLD)
icont = mpi_reduce([icont], 1, MPI_INT, MPI_SUM, main_node, MPI_COMM_WORLD)
if myid == main_node:
u = int(u[0])
print " Eigenvector: ",k," number changed ",int(icont[0])
else: u = 0
u = bcast_number_to_all(u, main_node)
cont = bool(u)
ITER += 1
del eig3Df, kb
mpi_barrier(MPI_COMM_WORLD)
del eigList, mask2d
if options.ave3D: del ave3D
if options.var2D:
from fundamentals import fpol
if myid == main_node:
km = 0
for i in xrange(number_of_proc):
if i == main_node :
for im in xrange(len(varList)):
tmpvol=fpol(varList[im], Tracker["nx"], Tracker["nx"],1)
tmpvol.write_image(options.var2D, km)
km += 1
else:
nl = mpi_recv(1, MPI_INT, i, MPI_TAG_UB, MPI_COMM_WORLD)
nl = int(nl[0])
for im in xrange(nl):
ave = recv_EMData(i, im+i+70000)
tmpvol=fpol(ave, Tracker["nx"], Tracker["nx"],1)
tmpvol.write_image(options.var2D, km)
km += 1
else:
mpi_send(len(varList), 1, MPI_INT, main_node, MPI_TAG_UB, MPI_COMM_WORLD)
for im in xrange(len(varList)):
send_EMData(varList[im], main_node, im+myid+70000)# What with the attributes??
mpi_barrier(MPI_COMM_WORLD)
if options.var3D:
if myid == main_node and options.VERBOSE:
print "Reconstructing 3D variability volume"
t6 = time()
radiusvar = options.radiusvar
if( radiusvar < 0 ): radiusvar = nx//2 -3
res = recons3d_4nn_MPI(myid, varList, symmetry=options.sym, npad=options.npad)
#res = recons3d_em_MPI(varList, vol_stack, options.iter, radiusvar, options.abs, True, options.sym, options.squ)
if myid == main_node:
from fundamentals import fpol
res =fpol(res, Tracker["nx"], Tracker["nx"], Tracker["nx"])
res.write_image(options.var3D)
if myid == main_node:
print_msg("%-70s: %.2f\n"%("Reconstructing 3D variability took [s]", time()-t6))
if options.VERBOSE:
print "Reconstruction took: %.2f [min]"%((time()-t6)/60)
if myid == main_node:
print_msg("%-70s: %.2f\n"%("Total time for these computations [s]", time()-t0))
if options.VERBOSE:
print "Total time for these computations: %.2f [min]"%((time()-t0)/60)
print_end_msg("sx3dvariability")
global_def.BATCH = False
from mpi import mpi_finalize
mpi_finalize()
def main():
def params_3D_2D_NEW(phi, theta, psi, s2x, s2y, mirror):
if mirror:
m = 1
alpha, sx, sy, scalen = compose_transform2(0, s2x, s2y, 1.0,
540.0 - psi, 0, 0, 1.0)
else:
m = 0
alpha, sx, sy, scalen = compose_transform2(0, s2x, s2y, 1.0,
360.0 - psi, 0, 0, 1.0)
return alpha, sx, sy, m
progname = os.path.basename(sys.argv[0])
usage = progname + " prj_stack --ave2D= --var2D= --ave3D= --var3D= --img_per_grp= --fl=15. --aa=0.01 --sym=symmetry --CTF"
parser = OptionParser(usage, version=SPARXVERSION)
parser.add_option("--output_dir",
type="string",
default="./",
help="output directory")
parser.add_option("--ave2D",
type="string",
default=False,
help="write to the disk a stack of 2D averages")
parser.add_option("--var2D",
type="string",
default=False,
help="write to the disk a stack of 2D variances")
parser.add_option("--ave3D",
type="string",
default=False,
help="write to the disk reconstructed 3D average")
parser.add_option("--var3D",
type="string",
default=False,
help="compute 3D variability (time consuming!)")
parser.add_option("--img_per_grp",
type="int",
default=10,
help="number of neighbouring projections")
parser.add_option("--no_norm",
action="store_true",
default=False,
help="do not use normalization")
#parser.add_option("--radius", type="int" , default=-1 , help="radius for 3D variability" )
parser.add_option("--npad",
type="int",
default=2,
help="number of time to pad the original images")
parser.add_option("--sym", type="string", default="c1", help="symmetry")
parser.add_option(
"--fl",
type="float",
default=0.0,
help=
"cutoff freqency in absolute frequency (0.0-0.5). (Default - no filtration)"
)
parser.add_option(
"--aa",
type="float",
default=0.0,
help=
"fall off of the filter. Put 0.01 if user has no clue about falloff (Default - no filtration)"
)
parser.add_option("--CTF",
action="store_true",
default=False,
help="use CFT correction")
parser.add_option("--VERBOSE",
action="store_true",
default=False,
help="Long output for debugging")
#parser.add_option("--MPI" , action="store_true", default=False, help="use MPI version")
#parser.add_option("--radiuspca", type="int" , default=-1 , help="radius for PCA" )
#parser.add_option("--iter", type="int" , default=40 , help="maximum number of iterations (stop criterion of reconstruction process)" )
#parser.add_option("--abs", type="float" , default=0.0 , help="minimum average absolute change of voxels' values (stop criterion of reconstruction process)" )
#parser.add_option("--squ", type="float" , default=0.0 , help="minimum average squared change of voxels' values (stop criterion of reconstruction process)" )
parser.add_option(
"--VAR",
action="store_true",
default=False,
help="stack on input consists of 2D variances (Default False)")
parser.add_option(
"--decimate",
type="float",
default=1.0,
help=
"image decimate rate, a number larger (expand image) or less (shrink image) than 1. default is 1"
)
parser.add_option(
"--window",
type="int",
default=0,
help=
"reduce images to a small image size without changing pixel_size. Default value is zero."
)
#parser.add_option("--SND", action="store_true", default=False, help="compute squared normalized differences (Default False)")
parser.add_option(
"--nvec",
type="int",
default=0,
help="number of eigenvectors, default = 0 meaning no PCA calculated")
parser.add_option(
"--symmetrize",
action="store_true",
default=False,
help="Prepare input stack for handling symmetry (Default False)")
(options, args) = parser.parse_args()
#####
from mpi import mpi_init, mpi_comm_rank, mpi_comm_size, mpi_recv, MPI_COMM_WORLD
from mpi import mpi_barrier, mpi_reduce, mpi_bcast, mpi_send, MPI_FLOAT, MPI_SUM, MPI_INT, MPI_MAX
from applications import MPI_start_end
from reconstruction import recons3d_em, recons3d_em_MPI
from reconstruction import recons3d_4nn_MPI, recons3d_4nn_ctf_MPI
from utilities import print_begin_msg, print_end_msg, print_msg
from utilities import read_text_row, get_image, get_im
from utilities import bcast_EMData_to_all, bcast_number_to_all
from utilities import get_symt
# This is code for handling symmetries by the above program. To be incorporated. PAP 01/27/2015
from EMAN2db import db_open_dict
# Set up global variables related to bdb cache
if global_def.CACHE_DISABLE:
from utilities import disable_bdb_cache
disable_bdb_cache()
# Set up global variables related to ERROR function
global_def.BATCH = True
# detect if program is running under MPI
RUNNING_UNDER_MPI = "OMPI_COMM_WORLD_SIZE" in os.environ
if RUNNING_UNDER_MPI:
global_def.MPI = True
if options.symmetrize:
if RUNNING_UNDER_MPI:
try:
sys.argv = mpi_init(len(sys.argv), sys.argv)
try:
number_of_proc = mpi_comm_size(MPI_COMM_WORLD)
if (number_of_proc > 1):
ERROR(
"Cannot use more than one CPU for symmetry prepration",
"sx3dvariability", 1)
except:
pass
except:
pass
if options.output_dir != "./" and not os.path.exists(
options.output_dir):
os.mkdir(options.output_dir)
# Input
#instack = "Clean_NORM_CTF_start_wparams.hdf"
#instack = "bdb:data"
from logger import Logger, BaseLogger_Files
if os.path.exists(os.path.join(options.output_dir, "log.txt")):
os.remove(os.path.join(options.output_dir, "log.txt"))
log_main = Logger(BaseLogger_Files())
log_main.prefix = os.path.join(options.output_dir, "./")
instack = args[0]
sym = options.sym.lower()
if (sym == "c1"):
ERROR("There is no need to symmetrize stack for C1 symmetry",
"sx3dvariability", 1)
line = ""
for a in sys.argv:
line += " " + a
log_main.add(line)
if (instack[:4] != "bdb:"):
if output_dir == "./": stack = "bdb:data"
else: stack = "bdb:" + options.output_dir + "/data"
delete_bdb(stack)
junk = cmdexecute("sxcpy.py " + instack + " " + stack)
else:
stack = instack
qt = EMUtil.get_all_attributes(stack, 'xform.projection')
na = len(qt)
ts = get_symt(sym)
ks = len(ts)
angsa = [None] * na
for k in xrange(ks):
#Qfile = "Q%1d"%k
if options.output_dir != "./":
Qfile = os.path.join(options.output_dir, "Q%1d" % k)
else:
Qfile = os.path.join(options.output_dir, "Q%1d" % k)
#delete_bdb("bdb:Q%1d"%k)
delete_bdb("bdb:" + Qfile)
#junk = cmdexecute("e2bdb.py "+stack+" --makevstack=bdb:Q%1d"%k)
junk = cmdexecute("e2bdb.py " + stack + " --makevstack=bdb:" +
Qfile)
#DB = db_open_dict("bdb:Q%1d"%k)
DB = db_open_dict("bdb:" + Qfile)
for i in xrange(na):
ut = qt[i] * ts[k]
DB.set_attr(i, "xform.projection", ut)
#bt = ut.get_params("spider")
#angsa[i] = [round(bt["phi"],3)%360.0, round(bt["theta"],3)%360.0, bt["psi"], -bt["tx"], -bt["ty"]]
#write_text_row(angsa, 'ptsma%1d.txt'%k)
#junk = cmdexecute("e2bdb.py "+stack+" --makevstack=bdb:Q%1d"%k)
#junk = cmdexecute("sxheader.py bdb:Q%1d --params=xform.projection --import=ptsma%1d.txt"%(k,k))
DB.close()
if options.output_dir == "./": delete_bdb("bdb:sdata")
else: delete_bdb("bdb:" + options.output_dir + "/" + "sdata")
#junk = cmdexecute("e2bdb.py . --makevstack=bdb:sdata --filt=Q")
sdata = "bdb:" + options.output_dir + "/" + "sdata"
print(sdata)
junk = cmdexecute("e2bdb.py " + options.output_dir +
" --makevstack=" + sdata + " --filt=Q")
#junk = cmdexecute("ls EMAN2DB/sdata*")
#a = get_im("bdb:sdata")
a = get_im(sdata)
a.set_attr("variabilitysymmetry", sym)
#a.write_image("bdb:sdata")
a.write_image(sdata)
else:
sys.argv = mpi_init(len(sys.argv), sys.argv)
myid = mpi_comm_rank(MPI_COMM_WORLD)
number_of_proc = mpi_comm_size(MPI_COMM_WORLD)
main_node = 0
if len(args) == 1:
stack = args[0]
else:
print(("usage: " + usage))
print(("Please run '" + progname + " -h' for detailed options"))
return 1
t0 = time()
# obsolete flags
options.MPI = True
options.nvec = 0
options.radiuspca = -1
options.iter = 40
options.abs = 0.0
options.squ = 0.0
if options.fl > 0.0 and options.aa == 0.0:
ERROR("Fall off has to be given for the low-pass filter",
"sx3dvariability", 1, myid)
if options.VAR and options.SND:
ERROR("Only one of var and SND can be set!", "sx3dvariability",
myid)
exit()
if options.VAR and (options.ave2D or options.ave3D or options.var2D):
ERROR(
"When VAR is set, the program cannot output ave2D, ave3D or var2D",
"sx3dvariability", 1, myid)
exit()
#if options.SND and (options.ave2D or options.ave3D):
# ERROR("When SND is set, the program cannot output ave2D or ave3D", "sx3dvariability", 1, myid)
# exit()
if options.nvec > 0:
ERROR("PCA option not implemented", "sx3dvariability", 1, myid)
exit()
if options.nvec > 0 and options.ave3D == None:
ERROR("When doing PCA analysis, one must set ave3D",
"sx3dvariability",
myid=myid)
exit()
import string
options.sym = options.sym.lower()
# if global_def.CACHE_DISABLE:
# from utilities import disable_bdb_cache
# disable_bdb_cache()
# global_def.BATCH = True
if myid == main_node:
if options.output_dir != "./" and not os.path.exists(
options.output_dir):
os.mkdir(options.output_dir)
img_per_grp = options.img_per_grp
nvec = options.nvec
radiuspca = options.radiuspca
from logger import Logger, BaseLogger_Files
#if os.path.exists(os.path.join(options.output_dir, "log.txt")): os.remove(os.path.join(options.output_dir, "log.txt"))
log_main = Logger(BaseLogger_Files())
log_main.prefix = os.path.join(options.output_dir, "./")
if myid == main_node:
line = ""
for a in sys.argv:
line += " " + a
log_main.add(line)
log_main.add("-------->>>Settings given by all options<<<-------")
log_main.add("instack :" + stack)
log_main.add("output_dir :" + options.output_dir)
log_main.add("var3d :" + options.var3D)
if myid == main_node:
line = strftime("%Y-%m-%d_%H:%M:%S", localtime()) + " =>"
#print_begin_msg("sx3dvariability")
msg = "sx3dvariability"
log_main.add(msg)
print(line, msg)
msg = ("%-70s: %s\n" % ("Input stack", stack))
log_main.add(msg)
print(line, msg)
symbaselen = 0
if myid == main_node:
nima = EMUtil.get_image_count(stack)
img = get_image(stack)
nx = img.get_xsize()
ny = img.get_ysize()
if options.sym != "c1":
imgdata = get_im(stack)
try:
i = imgdata.get_attr("variabilitysymmetry").lower()
if (i != options.sym):
ERROR(
"The symmetry provided does not agree with the symmetry of the input stack",
"sx3dvariability",
myid=myid)
except:
ERROR(
"Input stack is not prepared for symmetry, please follow instructions",
"sx3dvariability",
myid=myid)
from utilities import get_symt
i = len(get_symt(options.sym))
if ((nima / i) * i != nima):
ERROR(
"The length of the input stack is incorrect for symmetry processing",
"sx3dvariability",
myid=myid)
symbaselen = nima / i
else:
symbaselen = nima
else:
nima = 0
nx = 0
ny = 0
nima = bcast_number_to_all(nima)
nx = bcast_number_to_all(nx)
ny = bcast_number_to_all(ny)
Tracker = {}
Tracker["total_stack"] = nima
if options.decimate == 1.:
if options.window != 0:
nx = options.window
ny = options.window
else:
if options.window == 0:
nx = int(nx * options.decimate)
ny = int(ny * options.decimate)
else:
nx = int(options.window * options.decimate)
ny = nx
Tracker["nx"] = nx
Tracker["ny"] = ny
Tracker["nz"] = nx
symbaselen = bcast_number_to_all(symbaselen)
if radiuspca == -1: radiuspca = nx / 2 - 2
if myid == main_node:
line = strftime("%Y-%m-%d_%H:%M:%S", localtime()) + " =>"
msg = "%-70s: %d\n" % ("Number of projection", nima)
log_main.add(msg)
print(line, msg)
img_begin, img_end = MPI_start_end(nima, number_of_proc, myid)
"""
if options.SND:
from projection import prep_vol, prgs
from statistics import im_diff
from utilities import get_im, model_circle, get_params_proj, set_params_proj
from utilities import get_ctf, generate_ctf
from filter import filt_ctf
imgdata = EMData.read_images(stack, range(img_begin, img_end))
if options.CTF:
vol = recons3d_4nn_ctf_MPI(myid, imgdata, 1.0, symmetry=options.sym, npad=options.npad, xysize=-1, zsize=-1)
else:
vol = recons3d_4nn_MPI(myid, imgdata, symmetry=options.sym, npad=options.npad, xysize=-1, zsize=-1)
bcast_EMData_to_all(vol, myid)
volft, kb = prep_vol(vol)
mask = model_circle(nx/2-2, nx, ny)
varList = []
for i in xrange(img_begin, img_end):
phi, theta, psi, s2x, s2y = get_params_proj(imgdata[i-img_begin])
ref_prj = prgs(volft, kb, [phi, theta, psi, -s2x, -s2y])
if options.CTF:
ctf_params = get_ctf(imgdata[i-img_begin])
ref_prj = filt_ctf(ref_prj, generate_ctf(ctf_params))
diff, A, B = im_diff(ref_prj, imgdata[i-img_begin], mask)
diff2 = diff*diff
set_params_proj(diff2, [phi, theta, psi, s2x, s2y])
varList.append(diff2)
mpi_barrier(MPI_COMM_WORLD)
"""
if options.VAR:
#varList = EMData.read_images(stack, range(img_begin, img_end))
varList = []
this_image = EMData()
for index_of_particle in xrange(img_begin, img_end):
this_image.read_image(stack, index_of_particle)
varList.append(
image_decimate_window_xform_ctf(this_image,
options.decimate,
options.window,
options.CTF))
else:
from utilities import bcast_number_to_all, bcast_list_to_all, send_EMData, recv_EMData
from utilities import set_params_proj, get_params_proj, params_3D_2D, get_params2D, set_params2D, compose_transform2
from utilities import model_blank, nearest_proj, model_circle
from applications import pca
from statistics import avgvar, avgvar_ctf, ccc
from filter import filt_tanl
from morphology import threshold, square_root
from projection import project, prep_vol, prgs
from sets import Set
if myid == main_node:
t1 = time()
proj_angles = []
aveList = []
tab = EMUtil.get_all_attributes(stack, 'xform.projection')
for i in xrange(nima):
t = tab[i].get_params('spider')
phi = t['phi']
theta = t['theta']
psi = t['psi']
x = theta
if x > 90.0: x = 180.0 - x
x = x * 10000 + psi
proj_angles.append([x, t['phi'], t['theta'], t['psi'], i])
t2 = time()
line = strftime("%Y-%m-%d_%H:%M:%S", localtime()) + " =>"
msg = "%-70s: %d\n" % ("Number of neighboring projections",
img_per_grp)
log_main.add(msg)
print(line, msg)
msg = "...... Finding neighboring projections\n"
log_main.add(msg)
print(line, msg)
if options.VERBOSE:
msg = "Number of images per group: %d" % img_per_grp
log_main.add(msg)
print(line, msg)
msg = "Now grouping projections"
log_main.add(msg)
print(line, msg)
proj_angles.sort()
proj_angles_list = [0.0] * (nima * 4)
if myid == main_node:
for i in xrange(nima):
proj_angles_list[i * 4] = proj_angles[i][1]
proj_angles_list[i * 4 + 1] = proj_angles[i][2]
proj_angles_list[i * 4 + 2] = proj_angles[i][3]
proj_angles_list[i * 4 + 3] = proj_angles[i][4]
proj_angles_list = bcast_list_to_all(proj_angles_list, myid,
main_node)
proj_angles = []
for i in xrange(nima):
proj_angles.append([
proj_angles_list[i * 4], proj_angles_list[i * 4 + 1],
proj_angles_list[i * 4 + 2],
int(proj_angles_list[i * 4 + 3])
])
del proj_angles_list
proj_list, mirror_list = nearest_proj(proj_angles, img_per_grp,
range(img_begin, img_end))
all_proj = Set()
for im in proj_list:
for jm in im:
all_proj.add(proj_angles[jm][3])
all_proj = list(all_proj)
if options.VERBOSE:
print("On node %2d, number of images needed to be read = %5d" %
(myid, len(all_proj)))
index = {}
for i in xrange(len(all_proj)):
index[all_proj[i]] = i
mpi_barrier(MPI_COMM_WORLD)
if myid == main_node:
line = strftime("%Y-%m-%d_%H:%M:%S", localtime()) + " =>"
msg = ("%-70s: %.2f\n" %
("Finding neighboring projections lasted [s]",
time() - t2))
log_main.add(msg)
print(msg)
msg = ("%-70s: %d\n" %
("Number of groups processed on the main node",
len(proj_list)))
log_main.add(msg)
print(line, msg)
if options.VERBOSE:
print("Grouping projections took: ", (time() - t2) / 60,
"[min]")
print("Number of groups on main node: ", len(proj_list))
mpi_barrier(MPI_COMM_WORLD)
if myid == main_node:
line = strftime("%Y-%m-%d_%H:%M:%S", localtime()) + " =>"
msg = ("...... calculating the stack of 2D variances \n")
log_main.add(msg)
print(line, msg)
if options.VERBOSE:
print("Now calculating the stack of 2D variances")
proj_params = [0.0] * (nima * 5)
aveList = []
varList = []
if nvec > 0:
eigList = [[] for i in xrange(nvec)]
if options.VERBOSE:
print("Begin to read images on processor %d" % (myid))
ttt = time()
#imgdata = EMData.read_images(stack, all_proj)
imgdata = []
for index_of_proj in xrange(len(all_proj)):
#img = EMData()
#img.read_image(stack, all_proj[index_of_proj])
dmg = image_decimate_window_xform_ctf(
get_im(stack, all_proj[index_of_proj]), options.decimate,
options.window, options.CTF)
#print dmg.get_xsize(), "init"
imgdata.append(dmg)
if options.VERBOSE:
print("Reading images on processor %d done, time = %.2f" %
(myid, time() - ttt))
print("On processor %d, we got %d images" %
(myid, len(imgdata)))
mpi_barrier(MPI_COMM_WORLD)
'''
imgdata2 = EMData.read_images(stack, range(img_begin, img_end))
if options.fl > 0.0:
for k in xrange(len(imgdata2)):
imgdata2[k] = filt_tanl(imgdata2[k], options.fl, options.aa)
if options.CTF:
vol = recons3d_4nn_ctf_MPI(myid, imgdata2, 1.0, symmetry=options.sym, npad=options.npad, xysize=-1, zsize=-1)
else:
vol = recons3d_4nn_MPI(myid, imgdata2, symmetry=options.sym, npad=options.npad, xysize=-1, zsize=-1)
if myid == main_node:
vol.write_image("vol_ctf.hdf")
print_msg("Writing to the disk volume reconstructed from averages as : %s\n"%("vol_ctf.hdf"))
del vol, imgdata2
mpi_barrier(MPI_COMM_WORLD)
'''
from applications import prepare_2d_forPCA
from utilities import model_blank
for i in xrange(len(proj_list)):
ki = proj_angles[proj_list[i][0]][3]
if ki >= symbaselen: continue
mi = index[ki]
phiM, thetaM, psiM, s2xM, s2yM = get_params_proj(imgdata[mi])
grp_imgdata = []
for j in xrange(img_per_grp):
mj = index[proj_angles[proj_list[i][j]][3]]
phi, theta, psi, s2x, s2y = get_params_proj(imgdata[mj])
alpha, sx, sy, mirror = params_3D_2D_NEW(
phi, theta, psi, s2x, s2y, mirror_list[i][j])
if thetaM <= 90:
if mirror == 0:
alpha, sx, sy, scale = compose_transform2(
alpha, sx, sy, 1.0, phiM - phi, 0.0, 0.0, 1.0)
else:
alpha, sx, sy, scale = compose_transform2(
alpha, sx, sy, 1.0, 180 - (phiM - phi), 0.0,
0.0, 1.0)
else:
if mirror == 0:
alpha, sx, sy, scale = compose_transform2(
alpha, sx, sy, 1.0, -(phiM - phi), 0.0, 0.0,
1.0)
else:
alpha, sx, sy, scale = compose_transform2(
alpha, sx, sy, 1.0, -(180 - (phiM - phi)), 0.0,
0.0, 1.0)
set_params2D(imgdata[mj], [alpha, sx, sy, mirror, 1.0])
grp_imgdata.append(imgdata[mj])
#print grp_imgdata[j].get_xsize(), imgdata[mj].get_xsize()
if not options.no_norm:
#print grp_imgdata[j].get_xsize()
mask = model_circle(nx / 2 - 2, nx, nx)
for k in xrange(img_per_grp):
ave, std, minn, maxx = Util.infomask(
grp_imgdata[k], mask, False)
grp_imgdata[k] -= ave
grp_imgdata[k] /= std
del mask
if options.fl > 0.0:
from filter import filt_ctf, filt_table
from fundamentals import fft, window2d
nx2 = 2 * nx
ny2 = 2 * ny
if options.CTF:
from utilities import pad
for k in xrange(img_per_grp):
grp_imgdata[k] = window2d(
fft(
filt_tanl(
filt_ctf(
fft(
pad(grp_imgdata[k], nx2, ny2,
1, 0.0)),
grp_imgdata[k].get_attr("ctf"),
binary=1), options.fl,
options.aa)), nx, ny)
#grp_imgdata[k] = window2d(fft( filt_table( filt_tanl( filt_ctf(fft(pad(grp_imgdata[k], nx2, ny2, 1,0.0)), grp_imgdata[k].get_attr("ctf"), binary=1), options.fl, options.aa), fifi) ),nx,ny)
#grp_imgdata[k] = filt_tanl(grp_imgdata[k], options.fl, options.aa)
else:
for k in xrange(img_per_grp):
grp_imgdata[k] = filt_tanl(grp_imgdata[k],
options.fl, options.aa)
#grp_imgdata[k] = window2d(fft( filt_table( filt_tanl( filt_ctf(fft(pad(grp_imgdata[k], nx2, ny2, 1,0.0)), grp_imgdata[k].get_attr("ctf"), binary=1), options.fl, options.aa), fifi) ),nx,ny)
#grp_imgdata[k] = filt_tanl(grp_imgdata[k], options.fl, options.aa)
else:
from utilities import pad, read_text_file
from filter import filt_ctf, filt_table
from fundamentals import fft, window2d
nx2 = 2 * nx
ny2 = 2 * ny
if options.CTF:
from utilities import pad
for k in xrange(img_per_grp):
grp_imgdata[k] = window2d(
fft(
filt_ctf(fft(
pad(grp_imgdata[k], nx2, ny2, 1, 0.0)),
grp_imgdata[k].get_attr("ctf"),
binary=1)), nx, ny)
#grp_imgdata[k] = window2d(fft( filt_table( filt_tanl( filt_ctf(fft(pad(grp_imgdata[k], nx2, ny2, 1,0.0)), grp_imgdata[k].get_attr("ctf"), binary=1), options.fl, options.aa), fifi) ),nx,ny)
#grp_imgdata[k] = filt_tanl(grp_imgdata[k], options.fl, options.aa)
'''
if i < 10 and myid == main_node:
for k in xrange(10):
grp_imgdata[k].write_image("grp%03d.hdf"%i, k)
'''
"""
if myid == main_node and i==0:
for pp in xrange(len(grp_imgdata)):
grp_imgdata[pp].write_image("pp.hdf", pp)
"""
ave, grp_imgdata = prepare_2d_forPCA(grp_imgdata)
"""
if myid == main_node and i==0:
for pp in xrange(len(grp_imgdata)):
grp_imgdata[pp].write_image("qq.hdf", pp)
"""
var = model_blank(nx, ny)
for q in grp_imgdata:
Util.add_img2(var, q)
Util.mul_scalar(var, 1.0 / (len(grp_imgdata) - 1))
# Switch to std dev
var = square_root(threshold(var))
#if options.CTF: ave, var = avgvar_ctf(grp_imgdata, mode="a")
#else: ave, var = avgvar(grp_imgdata, mode="a")
"""
if myid == main_node:
ave.write_image("avgv.hdf",i)
var.write_image("varv.hdf",i)
"""
set_params_proj(ave, [phiM, thetaM, 0.0, 0.0, 0.0])
set_params_proj(var, [phiM, thetaM, 0.0, 0.0, 0.0])
aveList.append(ave)
varList.append(var)
if options.VERBOSE:
print("%5.2f%% done on processor %d" %
(i * 100.0 / len(proj_list), myid))
if nvec > 0:
eig = pca(input_stacks=grp_imgdata,
subavg="",
mask_radius=radiuspca,
nvec=nvec,
incore=True,
shuffle=False,
genbuf=True)
for k in xrange(nvec):
set_params_proj(eig[k], [phiM, thetaM, 0.0, 0.0, 0.0])
eigList[k].append(eig[k])
"""
if myid == 0 and i == 0:
for k in xrange(nvec):
eig[k].write_image("eig.hdf", k)
"""
del imgdata
# To this point, all averages, variances, and eigenvectors are computed
if options.ave2D:
from fundamentals import fpol
if myid == main_node:
km = 0
for i in xrange(number_of_proc):
if i == main_node:
for im in xrange(len(aveList)):
aveList[im].write_image(
os.path.join(options.output_dir,
options.ave2D), km)
km += 1
else:
nl = mpi_recv(1, MPI_INT, i,
SPARX_MPI_TAG_UNIVERSAL,
MPI_COMM_WORLD)
nl = int(nl[0])
for im in xrange(nl):
ave = recv_EMData(i, im + i + 70000)
"""
nm = mpi_recv(1, MPI_INT, i, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
nm = int(nm[0])
members = mpi_recv(nm, MPI_INT, i, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
ave.set_attr('members', map(int, members))
members = mpi_recv(nm, MPI_FLOAT, i, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
ave.set_attr('pix_err', map(float, members))
members = mpi_recv(3, MPI_FLOAT, i, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
ave.set_attr('refprojdir', map(float, members))
"""
tmpvol = fpol(ave, Tracker["nx"],
Tracker["nx"], 1)
tmpvol.write_image(
os.path.join(options.output_dir,
options.ave2D), km)
km += 1
else:
mpi_send(len(aveList), 1, MPI_INT, main_node,
SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
for im in xrange(len(aveList)):
send_EMData(aveList[im], main_node, im + myid + 70000)
"""
members = aveList[im].get_attr('members')
mpi_send(len(members), 1, MPI_INT, main_node, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
mpi_send(members, len(members), MPI_INT, main_node, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
members = aveList[im].get_attr('pix_err')
mpi_send(members, len(members), MPI_FLOAT, main_node, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
try:
members = aveList[im].get_attr('refprojdir')
mpi_send(members, 3, MPI_FLOAT, main_node, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
except:
mpi_send([-999.0,-999.0,-999.0], 3, MPI_FLOAT, main_node, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
"""
if options.ave3D:
from fundamentals import fpol
if options.VERBOSE:
print("Reconstructing 3D average volume")
ave3D = recons3d_4nn_MPI(myid,
aveList,
symmetry=options.sym,
npad=options.npad)
bcast_EMData_to_all(ave3D, myid)
if myid == main_node:
line = strftime("%Y-%m-%d_%H:%M:%S", localtime()) + " =>"
ave3D = fpol(ave3D, Tracker["nx"], Tracker["nx"],
Tracker["nx"])
ave3D.write_image(
os.path.join(options.output_dir, options.ave3D))
msg = ("%-70s: %s\n" % (
"Writing to the disk volume reconstructed from averages as",
options.ave3D))
log_main.add(msg)
print(line, msg)
del ave, var, proj_list, stack, phi, theta, psi, s2x, s2y, alpha, sx, sy, mirror, aveList
if nvec > 0:
for k in xrange(nvec):
if options.VERBOSE:
print("Reconstruction eigenvolumes", k)
cont = True
ITER = 0
mask2d = model_circle(radiuspca, nx, nx)
while cont:
#print "On node %d, iteration %d"%(myid, ITER)
eig3D = recons3d_4nn_MPI(myid,
eigList[k],
symmetry=options.sym,
npad=options.npad)
bcast_EMData_to_all(eig3D, myid, main_node)
if options.fl > 0.0:
eig3D = filt_tanl(eig3D, options.fl, options.aa)
if myid == main_node:
eig3D.write_image(
os.path.join(options.outpout_dir,
"eig3d_%03d.hdf" % (k, ITER)))
Util.mul_img(eig3D,
model_circle(radiuspca, nx, nx, nx))
eig3Df, kb = prep_vol(eig3D)
del eig3D
cont = False
icont = 0
for l in xrange(len(eigList[k])):
phi, theta, psi, s2x, s2y = get_params_proj(
eigList[k][l])
proj = prgs(eig3Df, kb,
[phi, theta, psi, s2x, s2y])
cl = ccc(proj, eigList[k][l], mask2d)
if cl < 0.0:
icont += 1
cont = True
eigList[k][l] *= -1.0
u = int(cont)
u = mpi_reduce([u], 1, MPI_INT, MPI_MAX, main_node,
MPI_COMM_WORLD)
icont = mpi_reduce([icont], 1, MPI_INT, MPI_SUM,
main_node, MPI_COMM_WORLD)
if myid == main_node:
line = strftime("%Y-%m-%d_%H:%M:%S",
localtime()) + " =>"
u = int(u[0])
msg = (" Eigenvector: ", k, " number changed ",
int(icont[0]))
log_main.add(msg)
print(line, msg)
else:
u = 0
u = bcast_number_to_all(u, main_node)
cont = bool(u)
ITER += 1
del eig3Df, kb
mpi_barrier(MPI_COMM_WORLD)
del eigList, mask2d
if options.ave3D: del ave3D
if options.var2D:
from fundamentals import fpol
if myid == main_node:
km = 0
for i in xrange(number_of_proc):
if i == main_node:
for im in xrange(len(varList)):
tmpvol = fpol(varList[im], Tracker["nx"],
Tracker["nx"], 1)
tmpvol.write_image(
os.path.join(options.output_dir,
options.var2D), km)
km += 1
else:
nl = mpi_recv(1, MPI_INT, i,
SPARX_MPI_TAG_UNIVERSAL,
MPI_COMM_WORLD)
nl = int(nl[0])
for im in xrange(nl):
ave = recv_EMData(i, im + i + 70000)
tmpvol = fpol(ave, Tracker["nx"],
Tracker["nx"], 1)
tmpvol.write_image(
os.path.join(options.output_dir,
options.var2D, km))
km += 1
else:
mpi_send(len(varList), 1, MPI_INT, main_node,
SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
for im in xrange(len(varList)):
send_EMData(varList[im], main_node, im + myid +
70000) # What with the attributes??
mpi_barrier(MPI_COMM_WORLD)
if options.var3D:
if myid == main_node and options.VERBOSE:
line = strftime("%Y-%m-%d_%H:%M:%S", localtime()) + " =>"
msg = ("Reconstructing 3D variability volume")
log_main.add(msg)
print(line, msg)
t6 = time()
# radiusvar = options.radius
# if( radiusvar < 0 ): radiusvar = nx//2 -3
res = recons3d_4nn_MPI(myid,
varList,
symmetry=options.sym,
npad=options.npad)
#res = recons3d_em_MPI(varList, vol_stack, options.iter, radiusvar, options.abs, True, options.sym, options.squ)
if myid == main_node:
from fundamentals import fpol
res = fpol(res, Tracker["nx"], Tracker["nx"], Tracker["nx"])
res.write_image(os.path.join(options.output_dir,
options.var3D))
if myid == main_node:
line = strftime("%Y-%m-%d_%H:%M:%S", localtime()) + " =>"
msg = ("%-70s: %.2f\n" %
("Reconstructing 3D variability took [s]", time() - t6))
log_main.add(msg)
print(line, msg)
if options.VERBOSE:
print("Reconstruction took: %.2f [min]" %
((time() - t6) / 60))
if myid == main_node:
line = strftime("%Y-%m-%d_%H:%M:%S", localtime()) + " =>"
msg = ("%-70s: %.2f\n" %
("Total time for these computations [s]", time() - t0))
print(line, msg)
log_main.add(msg)
if options.VERBOSE:
print("Total time for these computations: %.2f [min]" %
((time() - t0) / 60))
line = strftime("%Y-%m-%d_%H:%M:%S", localtime()) + " =>"
msg = ("sx3dvariability")
print(line, msg)
log_main.add(msg)
from mpi import mpi_finalize
mpi_finalize()
if RUNNING_UNDER_MPI:
global_def.MPI = False
global_def.BATCH = False
def ali3d_MPI(stack,
ref_vol,
outdir,
maskfile=None,
ir=1,
ou=-1,
rs=1,
xr="4 2 2 1",
yr="-1",
ts="1 1 0.5 0.25",
delta="10 6 4 4",
an="-1",
center=0,
maxit=5,
term=95,
CTF=False,
fourvar=False,
snr=1.0,
ref_a="S",
sym="c1",
sort=True,
cutoff=999.99,
pix_cutoff="0",
two_tail=False,
model_jump="1 1 1 1 1",
restart=False,
save_half=False,
protos=None,
oplane=None,
lmask=-1,
ilmask=-1,
findseam=False,
vertstep=None,
hpars="-1",
hsearch="0.0 50.0",
full_output=False,
compare_repro=False,
compare_ref_free="-1",
ref_free_cutoff="-1 -1 -1 -1",
wcmask=None,
debug=False,
recon_pad=4,
olmask=75):
from alignment import Numrinit, prepare_refrings
from utilities import model_circle, get_image, drop_image, get_input_from_string
from utilities import bcast_list_to_all, bcast_number_to_all, reduce_EMData_to_root, bcast_EMData_to_all
from utilities import send_attr_dict
from utilities import get_params_proj, file_type
from fundamentals import rot_avg_image
import os
import types
from utilities import print_begin_msg, print_end_msg, print_msg
from mpi import mpi_bcast, mpi_comm_size, mpi_comm_rank, MPI_FLOAT, MPI_COMM_WORLD, mpi_barrier, mpi_reduce
from mpi import mpi_reduce, MPI_INT, MPI_SUM, mpi_finalize
from filter import filt_ctf
from projection import prep_vol, prgs
from statistics import hist_list, varf3d_MPI, fsc_mask
from numpy import array, bincount, array2string, ones
number_of_proc = mpi_comm_size(MPI_COMM_WORLD)
myid = mpi_comm_rank(MPI_COMM_WORLD)
main_node = 0
if myid == main_node:
if os.path.exists(outdir):
ERROR(
'Output directory exists, please change the name and restart the program',
"ali3d_MPI", 1)
os.mkdir(outdir)
mpi_barrier(MPI_COMM_WORLD)
if debug:
from time import sleep
while not os.path.exists(outdir):
print "Node ", myid, " waiting..."
sleep(5)
info_file = os.path.join(outdir, "progress%04d" % myid)
finfo = open(info_file, 'w')
else:
finfo = None
mjump = get_input_from_string(model_jump)
xrng = get_input_from_string(xr)
if yr == "-1": yrng = xrng
else: yrng = get_input_from_string(yr)
step = get_input_from_string(ts)
delta = get_input_from_string(delta)
ref_free_cutoff = get_input_from_string(ref_free_cutoff)
pix_cutoff = get_input_from_string(pix_cutoff)
lstp = min(len(xrng), len(yrng), len(step), len(delta))
if an == "-1":
an = [-1] * lstp
else:
an = get_input_from_string(an)
# make sure pix_cutoff is set for all iterations
if len(pix_cutoff) < lstp:
for i in xrange(len(pix_cutoff), lstp):
pix_cutoff.append(pix_cutoff[-1])
# don't waste time on sub-pixel alignment for low-resolution ang incr
for i in range(len(step)):
if (delta[i] > 4 or delta[i] == -1) and step[i] < 1:
step[i] = 1
first_ring = int(ir)
rstep = int(rs)
last_ring = int(ou)
max_iter = int(maxit)
center = int(center)
nrefs = EMUtil.get_image_count(ref_vol)
nmasks = 0
if maskfile:
# read number of masks within each maskfile (mc)
nmasks = EMUtil.get_image_count(maskfile)
# open masks within maskfile (mc)
maskF = EMData.read_images(maskfile, xrange(nmasks))
vol = EMData.read_images(ref_vol, xrange(nrefs))
nx = vol[0].get_xsize()
## make sure box sizes are the same
if myid == main_node:
im = EMData.read_images(stack, [0])
bx = im[0].get_xsize()
if bx != nx:
print_msg(
"Error: Stack box size (%i) differs from initial model (%i)\n"
% (bx, nx))
sys.exit()
del im, bx
# for helical processing:
helicalrecon = False
if protos is not None or hpars != "-1" or findseam is True:
helicalrecon = True
# if no out-of-plane param set, use 5 degrees
if oplane is None:
oplane = 5.0
if protos is not None:
proto = get_input_from_string(protos)
if len(proto) != nrefs:
print_msg("Error: insufficient protofilament numbers supplied")
sys.exit()
if hpars != "-1":
hpars = get_input_from_string(hpars)
if len(hpars) != 2 * nrefs:
print_msg("Error: insufficient helical parameters supplied")
sys.exit()
## create helical parameter file for helical reconstruction
if helicalrecon is True and myid == main_node:
from hfunctions import createHpar
# create initial helical parameter files
dp = [0] * nrefs
dphi = [0] * nrefs
vdp = [0] * nrefs
vdphi = [0] * nrefs
for iref in xrange(nrefs):
hpar = os.path.join(outdir, "hpar%02d.spi" % (iref))
params = False
if hpars != "-1":
# if helical parameters explicitly given, set twist & rise
params = [float(hpars[iref * 2]), float(hpars[(iref * 2) + 1])]
dp[iref], dphi[iref], vdp[iref], vdphi[iref] = createHpar(
hpar, proto[iref], params, vertstep)
# get values for helical search parameters
hsearch = get_input_from_string(hsearch)
if len(hsearch) != 2:
print_msg("Error: specify outer and inner radii for helical search")
sys.exit()
if last_ring < 0 or last_ring > int(nx / 2) - 2:
last_ring = int(nx / 2) - 2
if myid == main_node:
# import user_functions
# user_func = user_functions.factory[user_func_name]
print_begin_msg("ali3d_MPI")
print_msg("Input stack : %s\n" % (stack))
print_msg("Reference volume : %s\n" % (ref_vol))
print_msg("Output directory : %s\n" % (outdir))
if nmasks > 0:
print_msg("Maskfile (number of masks) : %s (%i)\n" %
(maskfile, nmasks))
print_msg("Inner radius : %i\n" % (first_ring))
print_msg("Outer radius : %i\n" % (last_ring))
print_msg("Ring step : %i\n" % (rstep))
print_msg("X search range : %s\n" % (xrng))
print_msg("Y search range : %s\n" % (yrng))
print_msg("Translational step : %s\n" % (step))
print_msg("Angular step : %s\n" % (delta))
print_msg("Angular search range : %s\n" % (an))
print_msg("Maximum iteration : %i\n" % (max_iter))
print_msg("Center type : %i\n" % (center))
print_msg("CTF correction : %s\n" % (CTF))
print_msg("Signal-to-Noise Ratio : %f\n" % (snr))
print_msg("Reference projection method : %s\n" % (ref_a))
print_msg("Symmetry group : %s\n" % (sym))
print_msg("Fourier padding for 3D : %i\n" % (recon_pad))
print_msg("Number of reference models : %i\n" % (nrefs))
print_msg("Sort images between models : %s\n" % (sort))
print_msg("Allow images to jump : %s\n" % (mjump))
print_msg("CC cutoff standard dev : %f\n" % (cutoff))
print_msg("Two tail cutoff : %s\n" % (two_tail))
print_msg("Termination pix error : %f\n" % (term))
print_msg("Pixel error cutoff : %s\n" % (pix_cutoff))
print_msg("Restart : %s\n" % (restart))
print_msg("Full output : %s\n" % (full_output))
print_msg("Compare reprojections : %s\n" % (compare_repro))
print_msg("Compare ref free class avgs : %s\n" % (compare_ref_free))
print_msg("Use cutoff from ref free : %s\n" % (ref_free_cutoff))
if protos:
print_msg("Protofilament numbers : %s\n" % (proto))
print_msg("Using helical search range : %s\n" % hsearch)
if findseam is True:
print_msg("Using seam-based reconstruction\n")
if hpars != "-1":
print_msg("Using hpars : %s\n" % hpars)
if vertstep != None:
print_msg("Using vertical step : %.2f\n" % vertstep)
if save_half is True:
print_msg("Saving even/odd halves\n")
for i in xrange(100):
print_msg("*")
print_msg("\n\n")
if maskfile:
if type(maskfile) is types.StringType: mask3D = get_image(maskfile)
else: mask3D = maskfile
else: mask3D = model_circle(last_ring, nx, nx, nx)
numr = Numrinit(first_ring, last_ring, rstep, "F")
mask2D = model_circle(last_ring, nx, nx) - model_circle(first_ring, nx, nx)
fscmask = model_circle(last_ring, nx, nx, nx)
if CTF:
from filter import filt_ctf
from reconstruction_rjh import rec3D_MPI_noCTF
if myid == main_node:
active = EMUtil.get_all_attributes(stack, 'active')
list_of_particles = []
for im in xrange(len(active)):
if active[im]: list_of_particles.append(im)
del active
nima = len(list_of_particles)
else:
nima = 0
total_nima = bcast_number_to_all(nima, source_node=main_node)
if myid != main_node:
list_of_particles = [-1] * total_nima
list_of_particles = bcast_list_to_all(list_of_particles,
source_node=main_node)
image_start, image_end = MPI_start_end(total_nima, number_of_proc, myid)
# create a list of images for each node
list_of_particles = list_of_particles[image_start:image_end]
nima = len(list_of_particles)
if debug:
finfo.write("image_start, image_end: %d %d\n" %
(image_start, image_end))
finfo.flush()
data = EMData.read_images(stack, list_of_particles)
t_zero = Transform({
"type": "spider",
"phi": 0,
"theta": 0,
"psi": 0,
"tx": 0,
"ty": 0
})
transmulti = [[t_zero for i in xrange(nrefs)] for j in xrange(nima)]
for iref, im in ((iref, im) for iref in xrange(nrefs)
for im in xrange(nima)):
if nrefs == 1:
transmulti[im][iref] = data[im].get_attr("xform.projection")
else:
# if multi models, keep track of eulers for all models
try:
transmulti[im][iref] = data[im].get_attr("eulers_txty.%i" %
iref)
except:
data[im].set_attr("eulers_txty.%i" % iref, t_zero)
scoremulti = [[0.0 for i in xrange(nrefs)] for j in xrange(nima)]
pixelmulti = [[0.0 for i in xrange(nrefs)] for j in xrange(nima)]
ref_res = [0.0 for x in xrange(nrefs)]
apix = data[0].get_attr('apix_x')
# for oplane parameter, create cylindrical mask
if oplane is not None and myid == main_node:
from hfunctions import createCylMask
cmaskf = os.path.join(outdir, "mask3D_cyl.mrc")
mask3D = createCylMask(data, olmask, lmask, ilmask, cmaskf)
# if finding seam of helix, create wedge masks
if findseam is True:
wedgemask = []
for pf in xrange(nrefs):
wedgemask.append(EMData())
# wedgemask option
if wcmask is not None:
wcmask = get_input_from_string(wcmask)
if len(wcmask) != 3:
print_msg(
"Error: wcmask option requires 3 values: x y radius")
sys.exit()
# determine if particles have helix info:
try:
data[0].get_attr('h_angle')
original_data = []
boxmask = True
from hfunctions import createBoxMask
except:
boxmask = False
# prepare particles
for im in xrange(nima):
data[im].set_attr('ID', list_of_particles[im])
data[im].set_attr('pix_score', int(0))
if CTF:
# only phaseflip particles, not full CTF correction
ctf_params = data[im].get_attr("ctf")
st = Util.infomask(data[im], mask2D, False)
data[im] -= st[0]
data[im] = filt_ctf(data[im], ctf_params, sign=-1, binary=1)
data[im].set_attr('ctf_applied', 1)
# for window mask:
if boxmask is True:
h_angle = data[im].get_attr("h_angle")
original_data.append(data[im].copy())
bmask = createBoxMask(nx, apix, ou, lmask, h_angle)
data[im] *= bmask
del bmask
if debug:
finfo.write('%d loaded \n' % nima)
finfo.flush()
if myid == main_node:
# initialize data for the reference preparation function
ref_data = [mask3D, max(center, 0), None, None, None, None]
# for method -1, switch off centering in user function
from time import time
# this is needed for gathering of pixel errors
disps = []
recvcount = []
disps_score = []
recvcount_score = []
for im in xrange(number_of_proc):
if (im == main_node):
disps.append(0)
disps_score.append(0)
else:
disps.append(disps[im - 1] + recvcount[im - 1])
disps_score.append(disps_score[im - 1] + recvcount_score[im - 1])
ib, ie = MPI_start_end(total_nima, number_of_proc, im)
recvcount.append(ie - ib)
recvcount_score.append((ie - ib) * nrefs)
pixer = [0.0] * nima
cs = [0.0] * 3
total_iter = 0
volodd = EMData.read_images(ref_vol, xrange(nrefs))
voleve = EMData.read_images(ref_vol, xrange(nrefs))
if restart:
# recreate initial volumes from alignments stored in header
itout = "000_00"
for iref in xrange(nrefs):
if (nrefs == 1):
modout = ""
else:
modout = "_model_%02d" % (iref)
if (sort):
group = iref
for im in xrange(nima):
imgroup = data[im].get_attr('group')
if imgroup == iref:
data[im].set_attr('xform.projection',
transmulti[im][iref])
else:
group = int(999)
for im in xrange(nima):
data[im].set_attr('xform.projection', transmulti[im][iref])
fscfile = os.path.join(outdir, "fsc_%s%s" % (itout, modout))
vol[iref], fscc, volodd[iref], voleve[iref] = rec3D_MPI_noCTF(
data,
sym,
fscmask,
fscfile,
myid,
main_node,
index=group,
npad=recon_pad)
if myid == main_node:
if helicalrecon:
from hfunctions import processHelicalVol
vstep = None
if vertstep is not None:
vstep = (vdp[iref], vdphi[iref])
print_msg(
"Old rise and twist for model %i : %8.3f, %8.3f\n"
% (iref, dp[iref], dphi[iref]))
hvals = processHelicalVol(vol[iref], voleve[iref],
volodd[iref], iref, outdir,
itout, dp[iref], dphi[iref],
apix, hsearch, findseam, vstep,
wcmask)
(vol[iref], voleve[iref], volodd[iref], dp[iref],
dphi[iref], vdp[iref], vdphi[iref]) = hvals
print_msg(
"New rise and twist for model %i : %8.3f, %8.3f\n"
% (iref, dp[iref], dphi[iref]))
# get new FSC from symmetrized half volumes
fscc = fsc_mask(volodd[iref], voleve[iref], mask3D, rstep,
fscfile)
else:
vol[iref].write_image(
os.path.join(outdir, "vol_%s.hdf" % itout), -1)
if save_half is True:
volodd[iref].write_image(
os.path.join(outdir, "volodd_%s.hdf" % itout), -1)
voleve[iref].write_image(
os.path.join(outdir, "voleve_%s.hdf" % itout), -1)
if nmasks > 1:
# Read mask for multiplying
ref_data[0] = maskF[iref]
ref_data[2] = vol[iref]
ref_data[3] = fscc
# call user-supplied function to prepare reference image, i.e., center and filter it
vol[iref], cs, fl = ref_ali3d(ref_data)
vol[iref].write_image(
os.path.join(outdir, "volf_%s.hdf" % (itout)), -1)
if (apix == 1):
res_msg = "Models filtered at spatial frequency of:\t"
res = fl
else:
res_msg = "Models filtered at resolution of: \t"
res = apix / fl
ares = array2string(array(res), precision=2)
print_msg("%s%s\n\n" % (res_msg, ares))
bcast_EMData_to_all(vol[iref], myid, main_node)
# write out headers, under MPI writing has to be done sequentially
mpi_barrier(MPI_COMM_WORLD)
# projection matching
for N_step in xrange(lstp):
terminate = 0
Iter = -1
while (Iter < max_iter - 1 and terminate == 0):
Iter += 1
total_iter += 1
itout = "%03g_%02d" % (delta[N_step], Iter)
if myid == main_node:
print_msg(
"ITERATION #%3d, inner iteration #%3d\nDelta = %4.1f, an = %5.2f, xrange = %5.2f, yrange = %5.2f, step = %5.2f\n\n"
% (N_step, Iter, delta[N_step], an[N_step], xrng[N_step],
yrng[N_step], step[N_step]))
for iref in xrange(nrefs):
if myid == main_node: start_time = time()
volft, kb = prep_vol(vol[iref])
## constrain projections to out of plane parameter
theta1 = None
theta2 = None
if oplane is not None:
theta1 = 90 - oplane
theta2 = 90 + oplane
refrings = prepare_refrings(volft,
kb,
nx,
delta[N_step],
ref_a,
sym,
numr,
MPI=True,
phiEqpsi="Minus",
initial_theta=theta1,
delta_theta=theta2)
del volft, kb
if myid == main_node:
print_msg(
"Time to prepare projections for model %i: %s\n" %
(iref, legibleTime(time() - start_time)))
start_time = time()
for im in xrange(nima):
data[im].set_attr("xform.projection", transmulti[im][iref])
if an[N_step] == -1:
t1, peak, pixer[im] = proj_ali_incore(
data[im], refrings, numr, xrng[N_step],
yrng[N_step], step[N_step], finfo)
else:
t1, peak, pixer[im] = proj_ali_incore_local(
data[im], refrings, numr, xrng[N_step],
yrng[N_step], step[N_step], an[N_step], finfo)
#data[im].set_attr("xform.projection"%iref, t1)
if nrefs > 1:
data[im].set_attr("eulers_txty.%i" % iref, t1)
scoremulti[im][iref] = peak
from pixel_error import max_3D_pixel_error
# t1 is the current param, t2 is old
t2 = transmulti[im][iref]
pixelmulti[im][iref] = max_3D_pixel_error(t1, t2, numr[-3])
transmulti[im][iref] = t1
if myid == main_node:
print_msg("Time of alignment for model %i: %s\n" %
(iref, legibleTime(time() - start_time)))
start_time = time()
# gather scoring data from all processors
from mpi import mpi_gatherv
scoremultisend = sum(scoremulti, [])
pixelmultisend = sum(pixelmulti, [])
tmp = mpi_gatherv(scoremultisend, len(scoremultisend), MPI_FLOAT,
recvcount_score, disps_score, MPI_FLOAT,
main_node, MPI_COMM_WORLD)
tmp1 = mpi_gatherv(pixelmultisend, len(pixelmultisend), MPI_FLOAT,
recvcount_score, disps_score, MPI_FLOAT,
main_node, MPI_COMM_WORLD)
tmp = mpi_bcast(tmp, (total_nima * nrefs), MPI_FLOAT, 0,
MPI_COMM_WORLD)
tmp1 = mpi_bcast(tmp1, (total_nima * nrefs), MPI_FLOAT, 0,
MPI_COMM_WORLD)
tmp = map(float, tmp)
tmp1 = map(float, tmp1)
score = array(tmp).reshape(-1, nrefs)
pixelerror = array(tmp1).reshape(-1, nrefs)
score_local = array(scoremulti)
mean_score = score.mean(axis=0)
std_score = score.std(axis=0)
cut = mean_score - (cutoff * std_score)
cut2 = mean_score + (cutoff * std_score)
res_max = score_local.argmax(axis=1)
minus_cc = [0.0 for x in xrange(nrefs)]
minus_pix = [0.0 for x in xrange(nrefs)]
minus_ref = [0.0 for x in xrange(nrefs)]
#output pixel errors
if (myid == main_node):
from statistics import hist_list
lhist = 20
pixmin = pixelerror.min(axis=1)
region, histo = hist_list(pixmin, lhist)
if (region[0] < 0.0): region[0] = 0.0
print_msg(
"Histogram of pixel errors\n ERROR number of particles\n"
)
for lhx in xrange(lhist):
print_msg(" %10.3f %7d\n" % (region[lhx], histo[lhx]))
# Terminate if 95% within 1 pixel error
im = 0
for lhx in xrange(lhist):
if (region[lhx] > 1.0): break
im += histo[lhx]
print_msg("Percent of particles with pixel error < 1: %f\n\n" %
(im / float(total_nima) * 100))
term_cond = float(term) / 100
if (im / float(total_nima) > term_cond):
terminate = 1
print_msg("Terminating internal loop\n")
del region, histo
terminate = mpi_bcast(terminate, 1, MPI_INT, 0, MPI_COMM_WORLD)
terminate = int(terminate[0])
for im in xrange(nima):
if (sort == False):
data[im].set_attr('group', 999)
elif (mjump[N_step] == 1):
data[im].set_attr('group', int(res_max[im]))
pix_run = data[im].get_attr('pix_score')
if (pix_cutoff[N_step] == 1
and (terminate == 1 or Iter == max_iter - 1)):
if (pixelmulti[im][int(res_max[im])] > 1):
data[im].set_attr('pix_score', int(777))
if (score_local[im][int(res_max[im])] < cut[int(
res_max[im])]) or (two_tail and score_local[im][int(
res_max[im])] > cut2[int(res_max[im])]):
data[im].set_attr('group', int(888))
minus_cc[int(res_max[im])] = minus_cc[int(res_max[im])] + 1
if (pix_run == 777):
data[im].set_attr('group', int(777))
minus_pix[int(
res_max[im])] = minus_pix[int(res_max[im])] + 1
if (compare_ref_free != "-1") and (ref_free_cutoff[N_step] !=
-1) and (total_iter > 1):
id = data[im].get_attr('ID')
if id in rejects:
data[im].set_attr('group', int(666))
minus_ref[int(
res_max[im])] = minus_ref[int(res_max[im])] + 1
minus_cc_tot = mpi_reduce(minus_cc, nrefs, MPI_FLOAT, MPI_SUM, 0,
MPI_COMM_WORLD)
minus_pix_tot = mpi_reduce(minus_pix, nrefs, MPI_FLOAT, MPI_SUM, 0,
MPI_COMM_WORLD)
minus_ref_tot = mpi_reduce(minus_ref, nrefs, MPI_FLOAT, MPI_SUM, 0,
MPI_COMM_WORLD)
if (myid == main_node):
if (sort):
tot_max = score.argmax(axis=1)
res = bincount(tot_max)
else:
res = ones(nrefs) * total_nima
print_msg("Particle distribution: \t\t%s\n" % (res * 1.0))
afcut1 = res - minus_cc_tot
afcut2 = afcut1 - minus_pix_tot
afcut3 = afcut2 - minus_ref_tot
print_msg("Particle distribution after cc cutoff:\t\t%s\n" %
(afcut1))
print_msg("Particle distribution after pix cutoff:\t\t%s\n" %
(afcut2))
print_msg("Particle distribution after ref cutoff:\t\t%s\n\n" %
(afcut3))
res = [0.0 for i in xrange(nrefs)]
for iref in xrange(nrefs):
if (center == -1):
from utilities import estimate_3D_center_MPI, rotate_3D_shift
dummy = EMData()
cs[0], cs[1], cs[2], dummy, dummy = estimate_3D_center_MPI(
data, total_nima, myid, number_of_proc, main_node)
cs = mpi_bcast(cs, 3, MPI_FLOAT, main_node, MPI_COMM_WORLD)
cs = [-float(cs[0]), -float(cs[1]), -float(cs[2])]
rotate_3D_shift(data, cs)
if (sort):
group = iref
for im in xrange(nima):
imgroup = data[im].get_attr('group')
if imgroup == iref:
data[im].set_attr('xform.projection',
transmulti[im][iref])
else:
group = int(999)
for im in xrange(nima):
data[im].set_attr('xform.projection',
transmulti[im][iref])
if (nrefs == 1):
modout = ""
else:
modout = "_model_%02d" % (iref)
fscfile = os.path.join(outdir, "fsc_%s%s" % (itout, modout))
vol[iref], fscc, volodd[iref], voleve[iref] = rec3D_MPI_noCTF(
data,
sym,
fscmask,
fscfile,
myid,
main_node,
index=group,
npad=recon_pad)
if myid == main_node:
print_msg("3D reconstruction time for model %i: %s\n" %
(iref, legibleTime(time() - start_time)))
start_time = time()
# Compute Fourier variance
if fourvar:
outvar = os.path.join(outdir, "volVar_%s.hdf" % (itout))
ssnr_file = os.path.join(outdir, "ssnr_%s" % (itout))
varf = varf3d_MPI(data,
ssnr_text_file=ssnr_file,
mask2D=None,
reference_structure=vol[iref],
ou=last_ring,
rw=1.0,
npad=1,
CTF=None,
sign=1,
sym=sym,
myid=myid)
if myid == main_node:
print_msg(
"Time to calculate 3D Fourier variance for model %i: %s\n"
% (iref, legibleTime(time() - start_time)))
start_time = time()
varf = 1.0 / varf
varf.write_image(outvar, -1)
else:
varf = None
if myid == main_node:
if helicalrecon:
from hfunctions import processHelicalVol
vstep = None
if vertstep is not None:
vstep = (vdp[iref], vdphi[iref])
print_msg(
"Old rise and twist for model %i : %8.3f, %8.3f\n"
% (iref, dp[iref], dphi[iref]))
hvals = processHelicalVol(vol[iref], voleve[iref],
volodd[iref], iref, outdir,
itout, dp[iref], dphi[iref],
apix, hsearch, findseam,
vstep, wcmask)
(vol[iref], voleve[iref], volodd[iref], dp[iref],
dphi[iref], vdp[iref], vdphi[iref]) = hvals
print_msg(
"New rise and twist for model %i : %8.3f, %8.3f\n"
% (iref, dp[iref], dphi[iref]))
# get new FSC from symmetrized half volumes
fscc = fsc_mask(volodd[iref], voleve[iref], mask3D,
rstep, fscfile)
print_msg(
"Time to search and apply helical symmetry for model %i: %s\n\n"
% (iref, legibleTime(time() - start_time)))
start_time = time()
else:
vol[iref].write_image(
os.path.join(outdir, "vol_%s.hdf" % (itout)), -1)
if save_half is True:
volodd[iref].write_image(
os.path.join(outdir, "volodd_%s.hdf" % (itout)),
-1)
voleve[iref].write_image(
os.path.join(outdir, "voleve_%s.hdf" % (itout)),
-1)
if nmasks > 1:
# Read mask for multiplying
ref_data[0] = maskF[iref]
ref_data[2] = vol[iref]
ref_data[3] = fscc
ref_data[4] = varf
# call user-supplied function to prepare reference image, i.e., center and filter it
vol[iref], cs, fl = ref_ali3d(ref_data)
vol[iref].write_image(
os.path.join(outdir, "volf_%s.hdf" % (itout)), -1)
if (apix == 1):
res_msg = "Models filtered at spatial frequency of:\t"
res[iref] = fl
else:
res_msg = "Models filtered at resolution of: \t"
res[iref] = apix / fl
del varf
bcast_EMData_to_all(vol[iref], myid, main_node)
if compare_ref_free != "-1": compare_repro = True
if compare_repro:
outfile_repro = comp_rep(refrings, data, itout, modout,
vol[iref], group, nima, nx, myid,
main_node, outdir)
mpi_barrier(MPI_COMM_WORLD)
if compare_ref_free != "-1":
ref_free_output = os.path.join(
outdir, "ref_free_%s%s" % (itout, modout))
rejects = compare(compare_ref_free, outfile_repro,
ref_free_output, yrng[N_step],
xrng[N_step], rstep, nx, apix,
ref_free_cutoff[N_step],
number_of_proc, myid, main_node)
# retrieve alignment params from all processors
par_str = ['xform.projection', 'ID', 'group']
if nrefs > 1:
for iref in xrange(nrefs):
par_str.append('eulers_txty.%i' % iref)
if myid == main_node:
from utilities import recv_attr_dict
recv_attr_dict(main_node, stack, data, par_str, image_start,
image_end, number_of_proc)
else:
send_attr_dict(main_node, data, par_str, image_start,
image_end)
if myid == main_node:
ares = array2string(array(res), precision=2)
print_msg("%s%s\n\n" % (res_msg, ares))
dummy = EMData()
if full_output:
nimat = EMUtil.get_image_count(stack)
output_file = os.path.join(outdir, "paramout_%s" % itout)
foutput = open(output_file, 'w')
for im in xrange(nimat):
# save the parameters for each of the models
outstring = ""
dummy.read_image(stack, im, True)
param3d = dummy.get_attr('xform.projection')
g = dummy.get_attr("group")
# retrieve alignments in EMAN-format
pE = param3d.get_params('eman')
outstring += "%f\t%f\t%f\t%f\t%f\t%i\n" % (
pE["az"], pE["alt"], pE["phi"], pE["tx"], pE["ty"],
g)
foutput.write(outstring)
foutput.close()
del dummy
mpi_barrier(MPI_COMM_WORLD)
# mpi_finalize()
if myid == main_node: print_end_msg("ali3d_MPI")
def cml_find_structure2(Prj, Ori, Rot, outdir, outname, maxit, first_zero,
flag_weights, myid, main_node, number_of_proc):
from projection import cml_export_progress, cml_disc, cml_export_txtagls
import time, sys
from random import shuffle, random
from mpi import MPI_FLOAT, MPI_INT, MPI_SUM, MPI_COMM_WORLD
from mpi import mpi_reduce, mpi_bcast, mpi_barrier
# global vars
global g_i_prj, g_n_prj, g_n_anglst, g_anglst, g_d_psi, g_debug, g_n_lines, g_seq
# list of free orientation
ocp = [-1] * g_n_anglst
if first_zero:
listprj = range(1, g_n_prj)
ocp[0] = 0
else:
listprj = range(g_n_prj)
# to stop when the solution oscillates
period_disc = [0, 0, 0]
period_ct = 0
period_th = 2
#if not flag_weights: weights = [1.0] * g_n_lines
# iteration loop
for ite in xrange(maxit):
#print ">>>>>>>>>>>>>>>>>>>>>>>>>>>>>> ite = ", ite, " myid = ", myid
t_start = time.time()
# loop over i prj
change = False
tlistprj = listprj[:]
shuffle(tlistprj)
nnn = len(tlistprj)
tlistprj = mpi_bcast(tlistprj, nnn, MPI_INT, main_node, MPI_COMM_WORLD)
tlistprj = map(int, tlistprj)
"""
if(ite>1 and ite%5 == 0 and ite<140):
if(myid == main_node):
for i in xrange(0,len(tlistprj),5):
ind = 4*i
Ori[ind] = 360.*random()
Ori[ind+1] = 180.*random()
Ori[ind+2] = 360.*random()
Ori[ind+3] = -1
for i in xrange(len(tlistprj)):
ind = 4*i
Ori[ind+3] = float(Ori[ind+3])
nnn = len(Ori)
Ori = mpi_bcast(Ori, nnn, MPI_FLOAT, main_node, MPI_COMM_WORLD)
Ori = map(float, Ori)
for i in xrange(len(tlistprj)):
ind = 4*i
Ori[ind+3] = int(Ori[ind+3])
"""
for iprj in tlistprj:
#print "********************************** iprj = ", iprj, g_n_anglst
# Store current the current orientation
ind = 4 * iprj
store_phi = Ori[ind]
store_theta = Ori[ind + 1]
store_psi = Ori[ind + 2]
cur_agl = Ori[ind + 3]
if cur_agl != -1: ocp[cur_agl] = -1
# prepare active index of cml for weighting in order to earn time later
iw = [0] * (g_n_prj - 1)
c = 0
ct = 0
for i in xrange(g_n_prj):
for j in xrange(i + 1, g_n_prj):
if i == iprj or j == iprj:
iw[ct] = c
ct += 1
c += 1
# loop over all angles
best_disc_list = [0] * g_n_anglst
best_psi_list = [0] * g_n_anglst
for iagl in xrange(myid, g_n_anglst, number_of_proc):
# if orientation is free
if ocp[iagl] == -1:
# assign new orientation
Ori[ind] = g_anglst[iagl][0]
Ori[ind + 1] = g_anglst[iagl][1]
Rot = Util.cml_update_rot(Rot, iprj, Ori[ind],
Ori[ind + 1], 0.0)
# weights
if flag_weights:
cml = Util.cml_line_in3d(Ori, g_seq, g_n_prj,
g_n_lines)
weights = Util.cml_weights(cml)
mw = max(weights)
for i in xrange(g_n_lines):
weights[i] = mw - weights[i]
sw = sum(weights)
if sw == 0:
weights = [6.28 / float(g_n_lines)] * g_n_lines
else:
for i in xrange(g_n_lines):
weights[i] /= sw
weights[i] *= weights[i]
# spin all psi
com = Util.cml_line_insino(Rot, iprj, g_n_prj)
if flag_weights:
res = Util.cml_spin_psi(Prj, com, weights, iprj, iw,
g_n_psi, g_d_psi, g_n_prj)
else:
res = Util.cml_spin_psi_now(Prj, com, iprj, iw,
g_n_psi, g_d_psi, g_n_prj)
# select the best
best_disc_list[iagl] = res[0]
best_psi_list[iagl] = res[1]
if g_debug:
cml_export_progress(outdir, ite, iprj, iagl, res[1],
res[0], 'progress')
else:
if g_debug:
cml_export_progress(outdir, ite, iprj, iagl, -1, -1,
'progress')
best_disc_list = mpi_reduce(best_disc_list, g_n_anglst, MPI_FLOAT,
MPI_SUM, main_node, MPI_COMM_WORLD)
best_psi_list = mpi_reduce(best_psi_list, g_n_anglst, MPI_FLOAT,
MPI_SUM, main_node, MPI_COMM_WORLD)
best_psi = -1
best_iagl = -1
if myid == main_node:
best_disc = 1.0e20
for iagl in xrange(g_n_anglst):
if best_disc_list[iagl] > 0.0 and best_disc_list[
iagl] < best_disc:
best_disc = best_disc_list[iagl]
best_psi = best_psi_list[iagl]
best_iagl = iagl
best_psi = mpi_bcast(best_psi, 1, MPI_FLOAT, main_node,
MPI_COMM_WORLD)
best_iagl = mpi_bcast(best_iagl, 1, MPI_INT, main_node,
MPI_COMM_WORLD)
best_psi = float(best_psi[0])
best_iagl = int(best_iagl[0])
#print "xxxxx myid = ", myid, " best_psi = ", best_psi, " best_ialg = ", best_iagl
# if change, assign
if best_iagl != cur_agl:
ocp[best_iagl] = iprj
Ori[ind] = g_anglst[best_iagl][0] # phi
Ori[ind + 1] = g_anglst[best_iagl][1] # theta
Ori[ind + 2] = best_psi * g_d_psi # psi
Ori[ind + 3] = best_iagl # index
change = True
else:
if cur_agl != -1: ocp[cur_agl] = iprj
Ori[ind] = store_phi
Ori[ind + 1] = store_theta
Ori[ind + 2] = store_psi
Ori[ind + 3] = cur_agl
Rot = Util.cml_update_rot(Rot, iprj, Ori[ind], Ori[ind + 1],
Ori[ind + 2])
if g_debug:
cml_export_progress(outdir, ite, iprj, best_iagl,
best_psi * g_d_psi, best_disc, 'choose')
# if one change, compute new full disc
disc = cml_disc(Prj, Ori, Rot, flag_weights)
# display in the progress file
if myid == main_node:
cml_export_txtagls(outdir, outname, Ori, disc,
'Ite: %03i' % (ite + 1))
if not change: break
# to stop when the solution oscillates
period_disc.pop(0)
period_disc.append(disc)
if period_disc[0] == period_disc[2]:
period_ct += 1
if period_ct >= period_th and min(
period_disc) == disc and myid == main_node:
angfile = open(outdir + '/' + outname, 'a')
angfile.write('\nSTOP SOLUTION UNSTABLE\n')
angfile.write('Discrepancy period: %s\n' % period_disc)
angfile.close()
break
else:
period_ct = 0
mpi_barrier(MPI_COMM_WORLD)
return Ori, disc, ite
def main():
progname = os.path.basename(sys.argv[0])
usage = progname + """ input_micrograph_list_file input_micrograph_pattern input_coordinates_pattern output_directory --coordinates_format --box_size=box_size --invert --import_ctf=ctf_file --limit_ctf --resample_ratio=resample_ratio --defocus_error=defocus_error --astigmatism_error=astigmatism_error
Window particles from micrographs in input list file. The coordinates of the particles should be given as input.
Please specify name pattern of input micrographs and coordinates files with a wild card (*). Use the wild card to indicate the place of micrograph ID (e.g. serial number, time stamp, and etc).
The name patterns must be enclosed by single quotes (') or double quotes ("). (Note: sxgui.py automatically adds single quotes (')).
BDB files can not be selected as input micrographs.
sxwindow.py mic_list.txt ./mic*.hdf info/mic*_info.json particles --coordinates_format=eman2 --box_size=64 --invert --import_ctf=outdir_cter/partres/partres.txt
If micrograph list file name is not provided, all files matched with the micrograph name pattern will be processed.
sxwindow.py ./mic*.hdf info/mic*_info.json particles --coordinates_format=eman2 --box_size=64 --invert --import_ctf=outdir_cter/partres/partres.txt
"""
parser = OptionParser(usage, version=SPARXVERSION)
parser.add_option(
"--coordinates_format",
type="string",
default="eman1",
help=
"format of input coordinates files: 'sparx', 'eman1', 'eman2', or 'spider'. the coordinates of sparx, eman2, and spider format is particle center. the coordinates of eman1 format is particle box conner associated with the original box size. (default eman1)"
)
parser.add_option(
"--box_size",
type="int",
default=256,
help=
"x and y dimension of square area to be windowed (in pixels): pixel size after resampling is assumed when resample_ratio < 1.0 (default 256)"
)
parser.add_option(
"--invert",
action="store_true",
default=False,
help="invert image contrast: recommended for cryo data (default False)"
)
parser.add_option(
"--import_ctf",
type="string",
default="",
help="file name of sxcter output: normally partres.txt (default none)")
parser.add_option(
"--limit_ctf",
action="store_true",
default=False,
help=
"filter micrographs based on the CTF limit: this option requires --import_ctf. (default False)"
)
parser.add_option(
"--resample_ratio",
type="float",
default=1.0,
help=
"ratio of new to old image size (or old to new pixel size) for resampling: Valid range is 0.0 < resample_ratio <= 1.0. (default 1.0)"
)
parser.add_option(
"--defocus_error",
type="float",
default=1000000.0,
help=
"defocus errror limit: exclude micrographs whose relative defocus error as estimated by sxcter is larger than defocus_error percent. the error is computed as (std dev defocus)/defocus*100%. (default 1000000.0)"
)
parser.add_option(
"--astigmatism_error",
type="float",
default=360.0,
help=
"astigmatism error limit: Set to zero astigmatism for micrographs whose astigmatism angular error as estimated by sxcter is larger than astigmatism_error degrees. (default 360.0)"
)
### detect if program is running under MPI
RUNNING_UNDER_MPI = "OMPI_COMM_WORLD_SIZE" in os.environ
main_node = 0
if RUNNING_UNDER_MPI:
from mpi import mpi_init
from mpi import MPI_COMM_WORLD, mpi_comm_rank, mpi_comm_size, mpi_barrier, mpi_reduce, MPI_INT, MPI_SUM
mpi_init(0, [])
myid = mpi_comm_rank(MPI_COMM_WORLD)
number_of_processes = mpi_comm_size(MPI_COMM_WORLD)
else:
number_of_processes = 1
myid = 0
(options, args) = parser.parse_args(sys.argv[1:])
mic_list_file_path = None
mic_pattern = None
coords_pattern = None
error_status = None
while True:
if len(args) < 3 or len(args) > 4:
error_status = (
"Please check usage for number of arguments.\n Usage: " +
usage + "\n" + "Please run %s -h for help." % (progname),
getframeinfo(currentframe()))
break
if len(args) == 3:
mic_pattern = args[0]
coords_pattern = args[1]
out_dir = args[2]
else: # assert(len(args) == 4)
mic_list_file_path = args[0]
mic_pattern = args[1]
coords_pattern = args[2]
out_dir = args[3]
if mic_list_file_path != None:
if os.path.splitext(mic_list_file_path)[1] != ".txt":
error_status = (
"Extension of input micrograph list file must be \".txt\". Please check input_micrograph_list_file argument. Run %s -h for help."
% (progname), getframeinfo(currentframe()))
break
if mic_pattern[:len("bdb:")].lower() == "bdb":
error_status = (
"BDB file can not be selected as input micrographs. Please convert the format, and restart the program. Run %s -h for help."
% (progname), getframeinfo(currentframe()))
break
if mic_pattern.find("*") == -1:
error_status = (
"Input micrograph file name pattern must contain wild card (*). Please check input_micrograph_pattern argument. Run %s -h for help."
% (progname), getframeinfo(currentframe()))
break
if coords_pattern.find("*") == -1:
error_status = (
"Input coordinates file name pattern must contain wild card (*). Please check input_coordinates_pattern argument. Run %s -h for help."
% (progname), getframeinfo(currentframe()))
break
if myid == main_node:
if os.path.exists(out_dir):
error_status = (
"Output directory exists. Please change the name and restart the program.",
getframeinfo(currentframe()))
break
break
if_error_then_all_processes_exit_program(error_status)
# Check invalid conditions of options
check_options(options, progname)
mic_name_list = None
error_status = None
if myid == main_node:
if mic_list_file_path != None:
print("Loading micrograph list from %s file ..." %
(mic_list_file_path))
mic_name_list = read_text_file(mic_list_file_path)
if len(mic_name_list) == 0:
print("Directory of first micrograph entry is " %
(os.path.dirname(mic_name_list[0])))
else: # assert (mic_list_file_path == None)
print("Generating micrograph list in %s directory..." %
(os.path.dirname(mic_pattern)))
mic_name_list = glob.glob(mic_pattern)
if len(mic_name_list) == 0:
error_status = (
"No micrograph file is found. Please check input_micrograph_pattern and/or input_micrograph_list_file argument. Run %s -h for help."
% (progname), getframeinfo(currentframe()))
else:
print("Found %d microgarphs" % len(mic_name_list))
if_error_then_all_processes_exit_program(error_status)
if RUNNING_UNDER_MPI:
mic_name_list = wrap_mpi_bcast(mic_name_list, main_node)
coords_name_list = None
error_status = None
if myid == main_node:
coords_name_list = glob.glob(coords_pattern)
if len(coords_name_list) == 0:
error_status = (
"No coordinates file is found. Please check input_coordinates_pattern argument. Run %s -h for help."
% (progname), getframeinfo(currentframe()))
if_error_then_all_processes_exit_program(error_status)
if RUNNING_UNDER_MPI:
coords_name_list = wrap_mpi_bcast(coords_name_list, main_node)
##################################################################################################################################################################################################################
##################################################################################################################################################################################################################
##################################################################################################################################################################################################################
# all processes must have access to indices
if options.import_ctf:
i_enum = -1
i_enum += 1
idx_cter_def = i_enum # defocus [um]; index must be same as ctf object format
i_enum += 1
idx_cter_cs = i_enum # Cs [mm]; index must be same as ctf object format
i_enum += 1
idx_cter_vol = i_enum # voltage[kV]; index must be same as ctf object format
i_enum += 1
idx_cter_apix = i_enum # pixel size [A]; index must be same as ctf object format
i_enum += 1
idx_cter_bfactor = i_enum # B-factor [A^2]; index must be same as ctf object format
i_enum += 1
idx_cter_ac = i_enum # amplitude contrast [%]; index must be same as ctf object format
i_enum += 1
idx_cter_astig_amp = i_enum # astigmatism amplitude [um]; index must be same as ctf object format
i_enum += 1
idx_cter_astig_ang = i_enum # astigmatism angle [degree]; index must be same as ctf object format
i_enum += 1
idx_cter_sd_def = i_enum # std dev of defocus [um]
i_enum += 1
idx_cter_sd_astig_amp = i_enum # std dev of ast amp [A]
i_enum += 1
idx_cter_sd_astig_ang = i_enum # std dev of ast angle [degree]
i_enum += 1
idx_cter_cv_def = i_enum # coefficient of variation of defocus [%]
i_enum += 1
idx_cter_cv_astig_amp = i_enum # coefficient of variation of ast amp [%]
i_enum += 1
idx_cter_spectra_diff = i_enum # average of differences between with- and without-astig. experimental 1D spectra at extrema
i_enum += 1
idx_cter_error_def = i_enum # frequency at which signal drops by 50% due to estimated error of defocus alone [1/A]
i_enum += 1
idx_cter_error_astig = i_enum # frequency at which signal drops by 50% due to estimated error of defocus and astigmatism [1/A]
i_enum += 1
idx_cter_error_ctf = i_enum # limit frequency by CTF error [1/A]
i_enum += 1
idx_cter_mic_name = i_enum # micrograph name
i_enum += 1
n_idx_cter = i_enum
# Prepare loop variables
mic_basename_pattern = os.path.basename(
mic_pattern) # file pattern without path
mic_baseroot_pattern = os.path.splitext(mic_basename_pattern)[
0] # file pattern without path and extension
coords_format = options.coordinates_format.lower()
box_size = options.box_size
box_half = box_size // 2
mask2d = model_circle(
box_size // 2, box_size, box_size
) # Create circular 2D mask to Util.infomask of particle images
resample_ratio = options.resample_ratio
n_mic_process = 0
n_mic_reject_no_coords = 0
n_mic_reject_no_cter_entry = 0
n_global_coords_detect = 0
n_global_coords_process = 0
n_global_coords_reject_out_of_boundary = 0
serial_id_list = []
error_status = None
## not a real while, an if with the opportunity to use break when errors need to be reported
while myid == main_node:
#
# NOTE: 2016/05/24 Toshio Moriya
# Now, ignores the path in mic_pattern and entries of mic_name_list to create serial ID
# Only the basename (file name) in micrograph path must be match
#
# Create list of micrograph serial ID
# Break micrograph name pattern into prefix and suffix to find the head index of the micrograph serial id
#
mic_basename_tokens = mic_basename_pattern.split('*')
# assert (len(mic_basename_tokens) == 2)
serial_id_head_index = len(mic_basename_tokens[0])
# Loop through micrograph names
for mic_name in mic_name_list:
# Find the tail index of the serial id and extract serial id from the micrograph name
mic_basename = os.path.basename(mic_name)
serial_id_tail_index = mic_basename.index(mic_basename_tokens[1])
serial_id = mic_basename[serial_id_head_index:serial_id_tail_index]
serial_id_list.append(serial_id)
# assert (len(serial_id_list) == len(mic_name))
del mic_name_list # Do not need this anymore
# Load CTFs if necessary
if options.import_ctf:
ctf_list = read_text_row(options.import_ctf)
# print("Detected CTF entries : %6d ..." % (len(ctf_list)))
if len(ctf_list) == 0:
error_status = (
"No CTF entry is found in %s. Please check --import_ctf option. Run %s -h for help."
% (options.import_ctf, progname),
getframeinfo(currentframe()))
break
if (len(ctf_list[0]) != n_idx_cter):
error_status = (
"Number of columns (%d) must be %d in %s. The format might be old. Please run sxcter.py again."
% (len(ctf_list[0]), n_idx_cter, options.import_ctf),
getframeinfo(currentframe()))
break
ctf_dict = {}
n_reject_defocus_error = 0
ctf_error_limit = [
options.defocus_error / 100.0, options.astigmatism_error
]
for ctf_params in ctf_list:
assert (len(ctf_params) == n_idx_cter)
# mic_baseroot is name of micrograph minus the path and extension
mic_baseroot = os.path.splitext(
os.path.basename(ctf_params[idx_cter_mic_name]))[0]
if (ctf_params[idx_cter_sd_def] / ctf_params[idx_cter_def] >
ctf_error_limit[0]):
print(
"Defocus error %f exceeds the threshold. Micrograph %s is rejected."
% (ctf_params[idx_cter_sd_def] /
ctf_params[idx_cter_def], mic_baseroot))
n_reject_defocus_error += 1
else:
if (ctf_params[idx_cter_sd_astig_ang] >
ctf_error_limit[1]):
ctf_params[idx_cter_astig_amp] = 0.0
ctf_params[idx_cter_astig_ang] = 0.0
ctf_dict[mic_baseroot] = ctf_params
del ctf_list # Do not need this anymore
break
if_error_then_all_processes_exit_program(error_status)
if options.import_ctf:
if options.limit_ctf:
cutoff_histogram = [
] #@ming compute the histogram for micrographs cut of by ctf_params limit.
##################################################################################################################################################################################################################
##################################################################################################################################################################################################################
##################################################################################################################################################################################################################
restricted_serial_id_list = []
if myid == main_node:
# Loop over serial IDs of micrographs
for serial_id in serial_id_list:
# mic_baseroot is name of micrograph minus the path and extension
mic_baseroot = mic_baseroot_pattern.replace("*", serial_id)
mic_name = mic_pattern.replace("*", serial_id)
coords_name = coords_pattern.replace("*", serial_id)
########### # CHECKS: BEGIN
if coords_name not in coords_name_list:
print(" Cannot read %s. Skipping %s ..." %
(coords_name, mic_baseroot))
n_mic_reject_no_coords += 1
continue
# IF mic is in CTER results
if options.import_ctf:
if mic_baseroot not in ctf_dict:
print(
" Is not listed in CTER results. Skipping %s ..." %
(mic_baseroot))
n_mic_reject_no_cter_entry += 1
continue
else:
ctf_params = ctf_dict[mic_baseroot]
# CHECKS: END
n_mic_process += 1
restricted_serial_id_list.append(serial_id)
# restricted_serial_id_list = restricted_serial_id_list[:128] ## for testing against the nonMPI version
if myid != main_node:
if options.import_ctf:
ctf_dict = None
error_status = None
if len(restricted_serial_id_list) < number_of_processes:
error_status = (
'Number of processes (%d) supplied by --np in mpirun cannot be greater than %d (number of micrographs that satisfy all criteria to be processed) '
% (number_of_processes, len(restricted_serial_id_list)),
getframeinfo(currentframe()))
if_error_then_all_processes_exit_program(error_status)
## keep a copy of the original output directory where the final bdb will be created
original_out_dir = out_dir
if RUNNING_UNDER_MPI:
mpi_barrier(MPI_COMM_WORLD)
restricted_serial_id_list = wrap_mpi_bcast(restricted_serial_id_list,
main_node)
mic_start, mic_end = MPI_start_end(len(restricted_serial_id_list),
number_of_processes, myid)
restricted_serial_id_list_not_sliced = restricted_serial_id_list
restricted_serial_id_list = restricted_serial_id_list[
mic_start:mic_end]
if options.import_ctf:
ctf_dict = wrap_mpi_bcast(ctf_dict, main_node)
# generate subdirectories of out_dir, one for each process
out_dir = os.path.join(out_dir, "%03d" % myid)
if myid == main_node:
print(
"Micrographs processed by main process (including percent complete):"
)
len_processed_by_main_node_divided_by_100 = len(
restricted_serial_id_list) / 100.0
##################################################################################################################################################################################################################
##################################################################################################################################################################################################################
##################################################################################################################################################################################################################
##### Starting main parallel execution
for my_idx, serial_id in enumerate(restricted_serial_id_list):
mic_baseroot = mic_baseroot_pattern.replace("*", serial_id)
mic_name = mic_pattern.replace("*", serial_id)
coords_name = coords_pattern.replace("*", serial_id)
if myid == main_node:
print(
mic_name, " ---> % 2.2f%%" %
(my_idx / len_processed_by_main_node_divided_by_100))
mic_img = get_im(mic_name)
# Read coordinates according to the specified format and
# make the coordinates the center of particle image
if coords_format == "sparx":
coords_list = read_text_row(coords_name)
elif coords_format == "eman1":
coords_list = read_text_row(coords_name)
for i in xrange(len(coords_list)):
coords_list[i] = [(coords_list[i][0] + coords_list[i][2] // 2),
(coords_list[i][1] + coords_list[i][3] // 2)]
elif coords_format == "eman2":
coords_list = js_open_dict(coords_name)["boxes"]
for i in xrange(len(coords_list)):
coords_list[i] = [coords_list[i][0], coords_list[i][1]]
elif coords_format == "spider":
coords_list = read_text_row(coords_name)
for i in xrange(len(coords_list)):
coords_list[i] = [coords_list[i][2], coords_list[i][3]]
# else: assert (False) # Unreachable code
# Calculate the new pixel size
if options.import_ctf:
ctf_params = ctf_dict[mic_baseroot]
pixel_size_origin = ctf_params[idx_cter_apix]
if resample_ratio < 1.0:
# assert (resample_ratio > 0.0)
new_pixel_size = pixel_size_origin / resample_ratio
print(
"Resample micrograph to pixel size %6.4f and window segments from resampled micrograph."
% new_pixel_size)
else:
# assert (resample_ratio == 1.0)
new_pixel_size = pixel_size_origin
# Set ctf along with new pixel size in resampled micrograph
ctf_params[idx_cter_apix] = new_pixel_size
else:
# assert (not options.import_ctf)
if resample_ratio < 1.0:
# assert (resample_ratio > 0.0)
print(
"Resample micrograph with ratio %6.4f and window segments from resampled micrograph."
% resample_ratio)
# else:
# assert (resample_ratio == 1.0)
# Apply filters to micrograph
fftip(mic_img)
if options.limit_ctf:
# assert (options.import_ctf)
# Cut off frequency components higher than CTF limit
q1, q2 = ctflimit(box_size, ctf_params[idx_cter_def],
ctf_params[idx_cter_cs],
ctf_params[idx_cter_vol], new_pixel_size)
# This is absolute frequency of CTF limit in scale of original micrograph
if resample_ratio < 1.0:
# assert (resample_ratio > 0.0)
q1 = resample_ratio * q1 / float(
box_size
) # q1 = (pixel_size_origin / new_pixel_size) * q1/float(box_size)
else:
# assert (resample_ratio == 1.0) -> pixel_size_origin == new_pixel_size -> pixel_size_origin / new_pixel_size == 1.0
q1 = q1 / float(box_size)
if q1 < 0.5:
mic_img = filt_tanl(mic_img, q1, 0.01)
cutoff_histogram.append(q1)
# Cut off frequency components lower than the box size can express
mic_img = fft(filt_gaussh(mic_img, resample_ratio / box_size))
# Resample micrograph, map coordinates, and window segments from resampled micrograph using new coordinates
# after resampling by resample_ratio, new pixel size will be pixel_size/resample_ratio = new_pixel_size
# NOTE: 2015/04/13 Toshio Moriya
# resample() efficiently takes care of the case resample_ratio = 1.0 but
# it does not set apix_*. Even though it sets apix_* when resample_ratio < 1.0 ...
mic_img = resample(mic_img, resample_ratio)
if options.invert:
mic_stats = Util.infomask(
mic_img, None, True) # mic_stat[0:mean, 1:SD, 2:min, 3:max]
Util.mul_scalar(mic_img, -1.0)
mic_img += 2 * mic_stats[0]
if options.import_ctf:
from utilities import generate_ctf
ctf_obj = generate_ctf(
ctf_params
) # indexes 0 to 7 (idx_cter_def to idx_cter_astig_ang) must be same in cter format & ctf object format.
# Prepare loop variables
nx = mic_img.get_xsize()
ny = mic_img.get_ysize()
x0 = nx // 2
y0 = ny // 2
n_coords_reject_out_of_boundary = 0
local_stack_name = "bdb:%s#" % out_dir + mic_baseroot + '_ptcls'
local_particle_id = 0 # can be different from coordinates_id
# Loop over coordinates
for coords_id in xrange(len(coords_list)):
x = int(coords_list[coords_id][0])
y = int(coords_list[coords_id][1])
if resample_ratio < 1.0:
# assert (resample_ratio > 0.0)
x = int(x * resample_ratio)
y = int(y * resample_ratio)
# else:
# assert(resample_ratio == 1.0)
if ((0 <= x - box_half) and (x + box_half <= nx)
and (0 <= y - box_half) and (y + box_half <= ny)):
particle_img = Util.window(mic_img, box_size, box_size, 1,
x - x0, y - y0)
else:
print(
"In %s, coordinates ID = %04d (x = %4d, y = %4d, box_size = %4d) is out of micrograph bound, skipping ..."
% (mic_baseroot, coords_id, x, y, box_size))
n_coords_reject_out_of_boundary += 1
continue
particle_img = ramp(particle_img)
particle_stats = Util.infomask(
particle_img, mask2d,
False) # particle_stats[0:mean, 1:SD, 2:min, 3:max]
particle_img -= particle_stats[0]
particle_img /= particle_stats[1]
# NOTE: 2015/04/09 Toshio Moriya
# ptcl_source_image might be redundant information ...
# Consider re-organizing header entries...
particle_img.set_attr("ptcl_source_image", mic_name)
particle_img.set_attr("ptcl_source_coord_id", coords_id)
particle_img.set_attr("ptcl_source_coord", [
int(coords_list[coords_id][0]),
int(coords_list[coords_id][1])
])
particle_img.set_attr("resample_ratio", resample_ratio)
# NOTE: 2015/04/13 Toshio Moriya
# apix_* attributes are updated by resample() only when resample_ratio != 1.0
# Let's make sure header info is consistent by setting apix_* = 1.0
# regardless of options, so it is not passed down the processing line
particle_img.set_attr("apix_x", 1.0)
particle_img.set_attr("apix_y", 1.0)
particle_img.set_attr("apix_z", 1.0)
if options.import_ctf:
particle_img.set_attr("ctf", ctf_obj)
particle_img.set_attr("ctf_applied", 0)
particle_img.set_attr("pixel_size_origin", pixel_size_origin)
# particle_img.set_attr("apix_x", new_pixel_size)
# particle_img.set_attr("apix_y", new_pixel_size)
# particle_img.set_attr("apix_z", new_pixel_size)
# NOTE: 2015/04/13 Toshio Moriya
# Pawel Comment: Micrograph is not supposed to have CTF header info.
# So, let's assume it does not exist & ignore its presence.
# Note that resample() "correctly" updates pixel size of CTF header info if it exists
# elif (particle_img.has_ctff()):
# assert(not options.import_ctf)
# ctf_origin = particle_img.get_attr("ctf_obj")
# pixel_size_origin = round(ctf_origin.apix, 5) # Because SXCTER ouputs up to 5 digits
# particle_img.set_attr("apix_x",pixel_size_origin)
# particle_img.set_attr("apix_y",pixel_size_origin)
# particle_img.set_attr("apix_z",pixel_size_origin)
# print("local_stack_name, local_particle_id", local_stack_name, local_particle_id)
particle_img.write_image(local_stack_name, local_particle_id)
local_particle_id += 1
n_global_coords_detect += len(coords_list)
n_global_coords_process += local_particle_id
n_global_coords_reject_out_of_boundary += n_coords_reject_out_of_boundary
# # MRK_DEBUG: Toshio Moriya 2016/05/03
# # Following codes are for debugging bdb. Delete in future
# result = db_check_dict(local_stack_name)
# print('# MRK_DEBUG: result = db_check_dict(local_stack_name): %s' % (result))
# result = db_list_dicts('bdb:%s' % out_dir)
# print('# MRK_DEBUG: result = db_list_dicts(out_dir): %s' % (result))
# result = db_get_image_info(local_stack_name)
# print('# MRK_DEBUG: result = db_get_image_info(local_stack_name)', result)
# Release the data base of local stack from this process
# so that the subprocess can access to the data base
db_close_dict(local_stack_name)
# # MRK_DEBUG: Toshio Moriya 2016/05/03
# # Following codes are for debugging bdb. Delete in future
# cmd_line = "e2iminfo.py %s" % (local_stack_name)
# print('# MRK_DEBUG: Executing the command: %s' % (cmd_line))
# cmdexecute(cmd_line)
# # MRK_DEBUG: Toshio Moriya 2016/05/03
# # Following codes are for debugging bdb. Delete in future
# cmd_line = "e2iminfo.py bdb:%s#data" % (out_dir)
# print('# MRK_DEBUG: Executing the command: %s' % (cmd_line))
# cmdexecute(cmd_line)
if RUNNING_UNDER_MPI:
if options.import_ctf:
if options.limit_ctf:
cutoff_histogram = wrap_mpi_gatherv(cutoff_histogram,
main_node)
if myid == main_node:
if options.limit_ctf:
# Print out the summary of CTF-limit filtering
print(" ")
print("Global summary of CTF-limit filtering (--limit_ctf) ...")
print("Percentage of filtered micrographs: %8.2f\n" %
(len(cutoff_histogram) * 100.0 /
len(restricted_serial_id_list_not_sliced)))
n_bins = 10
if len(cutoff_histogram) >= n_bins:
from statistics import hist_list
cutoff_region, cutoff_counts = hist_list(
cutoff_histogram, n_bins)
print(" Histogram of cut-off frequency")
print(" cut-off counts")
for bin_id in xrange(n_bins):
print(" %14.7f %7d" %
(cutoff_region[bin_id], cutoff_counts[bin_id]))
else:
print(
"The number of filtered micrographs (%d) is less than the number of bins (%d). No histogram is produced."
% (len(cutoff_histogram), n_bins))
n_mic_process = mpi_reduce(n_mic_process, 1, MPI_INT, MPI_SUM, main_node,
MPI_COMM_WORLD)
n_mic_reject_no_coords = mpi_reduce(n_mic_reject_no_coords, 1, MPI_INT,
MPI_SUM, main_node, MPI_COMM_WORLD)
n_mic_reject_no_cter_entry = mpi_reduce(n_mic_reject_no_cter_entry, 1,
MPI_INT, MPI_SUM, main_node,
MPI_COMM_WORLD)
n_global_coords_detect = mpi_reduce(n_global_coords_detect, 1, MPI_INT,
MPI_SUM, main_node, MPI_COMM_WORLD)
n_global_coords_process = mpi_reduce(n_global_coords_process, 1, MPI_INT,
MPI_SUM, main_node, MPI_COMM_WORLD)
n_global_coords_reject_out_of_boundary = mpi_reduce(
n_global_coords_reject_out_of_boundary, 1, MPI_INT, MPI_SUM, main_node,
MPI_COMM_WORLD)
# Print out the summary of all micrographs
if main_node == myid:
print(" ")
print("Global summary of micrographs ...")
print("Detected : %6d" %
(len(restricted_serial_id_list_not_sliced)))
print("Processed : %6d" % (n_mic_process))
print("Rejected by no coordinates file : %6d" %
(n_mic_reject_no_coords))
print("Rejected by no CTER entry : %6d" %
(n_mic_reject_no_cter_entry))
print(" ")
print("Global summary of coordinates ...")
print("Detected : %6d" %
(n_global_coords_detect))
print("Processed : %6d" %
(n_global_coords_process))
print("Rejected by out of boundary : %6d" %
(n_global_coords_reject_out_of_boundary))
# print(" ")
# print("DONE!!!")
mpi_barrier(MPI_COMM_WORLD)
if main_node == myid:
import time
time.sleep(1)
print("\n Creating bdb:%s/data\n" % original_out_dir)
for proc_i in range(number_of_processes):
mic_start, mic_end = MPI_start_end(
len(restricted_serial_id_list_not_sliced), number_of_processes,
proc_i)
for serial_id in restricted_serial_id_list_not_sliced[
mic_start:mic_end]:
e2bdb_command = "e2bdb.py "
mic_baseroot = mic_baseroot_pattern.replace("*", serial_id)
if RUNNING_UNDER_MPI:
e2bdb_command += "bdb:" + os.path.join(
original_out_dir,
"%03d/" % proc_i) + mic_baseroot + "_ptcls "
else:
e2bdb_command += "bdb:" + os.path.join(
original_out_dir, mic_baseroot + "_ptcls ")
e2bdb_command += " --appendvstack=bdb:%s/data 1>/dev/null" % original_out_dir
cmdexecute(e2bdb_command, printing_on_success=False)
print("Done!\n")
if RUNNING_UNDER_MPI:
mpi_barrier(MPI_COMM_WORLD)
from mpi import mpi_finalize
mpi_finalize()
sys.stdout.flush()
sys.exit(0)
def helicalshiftali_MPI(stack,
maskfile=None,
maxit=100,
CTF=False,
snr=1.0,
Fourvar=False,
search_rng=-1):
nproc = mpi.mpi_comm_size(mpi.MPI_COMM_WORLD)
myid = mpi.mpi_comm_rank(mpi.MPI_COMM_WORLD)
main_node = 0
ftp = file_type(stack)
if myid == main_node:
print_begin_msg("helical-shiftali_MPI")
max_iter = int(maxit)
if (myid == main_node):
infils = EMUtil.get_all_attributes(stack, "filament")
ptlcoords = EMUtil.get_all_attributes(stack, 'ptcl_source_coord')
filaments = ordersegments(infils, ptlcoords)
total_nfils = len(filaments)
inidl = [0] * total_nfils
for i in range(total_nfils):
inidl[i] = len(filaments[i])
linidl = sum(inidl)
nima = linidl
tfilaments = []
for i in range(total_nfils):
tfilaments += filaments[i]
del filaments
else:
total_nfils = 0
linidl = 0
total_nfils = bcast_number_to_all(total_nfils, source_node=main_node)
if myid != main_node:
inidl = [-1] * total_nfils
inidl = bcast_list_to_all(inidl, myid, source_node=main_node)
linidl = bcast_number_to_all(linidl, source_node=main_node)
if myid != main_node:
tfilaments = [-1] * linidl
tfilaments = bcast_list_to_all(tfilaments, myid, source_node=main_node)
filaments = []
iendi = 0
for i in range(total_nfils):
isti = iendi
iendi = isti + inidl[i]
filaments.append(tfilaments[isti:iendi])
del tfilaments, inidl
if myid == main_node:
print_msg("total number of filaments: %d" % total_nfils)
if total_nfils < nproc:
ERROR(
'number of CPUs (%i) is larger than the number of filaments (%i), please reduce the number of CPUs used'
% (nproc, total_nfils),
myid=myid)
# balanced load
temp = chunks_distribution([[len(filaments[i]), i]
for i in range(len(filaments))],
nproc)[myid:myid + 1][0]
filaments = [filaments[temp[i][1]] for i in range(len(temp))]
nfils = len(filaments)
#filaments = [[0,1]]
#print "filaments",filaments
list_of_particles = []
indcs = []
k = 0
for i in range(nfils):
list_of_particles += filaments[i]
k1 = k + len(filaments[i])
indcs.append([k, k1])
k = k1
data = EMData.read_images(stack, list_of_particles)
ldata = len(data)
sxprint("ldata=", ldata)
nx = data[0].get_xsize()
ny = data[0].get_ysize()
if maskfile == None:
mrad = min(nx, ny) // 2 - 2
mask = pad(model_blank(2 * mrad + 1, ny, 1, 1.0), nx, ny, 1, 0.0)
else:
mask = get_im(maskfile)
# apply initial xform.align2d parameters stored in header
init_params = []
for im in range(ldata):
t = data[im].get_attr('xform.align2d')
init_params.append(t)
p = t.get_params("2d")
data[im] = rot_shift2D(data[im], p['alpha'], p['tx'], p['ty'],
p['mirror'], p['scale'])
if CTF:
from sp_filter import filt_ctf
from sp_morphology import ctf_img
ctf_abs_sum = EMData(nx, ny, 1, False)
ctf_2_sum = EMData(nx, ny, 1, False)
else:
ctf_2_sum = None
ctf_abs_sum = None
from sp_utilities import info
for im in range(ldata):
data[im].set_attr('ID', list_of_particles[im])
st = Util.infomask(data[im], mask, False)
data[im] -= st[0]
if CTF:
ctf_params = data[im].get_attr("ctf")
qctf = data[im].get_attr("ctf_applied")
if qctf == 0:
data[im] = filt_ctf(fft(data[im]), ctf_params)
data[im].set_attr('ctf_applied', 1)
elif qctf != 1:
ERROR('Incorrectly set qctf flag', myid=myid)
ctfimg = ctf_img(nx, ctf_params, ny=ny)
Util.add_img2(ctf_2_sum, ctfimg)
Util.add_img_abs(ctf_abs_sum, ctfimg)
else:
data[im] = fft(data[im])
del list_of_particles
if CTF:
reduce_EMData_to_root(ctf_2_sum, myid, main_node)
reduce_EMData_to_root(ctf_abs_sum, myid, main_node)
if CTF:
if myid != main_node:
del ctf_2_sum
del ctf_abs_sum
else:
temp = EMData(nx, ny, 1, False)
tsnr = 1. / snr
for i in range(0, nx + 2, 2):
for j in range(ny):
temp.set_value_at(i, j, tsnr)
temp.set_value_at(i + 1, j, 0.0)
#info(ctf_2_sum)
Util.add_img(ctf_2_sum, temp)
#info(ctf_2_sum)
del temp
total_iter = 0
shift_x = [0.0] * ldata
for Iter in range(max_iter):
if myid == main_node:
start_time = time()
print_msg("Iteration #%4d\n" % (total_iter))
total_iter += 1
avg = EMData(nx, ny, 1, False)
for im in range(ldata):
Util.add_img(avg, fshift(data[im], shift_x[im]))
reduce_EMData_to_root(avg, myid, main_node)
if myid == main_node:
if CTF: tavg = Util.divn_filter(avg, ctf_2_sum)
else: tavg = Util.mult_scalar(avg, 1.0 / float(nima))
else:
tavg = model_blank(nx, ny)
if Fourvar:
bcast_EMData_to_all(tavg, myid, main_node)
vav, rvar = varf2d_MPI(myid, data, tavg, mask, "a", CTF)
if myid == main_node:
if Fourvar:
tavg = fft(Util.divn_img(fft(tavg), vav))
vav_r = Util.pack_complex_to_real(vav)
# normalize and mask tavg in real space
tavg = fft(tavg)
stat = Util.infomask(tavg, mask, False)
tavg -= stat[0]
Util.mul_img(tavg, mask)
tavg.write_image("tavg.hdf", Iter)
# For testing purposes: shift tavg to some random place and see if the centering is still correct
#tavg = rot_shift3D(tavg,sx=3,sy=-4)
if Fourvar: del vav
bcast_EMData_to_all(tavg, myid, main_node)
tavg = fft(tavg)
sx_sum = 0.0
nxc = nx // 2
for ifil in range(nfils):
"""
# Calculate filament average
avg = EMData(nx, ny, 1, False)
filnima = 0
for im in xrange(indcs[ifil][0], indcs[ifil][1]):
Util.add_img(avg, data[im])
filnima += 1
tavg = Util.mult_scalar(avg, 1.0/float(filnima))
"""
# Calculate 1D ccf between each segment and filament average
nsegms = indcs[ifil][1] - indcs[ifil][0]
ctx = [None] * nsegms
pcoords = [None] * nsegms
for im in range(indcs[ifil][0], indcs[ifil][1]):
ctx[im - indcs[ifil][0]] = Util.window(ccf(tavg, data[im]), nx,
1)
pcoords[im - indcs[ifil][0]] = data[im].get_attr(
'ptcl_source_coord')
#ctx[im-indcs[ifil][0]].write_image("ctx.hdf",im-indcs[ifil][0])
#print " CTX ",myid,im,Util.infomask(ctx[im-indcs[ifil][0]], None, True)
# search for best x-shift
cents = nsegms // 2
dst = sqrt(
max((pcoords[cents][0] - pcoords[0][0])**2 +
(pcoords[cents][1] - pcoords[0][1])**2,
(pcoords[cents][0] - pcoords[-1][0])**2 +
(pcoords[cents][1] - pcoords[-1][1])**2))
maxincline = atan2(ny // 2 - 2 - float(search_rng), dst)
kang = int(dst * tan(maxincline) + 0.5)
#print " settings ",nsegms,cents,dst,search_rng,maxincline,kang
# ## C code for alignment. @ming
results = [0.0] * 3
results = Util.helixshiftali(ctx, pcoords, nsegms, maxincline,
kang, search_rng, nxc)
sib = int(results[0])
bang = results[1]
qm = results[2]
#print qm, sib, bang
# qm = -1.e23
#
# for six in xrange(-search_rng, search_rng+1,1):
# q0 = ctx[cents].get_value_at(six+nxc)
# for incline in xrange(kang+1):
# qt = q0
# qu = q0
# if(kang>0): tang = tan(maxincline/kang*incline)
# else: tang = 0.0
# for kim in xrange(cents+1,nsegms):
# dst = sqrt((pcoords[cents][0] - pcoords[kim][0])**2 + (pcoords[cents][1] - pcoords[kim][1])**2)
# xl = dst*tang+six+nxc
# ixl = int(xl)
# dxl = xl - ixl
# #print " A ", ifil,six,incline,kim,xl,ixl,dxl
# qt += (1.0-dxl)*ctx[kim].get_value_at(ixl) + dxl*ctx[kim].get_value_at(ixl+1)
# xl = -dst*tang+six+nxc
# ixl = int(xl)
# dxl = xl - ixl
# qu += (1.0-dxl)*ctx[kim].get_value_at(ixl) + dxl*ctx[kim].get_value_at(ixl+1)
# for kim in xrange(cents):
# dst = sqrt((pcoords[cents][0] - pcoords[kim][0])**2 + (pcoords[cents][1] - pcoords[kim][1])**2)
# xl = -dst*tang+six+nxc
# ixl = int(xl)
# dxl = xl - ixl
# qt += (1.0-dxl)*ctx[kim].get_value_at(ixl) + dxl*ctx[kim].get_value_at(ixl+1)
# xl = dst*tang+six+nxc
# ixl = int(xl)
# dxl = xl - ixl
# qu += (1.0-dxl)*ctx[kim].get_value_at(ixl) + dxl*ctx[kim].get_value_at(ixl+1)
# if( qt > qm ):
# qm = qt
# sib = six
# bang = tang
# if( qu > qm ):
# qm = qu
# sib = six
# bang = -tang
#if incline == 0: print "incline = 0 ",six,tang,qt,qu
#print qm,six,sib,bang
#print " got results ",indcs[ifil][0], indcs[ifil][1], ifil,myid,qm,sib,tang,bang,len(ctx),Util.infomask(ctx[0], None, True)
for im in range(indcs[ifil][0], indcs[ifil][1]):
kim = im - indcs[ifil][0]
dst = sqrt((pcoords[cents][0] - pcoords[kim][0])**2 +
(pcoords[cents][1] - pcoords[kim][1])**2)
if (kim < cents): xl = -dst * bang + sib
else: xl = dst * bang + sib
shift_x[im] = xl
# Average shift
sx_sum += shift_x[indcs[ifil][0] + cents]
# #print myid,sx_sum,total_nfils
sx_sum = mpi.mpi_reduce(sx_sum, 1, mpi.MPI_FLOAT, mpi.MPI_SUM,
main_node, mpi.MPI_COMM_WORLD)
if myid == main_node:
sx_sum = float(sx_sum[0]) / total_nfils
print_msg("Average shift %6.2f\n" % (sx_sum))
else:
sx_sum = 0.0
sx_sum = 0.0
sx_sum = bcast_number_to_all(sx_sum, source_node=main_node)
for im in range(ldata):
shift_x[im] -= sx_sum
#print " %3d %6.3f"%(im,shift_x[im])
#exit()
# combine shifts found with the original parameters
for im in range(ldata):
t1 = Transform()
##import random
##shix=random.randint(-10, 10)
##t1.set_params({"type":"2D","tx":shix})
t1.set_params({"type": "2D", "tx": shift_x[im]})
# combine t0 and t1
tt = t1 * init_params[im]
data[im].set_attr("xform.align2d", tt)
# write out headers and STOP, under MPI writing has to be done sequentially
mpi.mpi_barrier(mpi.MPI_COMM_WORLD)
par_str = ["xform.align2d", "ID"]
if myid == main_node:
from sp_utilities import file_type
if (file_type(stack) == "bdb"):
from sp_utilities import recv_attr_dict_bdb
recv_attr_dict_bdb(main_node, stack, data, par_str, 0, ldata,
nproc)
else:
from sp_utilities import recv_attr_dict
recv_attr_dict(main_node, stack, data, par_str, 0, ldata, nproc)
else:
send_attr_dict(main_node, data, par_str, 0, ldata)
if myid == main_node: print_end_msg("helical-shiftali_MPI")
def main():
def params_3D_2D_NEW(phi, theta, psi, s2x, s2y, mirror):
# the final ali2d parameters already combine shifts operation first and rotation operation second for parameters converted from 3D
if mirror:
m = 1
alpha, sx, sy, scalen = compose_transform2(0, s2x, s2y, 1.0, 540.0-psi, 0, 0, 1.0)
else:
m = 0
alpha, sx, sy, scalen = compose_transform2(0, s2x, s2y, 1.0, 360.0-psi, 0, 0, 1.0)
return alpha, sx, sy, m
progname = os.path.basename(sys.argv[0])
usage = progname + " prj_stack --ave2D= --var2D= --ave3D= --var3D= --img_per_grp= --fl= --aa= --sym=symmetry --CTF"
parser = OptionParser(usage, version=SPARXVERSION)
parser.add_option("--output_dir", type="string" , default="./", help="Output directory")
parser.add_option("--ave2D", type="string" , default=False, help="Write to the disk a stack of 2D averages")
parser.add_option("--var2D", type="string" , default=False, help="Write to the disk a stack of 2D variances")
parser.add_option("--ave3D", type="string" , default=False, help="Write to the disk reconstructed 3D average")
parser.add_option("--var3D", type="string" , default=False, help="Compute 3D variability (time consuming!)")
parser.add_option("--img_per_grp", type="int" , default=100, help="Number of neighbouring projections.(Default is 100)")
parser.add_option("--no_norm", action="store_true", default=False, help="Do not use normalization.(Default is to apply normalization)")
#parser.add_option("--radius", type="int" , default=-1 , help="radius for 3D variability" )
parser.add_option("--npad", type="int" , default=2 , help="Number of time to pad the original images.(Default is 2 times padding)")
parser.add_option("--sym" , type="string" , default="c1", help="Symmetry. (Default is no symmetry)")
parser.add_option("--fl", type="float" , default=0.0, help="Low pass filter cutoff in absolute frequency (0.0 - 0.5) and is applied to decimated images. (Default - no filtration)")
parser.add_option("--aa", type="float" , default=0.02 , help="Fall off of the filter. Use default value if user has no clue about falloff (Default value is 0.02)")
parser.add_option("--CTF", action="store_true", default=False, help="Use CFT correction.(Default is no CTF correction)")
#parser.add_option("--MPI" , action="store_true", default=False, help="use MPI version")
#parser.add_option("--radiuspca", type="int" , default=-1 , help="radius for PCA" )
#parser.add_option("--iter", type="int" , default=40 , help="maximum number of iterations (stop criterion of reconstruction process)" )
#parser.add_option("--abs", type="float" , default=0.0 , help="minimum average absolute change of voxels' values (stop criterion of reconstruction process)" )
#parser.add_option("--squ", type="float" , default=0.0 , help="minimum average squared change of voxels' values (stop criterion of reconstruction process)" )
parser.add_option("--VAR" , action="store_true", default=False, help="Stack of input consists of 2D variances (Default False)")
parser.add_option("--decimate", type ="float", default=0.25, help="Image decimate rate, a number less than 1. (Default is 0.25)")
parser.add_option("--window", type ="int", default=0, help="Target image size relative to original image size. (Default value is zero.)")
#parser.add_option("--SND", action="store_true", default=False, help="compute squared normalized differences (Default False)")
#parser.add_option("--nvec", type="int" , default=0 , help="Number of eigenvectors, (Default = 0 meaning no PCA calculated)")
parser.add_option("--symmetrize", action="store_true", default=False, help="Prepare input stack for handling symmetry (Default False)")
parser.add_option("--overhead", type ="float", default=0.5, help="python overhead per CPU.")
(options,args) = parser.parse_args()
#####
from mpi import mpi_init, mpi_comm_rank, mpi_comm_size, mpi_recv, MPI_COMM_WORLD
from mpi import mpi_barrier, mpi_reduce, mpi_bcast, mpi_send, MPI_FLOAT, MPI_SUM, MPI_INT, MPI_MAX
#from mpi import *
from applications import MPI_start_end
from reconstruction import recons3d_em, recons3d_em_MPI
from reconstruction import recons3d_4nn_MPI, recons3d_4nn_ctf_MPI
from utilities import print_begin_msg, print_end_msg, print_msg
from utilities import read_text_row, get_image, get_im, wrap_mpi_send, wrap_mpi_recv
from utilities import bcast_EMData_to_all, bcast_number_to_all
from utilities import get_symt
# This is code for handling symmetries by the above program. To be incorporated. PAP 01/27/2015
from EMAN2db import db_open_dict
# Set up global variables related to bdb cache
if global_def.CACHE_DISABLE:
from utilities import disable_bdb_cache
disable_bdb_cache()
# Set up global variables related to ERROR function
global_def.BATCH = True
# detect if program is running under MPI
RUNNING_UNDER_MPI = "OMPI_COMM_WORLD_SIZE" in os.environ
if RUNNING_UNDER_MPI: global_def.MPI = True
if options.output_dir =="./": current_output_dir = os.path.abspath(options.output_dir)
else: current_output_dir = options.output_dir
if options.symmetrize :
if RUNNING_UNDER_MPI:
try:
sys.argv = mpi_init(len(sys.argv), sys.argv)
try:
number_of_proc = mpi_comm_size(MPI_COMM_WORLD)
if( number_of_proc > 1 ):
ERROR("Cannot use more than one CPU for symmetry preparation","sx3dvariability",1)
except:
pass
except:
pass
if not os.path.exists(current_output_dir): os.mkdir(current_output_dir)
# Input
#instack = "Clean_NORM_CTF_start_wparams.hdf"
#instack = "bdb:data"
from logger import Logger,BaseLogger_Files
if os.path.exists(os.path.join(current_output_dir, "log.txt")): os.remove(os.path.join(current_output_dir, "log.txt"))
log_main=Logger(BaseLogger_Files())
log_main.prefix = os.path.join(current_output_dir, "./")
instack = args[0]
sym = options.sym.lower()
if( sym == "c1" ):
ERROR("There is no need to symmetrize stack for C1 symmetry","sx3dvariability",1)
line =""
for a in sys.argv:
line +=" "+a
log_main.add(line)
if(instack[:4] !="bdb:"):
#if output_dir =="./": stack = "bdb:data"
stack = "bdb:"+current_output_dir+"/data"
delete_bdb(stack)
junk = cmdexecute("sxcpy.py "+instack+" "+stack)
else: stack = instack
qt = EMUtil.get_all_attributes(stack,'xform.projection')
na = len(qt)
ts = get_symt(sym)
ks = len(ts)
angsa = [None]*na
for k in range(ks):
#Qfile = "Q%1d"%k
#if options.output_dir!="./": Qfile = os.path.join(options.output_dir,"Q%1d"%k)
Qfile = os.path.join(current_output_dir, "Q%1d"%k)
#delete_bdb("bdb:Q%1d"%k)
delete_bdb("bdb:"+Qfile)
#junk = cmdexecute("e2bdb.py "+stack+" --makevstack=bdb:Q%1d"%k)
junk = cmdexecute("e2bdb.py "+stack+" --makevstack=bdb:"+Qfile)
#DB = db_open_dict("bdb:Q%1d"%k)
DB = db_open_dict("bdb:"+Qfile)
for i in range(na):
ut = qt[i]*ts[k]
DB.set_attr(i, "xform.projection", ut)
#bt = ut.get_params("spider")
#angsa[i] = [round(bt["phi"],3)%360.0, round(bt["theta"],3)%360.0, bt["psi"], -bt["tx"], -bt["ty"]]
#write_text_row(angsa, 'ptsma%1d.txt'%k)
#junk = cmdexecute("e2bdb.py "+stack+" --makevstack=bdb:Q%1d"%k)
#junk = cmdexecute("sxheader.py bdb:Q%1d --params=xform.projection --import=ptsma%1d.txt"%(k,k))
DB.close()
#if options.output_dir =="./": delete_bdb("bdb:sdata")
delete_bdb("bdb:" + current_output_dir + "/"+"sdata")
#junk = cmdexecute("e2bdb.py . --makevstack=bdb:sdata --filt=Q")
sdata = "bdb:"+current_output_dir+"/"+"sdata"
print(sdata)
junk = cmdexecute("e2bdb.py " + current_output_dir +" --makevstack="+sdata +" --filt=Q")
#junk = cmdexecute("ls EMAN2DB/sdata*")
#a = get_im("bdb:sdata")
a = get_im(sdata)
a.set_attr("variabilitysymmetry",sym)
#a.write_image("bdb:sdata")
a.write_image(sdata)
else:
from fundamentals import window2d
sys.argv = mpi_init(len(sys.argv), sys.argv)
myid = mpi_comm_rank(MPI_COMM_WORLD)
number_of_proc = mpi_comm_size(MPI_COMM_WORLD)
main_node = 0
shared_comm = mpi_comm_split_type(MPI_COMM_WORLD, MPI_COMM_TYPE_SHARED, 0, MPI_INFO_NULL)
myid_on_node = mpi_comm_rank(shared_comm)
no_of_processes_per_group = mpi_comm_size(shared_comm)
masters_from_groups_vs_everything_else_comm = mpi_comm_split(MPI_COMM_WORLD, main_node == myid_on_node, myid_on_node)
color, no_of_groups, balanced_processor_load_on_nodes = get_colors_and_subsets(main_node, MPI_COMM_WORLD, myid, \
shared_comm, myid_on_node, masters_from_groups_vs_everything_else_comm)
overhead_loading = options.overhead*number_of_proc
#memory_per_node = options.memory_per_node
#if memory_per_node == -1.: memory_per_node = 2.*no_of_processes_per_group
keepgoing = 1
current_window = options.window
current_decimate = options.decimate
if len(args) == 1: stack = args[0]
else:
print(( "usage: " + usage))
print(( "Please run '" + progname + " -h' for detailed options"))
return 1
t0 = time()
# obsolete flags
options.MPI = True
#options.nvec = 0
options.radiuspca = -1
options.iter = 40
options.abs = 0.0
options.squ = 0.0
if options.fl > 0.0 and options.aa == 0.0:
ERROR("Fall off has to be given for the low-pass filter", "sx3dvariability", 1, myid)
#if options.VAR and options.SND:
# ERROR("Only one of var and SND can be set!", "sx3dvariability", myid)
if options.VAR and (options.ave2D or options.ave3D or options.var2D):
ERROR("When VAR is set, the program cannot output ave2D, ave3D or var2D", "sx3dvariability", 1, myid)
#if options.SND and (options.ave2D or options.ave3D):
# ERROR("When SND is set, the program cannot output ave2D or ave3D", "sx3dvariability", 1, myid)
#if options.nvec > 0 :
# ERROR("PCA option not implemented", "sx3dvariability", 1, myid)
#if options.nvec > 0 and options.ave3D == None:
# ERROR("When doing PCA analysis, one must set ave3D", "sx3dvariability", 1, myid)
if current_decimate>1.0 or current_decimate<0.0:
ERROR("Decimate rate should be a value between 0.0 and 1.0", "sx3dvariability", 1, myid)
if current_window < 0.0:
ERROR("Target window size should be always larger than zero", "sx3dvariability", 1, myid)
if myid == main_node:
img = get_image(stack, 0)
nx = img.get_xsize()
ny = img.get_ysize()
if(min(nx, ny) < current_window): keepgoing = 0
keepgoing = bcast_number_to_all(keepgoing, main_node, MPI_COMM_WORLD)
if keepgoing == 0: ERROR("The target window size cannot be larger than the size of decimated image", "sx3dvariability", 1, myid)
import string
options.sym = options.sym.lower()
# if global_def.CACHE_DISABLE:
# from utilities import disable_bdb_cache
# disable_bdb_cache()
# global_def.BATCH = True
if myid == main_node:
if not os.path.exists(current_output_dir): os.mkdir(current_output_dir)# Never delete output_dir in the program!
img_per_grp = options.img_per_grp
#nvec = options.nvec
radiuspca = options.radiuspca
from logger import Logger,BaseLogger_Files
#if os.path.exists(os.path.join(options.output_dir, "log.txt")): os.remove(os.path.join(options.output_dir, "log.txt"))
log_main=Logger(BaseLogger_Files())
log_main.prefix = os.path.join(current_output_dir, "./")
if myid == main_node:
line = ""
for a in sys.argv: line +=" "+a
log_main.add(line)
log_main.add("-------->>>Settings given by all options<<<-------")
log_main.add("Symmetry : %s"%options.sym)
log_main.add("Input stack : %s"%stack)
log_main.add("Output_dir : %s"%current_output_dir)
if options.ave3D: log_main.add("Ave3d : %s"%options.ave3D)
if options.var3D: log_main.add("Var3d : %s"%options.var3D)
if options.ave2D: log_main.add("Ave2D : %s"%options.ave2D)
if options.var2D: log_main.add("Var2D : %s"%options.var2D)
if options.VAR: log_main.add("VAR : True")
else: log_main.add("VAR : False")
if options.CTF: log_main.add("CTF correction : True ")
else: log_main.add("CTF correction : False ")
log_main.add("Image per group : %5d"%options.img_per_grp)
log_main.add("Image decimate rate : %4.3f"%current_decimate)
log_main.add("Low pass filter : %4.3f"%options.fl)
current_fl = options.fl
if current_fl == 0.0: current_fl = 0.5
log_main.add("Current low pass filter is equivalent to cutoff frequency %4.3f for original image size"%round((current_fl*current_decimate),3))
log_main.add("Window size : %5d "%current_window)
log_main.add("sx3dvariability begins")
symbaselen = 0
if myid == main_node:
nima = EMUtil.get_image_count(stack)
img = get_image(stack)
nx = img.get_xsize()
ny = img.get_ysize()
nnxo = nx
nnyo = ny
if options.sym != "c1" :
imgdata = get_im(stack)
try:
i = imgdata.get_attr("variabilitysymmetry").lower()
if(i != options.sym):
ERROR("The symmetry provided does not agree with the symmetry of the input stack", "sx3dvariability", 1, myid)
except:
ERROR("Input stack is not prepared for symmetry, please follow instructions", "sx3dvariability", 1, myid)
from utilities import get_symt
i = len(get_symt(options.sym))
if((nima/i)*i != nima):
ERROR("The length of the input stack is incorrect for symmetry processing", "sx3dvariability", 1, myid)
symbaselen = nima/i
else: symbaselen = nima
else:
nima = 0
nx = 0
ny = 0
nnxo = 0
nnyo = 0
nima = bcast_number_to_all(nima)
nx = bcast_number_to_all(nx)
ny = bcast_number_to_all(ny)
nnxo = bcast_number_to_all(nnxo)
nnyo = bcast_number_to_all(nnyo)
if current_window > max(nx, ny):
ERROR("Window size is larger than the original image size", "sx3dvariability", 1)
if current_decimate == 1.:
if current_window !=0:
nx = current_window
ny = current_window
else:
if current_window == 0:
nx = int(nx*current_decimate+0.5)
ny = int(ny*current_decimate+0.5)
else:
nx = int(current_window*current_decimate+0.5)
ny = nx
symbaselen = bcast_number_to_all(symbaselen)
# check FFT prime number
from fundamentals import smallprime
is_fft_friendly = (nx == smallprime(nx))
if not is_fft_friendly:
if myid == main_node:
log_main.add("The target image size is not a product of small prime numbers")
log_main.add("Program adjusts the input settings!")
### two cases
if current_decimate == 1.:
nx = smallprime(nx)
ny = nx
current_window = nx # update
if myid == main_node:
log_main.add("The window size is updated to %d."%current_window)
else:
if current_window == 0:
nx = smallprime(int(nx*current_decimate+0.5))
current_decimate = float(nx)/nnxo
ny = nx
if (myid == main_node):
log_main.add("The decimate rate is updated to %f."%current_decimate)
else:
nx = smallprime(int(current_window*current_decimate+0.5))
ny = nx
current_window = int(nx/current_decimate+0.5)
if (myid == main_node):
log_main.add("The window size is updated to %d."%current_window)
if myid == main_node:
log_main.add("The target image size is %d"%nx)
if radiuspca == -1: radiuspca = nx/2-2
if myid == main_node: log_main.add("%-70s: %d\n"%("Number of projection", nima))
img_begin, img_end = MPI_start_end(nima, number_of_proc, myid)
"""
if options.SND:
from projection import prep_vol, prgs
from statistics import im_diff
from utilities import get_im, model_circle, get_params_proj, set_params_proj
from utilities import get_ctf, generate_ctf
from filter import filt_ctf
imgdata = EMData.read_images(stack, range(img_begin, img_end))
if options.CTF:
vol = recons3d_4nn_ctf_MPI(myid, imgdata, 1.0, symmetry=options.sym, npad=options.npad, xysize=-1, zsize=-1)
else:
vol = recons3d_4nn_MPI(myid, imgdata, symmetry=options.sym, npad=options.npad, xysize=-1, zsize=-1)
bcast_EMData_to_all(vol, myid)
volft, kb = prep_vol(vol)
mask = model_circle(nx/2-2, nx, ny)
varList = []
for i in xrange(img_begin, img_end):
phi, theta, psi, s2x, s2y = get_params_proj(imgdata[i-img_begin])
ref_prj = prgs(volft, kb, [phi, theta, psi, -s2x, -s2y])
if options.CTF:
ctf_params = get_ctf(imgdata[i-img_begin])
ref_prj = filt_ctf(ref_prj, generate_ctf(ctf_params))
diff, A, B = im_diff(ref_prj, imgdata[i-img_begin], mask)
diff2 = diff*diff
set_params_proj(diff2, [phi, theta, psi, s2x, s2y])
varList.append(diff2)
mpi_barrier(MPI_COMM_WORLD)
"""
if options.VAR: # 2D variance images have no shifts
#varList = EMData.read_images(stack, range(img_begin, img_end))
from EMAN2 import Region
for index_of_particle in range(img_begin,img_end):
image = get_im(stack, index_of_proj)
if current_window > 0: varList.append(fdecimate(window2d(image,current_window,current_window), nx,ny))
else: varList.append(fdecimate(image, nx,ny))
else:
from utilities import bcast_number_to_all, bcast_list_to_all, send_EMData, recv_EMData
from utilities import set_params_proj, get_params_proj, params_3D_2D, get_params2D, set_params2D, compose_transform2
from utilities import model_blank, nearest_proj, model_circle, write_text_row
from applications import pca
from statistics import avgvar, avgvar_ctf, ccc
from filter import filt_tanl
from morphology import threshold, square_root
from projection import project, prep_vol, prgs
from sets import Set
from utilities import wrap_mpi_recv, wrap_mpi_bcast, wrap_mpi_send
import numpy as np
if myid == main_node:
t1 = time()
proj_angles = []
aveList = []
tab = EMUtil.get_all_attributes(stack, 'xform.projection')
for i in range(nima):
t = tab[i].get_params('spider')
phi = t['phi']
theta = t['theta']
psi = t['psi']
x = theta
if x > 90.0: x = 180.0 - x
x = x*10000+psi
proj_angles.append([x, t['phi'], t['theta'], t['psi'], i])
t2 = time()
log_main.add( "%-70s: %d\n"%("Number of neighboring projections", img_per_grp))
log_main.add("...... Finding neighboring projections\n")
log_main.add( "Number of images per group: %d"%img_per_grp)
log_main.add( "Now grouping projections")
proj_angles.sort()
proj_angles_list = np.full((nima, 4), 0.0, dtype=np.float32)
for i in range(nima):
proj_angles_list[i][0] = proj_angles[i][1]
proj_angles_list[i][1] = proj_angles[i][2]
proj_angles_list[i][2] = proj_angles[i][3]
proj_angles_list[i][3] = proj_angles[i][4]
else: proj_angles_list = 0
proj_angles_list = wrap_mpi_bcast(proj_angles_list, main_node, MPI_COMM_WORLD)
proj_angles = []
for i in range(nima):
proj_angles.append([proj_angles_list[i][0], proj_angles_list[i][1], proj_angles_list[i][2], int(proj_angles_list[i][3])])
del proj_angles_list
proj_list, mirror_list = nearest_proj(proj_angles, img_per_grp, range(img_begin, img_end))
all_proj = Set()
for im in proj_list:
for jm in im:
all_proj.add(proj_angles[jm][3])
all_proj = list(all_proj)
index = {}
for i in range(len(all_proj)): index[all_proj[i]] = i
mpi_barrier(MPI_COMM_WORLD)
if myid == main_node:
log_main.add("%-70s: %.2f\n"%("Finding neighboring projections lasted [s]", time()-t2))
log_main.add("%-70s: %d\n"%("Number of groups processed on the main node", len(proj_list)))
log_main.add("Grouping projections took: %12.1f [m]"%((time()-t2)/60.))
log_main.add("Number of groups on main node: ", len(proj_list))
mpi_barrier(MPI_COMM_WORLD)
if myid == main_node:
log_main.add("...... Calculating the stack of 2D variances \n")
# Memory estimation. There are two memory consumption peaks
# peak 1. Compute ave, var;
# peak 2. Var volume reconstruction;
# proj_params = [0.0]*(nima*5)
aveList = []
varList = []
#if nvec > 0: eigList = [[] for i in range(nvec)]
dnumber = len(all_proj)# all neighborhood set for assigned to myid
pnumber = len(proj_list)*2. + img_per_grp # aveList and varList
tnumber = dnumber+pnumber
vol_size2 = nx**3*4.*8/1.e9
vol_size1 = 2.*nnxo**3*4.*8/1.e9
proj_size = nnxo*nnyo*len(proj_list)*4.*2./1.e9 # both aveList and varList
orig_data_size = nnxo*nnyo*4.*tnumber/1.e9
reduced_data_size = nx*nx*4.*tnumber/1.e9
full_data = np.full((number_of_proc, 2), -1., dtype=np.float16)
full_data[myid] = orig_data_size, reduced_data_size
if myid != main_node: wrap_mpi_send(full_data, main_node, MPI_COMM_WORLD)
if myid == main_node:
for iproc in range(number_of_proc):
if iproc != main_node:
dummy = wrap_mpi_recv(iproc, MPI_COMM_WORLD)
full_data[np.where(dummy>-1)] = dummy[np.where(dummy>-1)]
del dummy
mpi_barrier(MPI_COMM_WORLD)
full_data = wrap_mpi_bcast(full_data, main_node, MPI_COMM_WORLD)
# find the CPU with heaviest load
minindx = np.argsort(full_data, 0)
heavy_load_myid = minindx[-1][1]
total_mem = sum(full_data)
if myid == main_node:
if current_window == 0:
log_main.add("Nx: current image size = %d. Decimated by %f from %d"%(nx, current_decimate, nnxo))
else:
log_main.add("Nx: current image size = %d. Windowed to %d, and decimated by %f from %d"%(nx, current_window, current_decimate, nnxo))
log_main.add("Nproj: number of particle images.")
log_main.add("Navg: number of 2D average images.")
log_main.add("Nvar: number of 2D variance images.")
log_main.add("Img_per_grp: user defined image per group for averaging = %d"%img_per_grp)
log_main.add("Overhead: total python overhead memory consumption = %f"%overhead_loading)
log_main.add("Total memory) = 4.0*nx^2*(nproj + navg +nvar+ img_per_grp)/1.0e9 + overhead: %12.3f [GB]"%\
(total_mem[1] + overhead_loading))
del full_data
mpi_barrier(MPI_COMM_WORLD)
if myid == heavy_load_myid:
log_main.add("Begin reading and preprocessing images on processor. Wait... ")
ttt = time()
#imgdata = EMData.read_images(stack, all_proj)
imgdata = [ None for im in range(len(all_proj))]
for index_of_proj in range(len(all_proj)):
#image = get_im(stack, all_proj[index_of_proj])
if( current_window > 0): imgdata[index_of_proj] = fdecimate(window2d(get_im(stack, all_proj[index_of_proj]),current_window,current_window), nx, ny)
else: imgdata[index_of_proj] = fdecimate(get_im(stack, all_proj[index_of_proj]), nx, ny)
if (current_decimate> 0.0 and options.CTF):
ctf = imgdata[index_of_proj].get_attr("ctf")
ctf.apix = ctf.apix/current_decimate
imgdata[index_of_proj].set_attr("ctf", ctf)
if myid == heavy_load_myid and index_of_proj%100 == 0:
log_main.add(" ...... %6.2f%% "%(index_of_proj/float(len(all_proj))*100.))
mpi_barrier(MPI_COMM_WORLD)
if myid == heavy_load_myid:
log_main.add("All_proj preprocessing cost %7.2f m"%((time()-ttt)/60.))
log_main.add("Wait untill reading on all CPUs done...")
'''
imgdata2 = EMData.read_images(stack, range(img_begin, img_end))
if options.fl > 0.0:
for k in xrange(len(imgdata2)):
imgdata2[k] = filt_tanl(imgdata2[k], options.fl, options.aa)
if options.CTF:
vol = recons3d_4nn_ctf_MPI(myid, imgdata2, 1.0, symmetry=options.sym, npad=options.npad, xysize=-1, zsize=-1)
else:
vol = recons3d_4nn_MPI(myid, imgdata2, symmetry=options.sym, npad=options.npad, xysize=-1, zsize=-1)
if myid == main_node:
vol.write_image("vol_ctf.hdf")
print_msg("Writing to the disk volume reconstructed from averages as : %s\n"%("vol_ctf.hdf"))
del vol, imgdata2
mpi_barrier(MPI_COMM_WORLD)
'''
from applications import prepare_2d_forPCA
from utilities import model_blank
from EMAN2 import Transform
if not options.no_norm:
mask = model_circle(nx/2-2, nx, nx)
if options.CTF:
from utilities import pad
from filter import filt_ctf
from filter import filt_tanl
if myid == heavy_load_myid:
log_main.add("Start computing 2D aveList and varList. Wait...")
ttt = time()
inner=nx//2-4
outer=inner+2
xform_proj_for_2D = [ None for i in range(len(proj_list))]
for i in range(len(proj_list)):
ki = proj_angles[proj_list[i][0]][3]
if ki >= symbaselen: continue
mi = index[ki]
dpar = Util.get_transform_params(imgdata[mi], "xform.projection", "spider")
phiM, thetaM, psiM, s2xM, s2yM = dpar["phi"],dpar["theta"],dpar["psi"],-dpar["tx"]*current_decimate,-dpar["ty"]*current_decimate
grp_imgdata = []
for j in range(img_per_grp):
mj = index[proj_angles[proj_list[i][j]][3]]
cpar = Util.get_transform_params(imgdata[mj], "xform.projection", "spider")
alpha, sx, sy, mirror = params_3D_2D_NEW(cpar["phi"], cpar["theta"],cpar["psi"], -cpar["tx"]*current_decimate, -cpar["ty"]*current_decimate, mirror_list[i][j])
if thetaM <= 90:
if mirror == 0: alpha, sx, sy, scale = compose_transform2(alpha, sx, sy, 1.0, phiM - cpar["phi"], 0.0, 0.0, 1.0)
else: alpha, sx, sy, scale = compose_transform2(alpha, sx, sy, 1.0, 180-(phiM - cpar["phi"]), 0.0, 0.0, 1.0)
else:
if mirror == 0: alpha, sx, sy, scale = compose_transform2(alpha, sx, sy, 1.0, -(phiM- cpar["phi"]), 0.0, 0.0, 1.0)
else: alpha, sx, sy, scale = compose_transform2(alpha, sx, sy, 1.0, -(180-(phiM - cpar["phi"])), 0.0, 0.0, 1.0)
imgdata[mj].set_attr("xform.align2d", Transform({"type":"2D","alpha":alpha,"tx":sx,"ty":sy,"mirror":mirror,"scale":1.0}))
grp_imgdata.append(imgdata[mj])
if not options.no_norm:
for k in range(img_per_grp):
ave, std, minn, maxx = Util.infomask(grp_imgdata[k], mask, False)
grp_imgdata[k] -= ave
grp_imgdata[k] /= std
if options.fl > 0.0:
for k in range(img_per_grp):
grp_imgdata[k] = filt_tanl(grp_imgdata[k], options.fl, options.aa)
# Because of background issues, only linear option works.
if options.CTF: ave, var = aves_wiener(grp_imgdata, SNR = 1.0e5, interpolation_method = "linear")
else: ave, var = ave_var(grp_imgdata)
# Switch to std dev
# threshold is not really needed,it is just in case due to numerical accuracy something turns out negative.
var = square_root(threshold(var))
set_params_proj(ave, [phiM, thetaM, 0.0, 0.0, 0.0])
set_params_proj(var, [phiM, thetaM, 0.0, 0.0, 0.0])
aveList.append(ave)
varList.append(var)
xform_proj_for_2D[i] = [phiM, thetaM, 0.0, 0.0, 0.0]
'''
if nvec > 0:
eig = pca(input_stacks=grp_imgdata, subavg="", mask_radius=radiuspca, nvec=nvec, incore=True, shuffle=False, genbuf=True)
for k in range(nvec):
set_params_proj(eig[k], [phiM, thetaM, 0.0, 0.0, 0.0])
eigList[k].append(eig[k])
"""
if myid == 0 and i == 0:
for k in xrange(nvec):
eig[k].write_image("eig.hdf", k)
"""
'''
if (myid == heavy_load_myid) and (i%100 == 0):
log_main.add(" ......%6.2f%% "%(i/float(len(proj_list))*100.))
del imgdata, grp_imgdata, cpar, dpar, all_proj, proj_angles, index
if not options.no_norm: del mask
if myid == main_node: del tab
# At this point, all averages and variances are computed
mpi_barrier(MPI_COMM_WORLD)
if (myid == heavy_load_myid):
log_main.add("Computing aveList and varList took %12.1f [m]"%((time()-ttt)/60.))
nproj = len(xform_proj_for_2D)
nproj = mpi_reduce(nproj, 1, MPI_INT, MPI_SUM, main_node, MPI_COMM_WORLD)
if myid == main_node:
txform_proj = [ None for i in range(nproj)]
txform_proj[0:len(xform_proj_for_2D)] = xform_proj_for_2D[:]
nc = len(xform_proj_for_2D)
else:
wrap_mpi_send(xform_proj_for_2D, main_node, MPI_COMM_WORLD)
if (myid == main_node):
for iproc in range(1, number_of_proc):
dummy = wrap_mpi_recv(iproc, MPI_COMM_WORLD)
for im in range(len(dummy)):
txform_proj[nc] = dummy[im]
nc +=1
write_text_row(txform_proj, os.path.join(current_output_dir, "params.txt"))
del txform_proj
del xform_proj_for_2D
mpi_barrier(MPI_COMM_WORLD)
if options.ave2D:
from fundamentals import fpol
from applications import header
if myid == main_node:
log_main.add("Compute ave2D ... ")
km = 0
for i in range(number_of_proc):
if i == main_node :
for im in range(len(aveList)):
aveList[im].write_image(os.path.join(current_output_dir, options.ave2D), km)
km += 1
else:
nl = mpi_recv(1, MPI_INT, i, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
nl = int(nl[0])
for im in range(nl):
ave = recv_EMData(i, im+i+70000)
"""
nm = mpi_recv(1, MPI_INT, i, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
nm = int(nm[0])
members = mpi_recv(nm, MPI_INT, i, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
ave.set_attr('members', map(int, members))
members = mpi_recv(nm, MPI_FLOAT, i, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
ave.set_attr('pix_err', map(float, members))
members = mpi_recv(3, MPI_FLOAT, i, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
ave.set_attr('refprojdir', map(float, members))
"""
tmpvol=fpol(ave, nx, nx,1)
tmpvol.write_image(os.path.join(current_output_dir, options.ave2D), km)
km += 1
else:
mpi_send(len(aveList), 1, MPI_INT, main_node, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
for im in range(len(aveList)):
send_EMData(aveList[im], main_node,im+myid+70000)
"""
members = aveList[im].get_attr('members')
mpi_send(len(members), 1, MPI_INT, main_node, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
mpi_send(members, len(members), MPI_INT, main_node, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
members = aveList[im].get_attr('pix_err')
mpi_send(members, len(members), MPI_FLOAT, main_node, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
try:
members = aveList[im].get_attr('refprojdir')
mpi_send(members, 3, MPI_FLOAT, main_node, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
except:
mpi_send([-999.0,-999.0,-999.0], 3, MPI_FLOAT, main_node, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
"""
if myid == main_node:
header(os.path.join(current_output_dir, options.ave2D), params='xform.projection', fimport = os.path.join(current_output_dir, "params.txt"))
mpi_barrier(MPI_COMM_WORLD)
if options.ave3D:
from fundamentals import fpol
t5 = time()
if myid == main_node: log_main.add("Reconstruct ave3D ... ")
ave3D = recons3d_4nn_MPI(myid, aveList, symmetry=options.sym, npad=options.npad)
bcast_EMData_to_all(ave3D, myid)
if myid == main_node:
if current_decimate != 1.0: ave3D = resample(ave3D, 1./current_decimate)
ave3D = fpol(ave3D, nnxo, nnxo, nnxo) # always to the orignal image size
set_pixel_size(ave3D, 1.0)
ave3D.write_image(os.path.join(current_output_dir, options.ave3D))
log_main.add("Ave3D reconstruction took %12.1f [m]"%((time()-t5)/60.0))
log_main.add("%-70s: %s\n"%("The reconstructed ave3D is saved as ", options.ave3D))
mpi_barrier(MPI_COMM_WORLD)
del ave, var, proj_list, stack, alpha, sx, sy, mirror, aveList
'''
if nvec > 0:
for k in range(nvec):
if myid == main_node:log_main.add("Reconstruction eigenvolumes", k)
cont = True
ITER = 0
mask2d = model_circle(radiuspca, nx, nx)
while cont:
#print "On node %d, iteration %d"%(myid, ITER)
eig3D = recons3d_4nn_MPI(myid, eigList[k], symmetry=options.sym, npad=options.npad)
bcast_EMData_to_all(eig3D, myid, main_node)
if options.fl > 0.0:
eig3D = filt_tanl(eig3D, options.fl, options.aa)
if myid == main_node:
eig3D.write_image(os.path.join(options.outpout_dir, "eig3d_%03d.hdf"%(k, ITER)))
Util.mul_img( eig3D, model_circle(radiuspca, nx, nx, nx) )
eig3Df, kb = prep_vol(eig3D)
del eig3D
cont = False
icont = 0
for l in range(len(eigList[k])):
phi, theta, psi, s2x, s2y = get_params_proj(eigList[k][l])
proj = prgs(eig3Df, kb, [phi, theta, psi, s2x, s2y])
cl = ccc(proj, eigList[k][l], mask2d)
if cl < 0.0:
icont += 1
cont = True
eigList[k][l] *= -1.0
u = int(cont)
u = mpi_reduce([u], 1, MPI_INT, MPI_MAX, main_node, MPI_COMM_WORLD)
icont = mpi_reduce([icont], 1, MPI_INT, MPI_SUM, main_node, MPI_COMM_WORLD)
if myid == main_node:
u = int(u[0])
log_main.add(" Eigenvector: ",k," number changed ",int(icont[0]))
else: u = 0
u = bcast_number_to_all(u, main_node)
cont = bool(u)
ITER += 1
del eig3Df, kb
mpi_barrier(MPI_COMM_WORLD)
del eigList, mask2d
'''
if options.ave3D: del ave3D
if options.var2D:
from fundamentals import fpol
from applications import header
if myid == main_node:
log_main.add("Compute var2D...")
km = 0
for i in range(number_of_proc):
if i == main_node :
for im in range(len(varList)):
tmpvol=fpol(varList[im], nx, nx,1)
tmpvol.write_image(os.path.join(current_output_dir, options.var2D), km)
km += 1
else:
nl = mpi_recv(1, MPI_INT, i, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
nl = int(nl[0])
for im in range(nl):
ave = recv_EMData(i, im+i+70000)
tmpvol=fpol(ave, nx, nx,1)
tmpvol.write_image(os.path.join(current_output_dir, options.var2D), km)
km += 1
else:
mpi_send(len(varList), 1, MPI_INT, main_node, SPARX_MPI_TAG_UNIVERSAL, MPI_COMM_WORLD)
for im in range(len(varList)):
send_EMData(varList[im], main_node, im+myid+70000)# What with the attributes??
mpi_barrier(MPI_COMM_WORLD)
if myid == main_node:
from applications import header
header(os.path.join(current_output_dir, options.var2D), params = 'xform.projection',fimport = os.path.join(current_output_dir, "params.txt"))
mpi_barrier(MPI_COMM_WORLD)
if options.var3D:
if myid == main_node: log_main.add("Reconstruct var3D ...")
t6 = time()
# radiusvar = options.radius
# if( radiusvar < 0 ): radiusvar = nx//2 -3
res = recons3d_4nn_MPI(myid, varList, symmetry = options.sym, npad=options.npad)
#res = recons3d_em_MPI(varList, vol_stack, options.iter, radiusvar, options.abs, True, options.sym, options.squ)
if myid == main_node:
from fundamentals import fpol
if current_decimate != 1.0: res = resample(res, 1./current_decimate)
res = fpol(res, nnxo, nnxo, nnxo)
set_pixel_size(res, 1.0)
res.write_image(os.path.join(current_output_dir, options.var3D))
log_main.add("%-70s: %s\n"%("The reconstructed var3D is saved as ", options.var3D))
log_main.add("Var3D reconstruction took %f12.1 [m]"%((time()-t6)/60.0))
log_main.add("Total computation time %f12.1 [m]"%((time()-t0)/60.0))
log_main.add("sx3dvariability finishes")
from mpi import mpi_finalize
mpi_finalize()
if RUNNING_UNDER_MPI: global_def.MPI = False
global_def.BATCH = False