scaler.rejected_fractions = scaler.n_rejected[irej]/scaler.n_obs[irej]
# ---------- End of new code ----------------
if work_params.raw_data.sdfac_refine:
scaler.scale_errors()
if work_params.raw_data.errors_from_sample_residuals:
scaler.errors_from_residuals()
miller_set_avg = miller_set.customized_copy(
unit_cell=work_params.target_unit_cell)
table1 = show_overall_observations(
obs=miller_set_avg,
redundancy=scaler.completeness,
summed_wt_I=scaler.summed_wt_I,
summed_weight=scaler.summed_weight,
ISIGI=scaler.ISIGI,
n_bins=work_params.output.n_bins,
title="Statistics for all reflections",
out=out,
work_params=work_params)
print >> out, ""
if work_params.scaling.algorithm == 'mark0':
n_refl, corr = scaler.get_overall_correlation(sum_I)
else:
n_refl, corr = ((scaler.completeness > 0).count(True), 0)
print >> out, "\n"
table2 = show_overall_observations(
obs=miller_set_avg,
redundancy=scaler.summed_N,
summed_wt_I=scaler.summed_wt_I,
summed_weight=scaler.summed_weight,
def run(args):
phil = iotbx.phil.process_command_line(args=args,
master_string=master_phil).show()
work_params = phil.work.extract()
from xfel.merging.phil_validation import application, samosa
application(work_params)
samosa(work_params)
if ("--help" in args):
libtbx.phil.parse(master_phil.show())
return
if ((work_params.d_min is None) or (work_params.data is None)):
command_name = os.environ["LIBTBX_DISPATCHER_NAME"]
raise Usage(command_name + " "
"d_min=4.0 "
"data=~/scratch/r0220/006/strong/ "
"model=3bz1_3bz2_core.pdb")
if ((work_params.rescale_with_average_cell)
and (not work_params.set_average_unit_cell)):
raise Usage(
"If rescale_with_average_cell=True, you must also specify " +
"set_average_unit_cell=True.")
if work_params.raw_data.sdfac_auto and work_params.raw_data.sdfac_refine:
raise Usage("Cannot specify both sdfac_auto and sdfac_refine")
# Read Nat's reference model from an MTZ file. XXX The observation
# type is given as F, not I--should they be squared? Check with Nat!
log = open("%s.log" % work_params.output.prefix, "w")
out = multi_out()
out.register("log", log, atexit_send_to=None)
out.register("stdout", sys.stdout)
print >> out, "I model"
if work_params.model is not None:
from xfel.merging.general_fcalc import run
i_model = run(work_params)
work_params.target_unit_cell = i_model.unit_cell()
work_params.target_space_group = i_model.space_group_info()
i_model.show_summary()
else:
i_model = None
print >> out, "Target unit cell and space group:"
print >> out, " ", work_params.target_unit_cell
print >> out, " ", work_params.target_space_group
miller_set, i_model = consistent_set_and_model(work_params, i_model)
frame_files = get_observations(work_params)
scaler = scaling_manager(miller_set=miller_set,
i_model=i_model,
params=work_params,
log=out)
scaler.scale_all(frame_files)
if scaler.n_accepted == 0:
return None
scaler.show_unit_cell_histograms()
if (work_params.rescale_with_average_cell):
average_cell_abc = scaler.uc_values.get_average_cell_dimensions()
average_cell = uctbx.unit_cell(
list(average_cell_abc) +
list(work_params.target_unit_cell.parameters()[3:]))
work_params.target_unit_cell = average_cell
print >> out, ""
print >> out, "#" * 80
print >> out, "RESCALING WITH NEW TARGET CELL"
print >> out, " average cell: %g %g %g %g %g %g" % \
work_params.target_unit_cell.parameters()
print >> out, ""
scaler.reset()
scaler.scale_all(frame_files)
scaler.show_unit_cell_histograms()
if False: #(work_params.output.show_plots) :
try:
plot_overall_completeness(completeness)
except Exception as e:
print "ERROR: can't show plots"
print " %s" % str(e)
print >> out, "\n"
# Sum the observations of I and I/sig(I) for each reflection.
sum_I = flex.double(miller_set.size(), 0.)
sum_I_SIGI = flex.double(miller_set.size(), 0.)
for i in xrange(miller_set.size()):
index = miller_set.indices()[i]
if index in scaler.ISIGI:
for t in scaler.ISIGI[index]:
sum_I[i] += t[0]
sum_I_SIGI[i] += t[1]
miller_set_avg = miller_set.customized_copy(
unit_cell=work_params.target_unit_cell)
table1 = show_overall_observations(obs=miller_set_avg,
redundancy=scaler.completeness,
summed_wt_I=scaler.summed_wt_I,
summed_weight=scaler.summed_weight,
ISIGI=scaler.ISIGI,
n_bins=work_params.output.n_bins,
title="Statistics for all reflections",
out=out,
work_params=work_params)
print >> out, ""
n_refl, corr = ((scaler.completeness > 0).count(True), 0)
print >> out, "\n"
table2 = show_overall_observations(
obs=miller_set_avg,
redundancy=scaler.summed_N,
summed_wt_I=scaler.summed_wt_I,
summed_weight=scaler.summed_weight,
ISIGI=scaler.ISIGI,
n_bins=work_params.output.n_bins,
title="Statistics for reflections where I > 0",
out=out,
work_params=work_params)
#from libtbx import easy_pickle
#easy_pickle.dump(file_name="stats.pickle", obj=stats)
#stats.report(plot=work_params.plot)
#miller_counts = miller_set_p1.array(data=stats.counts.as_double()).select(
# stats.counts != 0)
#miller_counts.as_mtz_dataset(column_root_label="NOBS").mtz_object().write(
# file_name="nobs.mtz")
if work_params.data_subsubsets.subsubset is not None and work_params.data_subsubsets.subsubset_total is not None:
easy_pickle.dump(
"scaler_%d.pickle" % work_params.data_subsubsets.subsubset, scaler)
print >> out, ""
mtz_file, miller_array = scaler.finalize_and_save_data()
#table_pickle_file = "%s_graphs.pkl" % work_params.output.prefix
#easy_pickle.dump(table_pickle_file, [table1, table2])
loggraph_file = os.path.abspath("%s_graphs.log" %
work_params.output.prefix)
f = open(loggraph_file, "w")
f.write(table1.format_loggraph())
f.write("\n")
f.write(table2.format_loggraph())
f.close()
result = scaling_result(miller_array=miller_array,
plots=scaler.get_plot_statistics(),
mtz_file=mtz_file,
loggraph_file=loggraph_file,
obs_table=table1,
all_obs_table=table2,
n_reflections=n_refl,
overall_correlation=corr)
easy_pickle.dump("%s.pkl" % work_params.output.prefix, result)
return result
irej] / scaler.n_obs[irej]
# ---------- End of new code ----------------
if work_params.raw_data.sdfac_refine:
scaler.scale_errors()
if work_params.raw_data.errors_from_sample_residuals:
scaler.errors_from_residuals()
miller_set_avg = miller_set.customized_copy(
unit_cell=work_params.target_unit_cell)
table1 = show_overall_observations(
obs=miller_set_avg,
redundancy=scaler.completeness,
summed_wt_I=scaler.summed_wt_I,
summed_weight=scaler.summed_weight,
ISIGI=scaler.ISIGI,
n_bins=work_params.output.n_bins,
title="Statistics for all reflections",
out=out,
work_params=work_params)
print >> out, ""
if work_params.scaling.algorithm == 'mark0':
n_refl, corr = scaler.get_overall_correlation(sum_I)
else:
n_refl, corr = ((scaler.completeness > 0).count(True), 0)
print >> out, "\n"
table2 = show_overall_observations(
obs=miller_set_avg,
redundancy=scaler.summed_N,
summed_wt_I=scaler.summed_wt_I,
summed_weight=scaler.summed_weight,
def run(args):
phil = iotbx.phil.process_command_line(args=args,
master_string=master_phil).show()
work_params = phil.work.extract()
from xfel.merging.phil_validation import application
application(work_params)
if ("--help" in args):
libtbx.phil.parse(master_phil.show())
return
if ((work_params.d_min is None) or (work_params.data is None)
or ((work_params.model is None)
and work_params.scaling.algorithm != "mark1")):
raise Usage("cxi.merge "
"d_min=4.0 "
"data=~/scratch/r0220/006/strong/ "
"model=3bz1_3bz2_core.pdb")
if ((work_params.rescale_with_average_cell)
and (not work_params.set_average_unit_cell)):
raise Usage(
"If rescale_with_average_cell=True, you must also specify " +
"set_average_unit_cell=True.")
if work_params.raw_data.sdfac_auto and work_params.raw_data.sdfac_refine:
raise Usage("Cannot specify both sdfac_auto and sdfac_refine")
if not work_params.include_negatives_fix_27May2018:
work_params.include_negatives = False # use old behavior
log = open(
"%s_%s.log" %
(work_params.output.prefix, work_params.scaling.algorithm), "w")
out = multi_out()
out.register("log", log, atexit_send_to=None)
out.register("stdout", sys.stdout)
# Verify that the externally supplied isomorphous reference, if
# present, defines a suitable column of intensities, and exit with
# error if it does not. Then warn if it is necessary to generate
# Bijvoet mates. Failure to catch these issues here would lead to
# possibly obscure problems in cxi/cxi_cc.py later on.
try:
data_SR = mtz.object(work_params.scaling.mtz_file)
except RuntimeError:
pass
else:
array_SR = None
obs_labels = []
for array in data_SR.as_miller_arrays():
this_label = array.info().label_string().lower()
if array.observation_type() is not None:
obs_labels.append(this_label.split(',')[0])
if this_label.find('fobs') >= 0:
array_SR = array.as_intensity_array()
break
if this_label.find('imean') >= 0:
array_SR = array.as_intensity_array()
break
if this_label.find(work_params.scaling.mtz_column_F) == 0:
array_SR = array.as_intensity_array()
break
if array_SR is None:
known_labels = ['fobs', 'imean', work_params.scaling.mtz_column_F]
raise Usage(work_params.scaling.mtz_file +
" does not contain any observations labelled [" +
", ".join(known_labels) +
"]. Please set scaling.mtz_column_F to one of [" +
",".join(obs_labels) + "].")
elif not work_params.merge_anomalous and not array_SR.anomalous_flag():
print("Warning: Preserving anomalous contributors, but %s " \
"has anomalous contributors merged. Generating identical Bijvoet " \
"mates." % work_params.scaling.mtz_file, file=out)
# Read Nat's reference model from an MTZ file. XXX The observation
# type is given as F, not I--should they be squared? Check with Nat!
print("I model", file=out)
if work_params.model is not None:
from xfel.merging.general_fcalc import run
i_model = run(work_params)
work_params.target_unit_cell = i_model.unit_cell()
work_params.target_space_group = i_model.space_group_info()
i_model.show_summary()
else:
i_model = None
print("Target unit cell and space group:", file=out)
print(" ", work_params.target_unit_cell, file=out)
print(" ", work_params.target_space_group, file=out)
miller_set, i_model = consistent_set_and_model(work_params, i_model)
# ---- Augment this code with any special procedures for x scaling
scaler = xscaling_manager(miller_set=miller_set,
i_model=i_model,
params=work_params,
log=out)
scaler.scale_all()
if scaler.n_accepted == 0:
return None
# --- End of x scaling
scaler.uc_values = unit_cell_distribution()
for icell in range(len(scaler.frames["unit_cell"])):
if scaler.params.model is None:
scaler.uc_values.add_cell(
unit_cell=scaler.frames["unit_cell"][icell])
else:
scaler.uc_values.add_cell(
unit_cell=scaler.frames["unit_cell"][icell],
rejected=(scaler.frames["cc"][icell] < scaler.params.min_corr))
scaler.show_unit_cell_histograms()
if (work_params.rescale_with_average_cell):
average_cell_abc = scaler.uc_values.get_average_cell_dimensions()
average_cell = uctbx.unit_cell(
list(average_cell_abc) +
list(work_params.target_unit_cell.parameters()[3:]))
work_params.target_unit_cell = average_cell
print("", file=out)
print("#" * 80, file=out)
print("RESCALING WITH NEW TARGET CELL", file=out)
print(" average cell: %g %g %g %g %g %g" % \
work_params.target_unit_cell.parameters(), file=out)
print("", file=out)
scaler.reset()
scaler = xscaling_manager(miller_set=miller_set,
i_model=i_model,
params=work_params,
log=out)
scaler.scale_all()
scaler.uc_values = unit_cell_distribution()
for icell in range(len(scaler.frames["unit_cell"])):
if scaler.params.model is None:
scaler.uc_values.add_cell(
unit_cell=scaler.frames["unit_cell"][icell])
else:
scaler.uc_values.add_cell(
unit_cell=scaler.frames["unit_cell"][icell],
rejected=(scaler.frames["cc"][icell] <
scaler.params.min_corr))
scaler.show_unit_cell_histograms()
if False: #(work_params.output.show_plots) :
try:
plot_overall_completeness(completeness)
except Exception as e:
print("ERROR: can't show plots")
print(" %s" % str(e))
print("\n", file=out)
reserve_prefix = work_params.output.prefix
for data_subset in [1, 2, 0]:
work_params.data_subset = data_subset
work_params.output.prefix = "%s_s%1d_%s" % (
reserve_prefix, data_subset, work_params.scaling.algorithm)
if work_params.data_subset == 0:
scaler.frames["data_subset"] = flex.bool(
scaler.frames["frame_id"].size(), True)
elif work_params.data_subset == 1:
scaler.frames["data_subset"] = scaler.frames["odd_numbered"]
elif work_params.data_subset == 2:
scaler.frames["data_subset"] = scaler.frames[
"odd_numbered"] == False
# --------- New code ------------------
#sanity check
for mod, obs in zip(miller_set.indices(),
scaler.millers["merged_asu_hkl"]):
if mod != obs:
raise Exception(
"miller index lists inconsistent--check d_min are equal for merge and xmerge scripts"
)
assert mod == obs
"""Sum the observations of I and I/sig(I) for each reflection.
sum_I = flex.double(i_model.size(), 0.)
sum_I_SIGI = flex.double(i_model.size(), 0.)
scaler.completeness = flex.int(i_model.size(), 0)
scaler.summed_N = flex.int(i_model.size(), 0)
scaler.summed_wt_I = flex.double(i_model.size(), 0.)
scaler.summed_weight = flex.double(i_model.size(), 0.)
scaler.n_rejected = flex.double(scaler.frames["frame_id"].size(), 0.)
scaler.n_obs = flex.double(scaler.frames["frame_id"].size(), 0.)
scaler.d_min_values = flex.double(scaler.frames["frame_id"].size(), 0.)
scaler.ISIGI = {}"""
from xfel import scaling_results, get_scaling_results, get_isigi_dict
results = scaling_results(scaler._observations, scaler._frames,
scaler.millers["merged_asu_hkl"],
scaler.frames["data_subset"],
work_params.include_negatives)
results.__getattribute__(work_params.scaling.algorithm)(
scaler.params.min_corr, scaler.params.target_unit_cell)
sum_I, sum_I_SIGI, \
scaler.completeness, scaler.summed_N, \
scaler.summed_wt_I, scaler.summed_weight, scaler.n_rejected, scaler.n_obs, \
scaler.d_min_values, hkl_ids, i_sigi_list = get_scaling_results(results)
scaler.ISIGI = get_isigi_dict(results)
if work_params.merging.refine_G_Imodel:
from xfel.cxi.merging.refine import find_scale
my_find_scale = find_scale(scaler, work_params)
sum_I, sum_I_SIGI, \
scaler.completeness, scaler.summed_N, \
scaler.summed_wt_I, scaler.summed_weight, scaler.n_rejected, \
scaler.n_obs, scaler.d_min_values, hkl_ids, i_sigi_list \
= my_find_scale.get_scaling_results(results, scaler)
scaler.ISIGI = get_isigi_dict(results)
scaler.wavelength = scaler.frames["wavelength"]
scaler.corr_values = scaler.frames["cc"]
scaler.rejected_fractions = flex.double(
scaler.frames["frame_id"].size(), 0.)
for irej in range(len(scaler.rejected_fractions)):
if scaler.n_obs[irej] > 0:
scaler.rejected_fractions = scaler.n_rejected[
irej] / scaler.n_obs[irej]
# ---------- End of new code ----------------
if work_params.raw_data.sdfac_refine or work_params.raw_data.errors_from_sample_residuals:
if work_params.raw_data.sdfac_refine:
if work_params.raw_data.error_models.sdfac_refine.minimizer == 'simplex':
from xfel.merging.algorithms.error_model.sdfac_refine import sdfac_refine as error_modeler
elif work_params.raw_data.error_models.sdfac_refine.minimizer == 'lbfgs':
from xfel.merging.algorithms.error_model.sdfac_refine_lbfgs import sdfac_refine_refltable_lbfgs as error_modeler
elif self.params.raw_data.error_models.sdfac_refine.minimizer == 'LevMar':
from xfel.merging.algorithms.error_model.sdfac_refine_levmar import sdfac_refine_refltable_levmar as error_modeler
if work_params.raw_data.errors_from_sample_residuals:
from xfel.merging.algorithms.error_model.errors_from_residuals import errors_from_residuals as error_modeler
error_modeler(scaler).adjust_errors()
if work_params.raw_data.reduced_chi_squared_correction:
from xfel.merging.algorithms.error_model.reduced_chi_squared import reduced_chi_squared
reduced_chi_squared(scaler).compute()
miller_set_avg = miller_set.customized_copy(
unit_cell=work_params.target_unit_cell)
table1 = show_overall_observations(
obs=miller_set_avg,
redundancy=scaler.completeness,
redundancy_to_edge=None,
summed_wt_I=scaler.summed_wt_I,
summed_weight=scaler.summed_weight,
ISIGI=scaler.ISIGI,
n_bins=work_params.output.n_bins,
title="Statistics for all reflections",
out=out,
work_params=work_params)
if table1 is None:
raise Exception("table could not be constructed")
print("", file=out)
if work_params.scaling.algorithm == 'mark0':
n_refl, corr = scaler.get_overall_correlation(sum_I)
else:
n_refl, corr = ((scaler.completeness > 0).count(True), 0)
print("\n", file=out)
table2 = show_overall_observations(
obs=miller_set_avg,
redundancy=scaler.summed_N,
redundancy_to_edge=None,
summed_wt_I=scaler.summed_wt_I,
summed_weight=scaler.summed_weight,
ISIGI=scaler.ISIGI,
n_bins=work_params.output.n_bins,
title="Statistics for reflections where I > 0",
out=out,
work_params=work_params)
if table2 is None:
raise Exception("table could not be constructed")
print("", file=out)
mtz_file, miller_array = scaler.finalize_and_save_data()
loggraph_file = os.path.abspath("%s_graphs.log" %
work_params.output.prefix)
f = open(loggraph_file, "w")
f.write(table1.format_loggraph())
f.write("\n")
f.write(table2.format_loggraph())
f.close()
result = scaling_result(miller_array=miller_array,
plots=scaler.get_plot_statistics(),
mtz_file=mtz_file,
loggraph_file=loggraph_file,
obs_table=table1,
all_obs_table=table2,
n_reflections=n_refl,
overall_correlation=corr)
easy_pickle.dump("%s.pkl" % work_params.output.prefix, result)
work_params.output.prefix = reserve_prefix
# Output table with number of images contribution reflections per
# resolution bin.
from libtbx import table_utils
miller_set_avg.setup_binner(d_max=100000,
d_min=work_params.d_min,
n_bins=work_params.output.n_bins)
table_data = [["Bin", "Resolution Range", "# images", "%accept"]]
if work_params.model is None:
appropriate_min_corr = -1.1 # lowest possible c.c.
else:
appropriate_min_corr = work_params.min_corr
n_frames = (scaler.frames['cc'] > appropriate_min_corr).count(True)
iselect = 1
while iselect < work_params.output.n_bins:
col_count1 = results.count_frames(
appropriate_min_corr,
miller_set_avg.binner().selection(iselect))
print("colcount1", col_count1)
if col_count1 > 0: break
iselect += 1
if col_count1 == 0: raise Exception("no reflections in any bins")
for i_bin in miller_set_avg.binner().range_used():
col_count = '%8d' % results.count_frames(
appropriate_min_corr,
miller_set_avg.binner().selection(i_bin))
col_legend = '%-13s' % miller_set_avg.binner().bin_legend(
i_bin=i_bin,
show_bin_number=False,
show_bin_range=False,
show_d_range=True,
show_counts=False)
xpercent = results.count_frames(
appropriate_min_corr,
miller_set_avg.binner().selection(i_bin)) / float(n_frames)
percent = '%5.2f' % (100. * xpercent)
table_data.append(['%3d' % i_bin, col_legend, col_count, percent])
table_data.append([""] * len(table_data[0]))
table_data.append(["All", "", '%8d' % n_frames])
print(file=out)
print(table_utils.format(table_data,
has_header=1,
justify='center',
delim=' '),
file=out)
reindexing_ops = {
"h,k,l": 0
} # get a list of all reindexing ops for this dataset
if work_params.merging.reverse_lookup is not None:
for key in scaler.reverse_lookup:
if reindexing_ops.get(scaler.reverse_lookup[key], None) is None:
reindexing_ops[scaler.reverse_lookup[key]] = 0
reindexing_ops[scaler.reverse_lookup[key]] += 1
from xfel.cxi.cxi_cc import run_cc
for key in reindexing_ops.keys():
run_cc(work_params, reindexing_op=key, output=out)
if isinstance(scaler.ISIGI, dict):
from xfel.merging import isigi_dict_to_reflection_table
refls = isigi_dict_to_reflection_table(scaler.miller_set.indices(),
scaler.ISIGI)
else:
refls = scaler.ISIGI
easy_pickle.dump("%s.refl" % work_params.output.prefix, refls)
return result
def other(self, scaler):
out = self.out
work_params = self.work_params
miller_set = self.miller_set
if scaler.n_accepted == 0:
return None
scaler.show_unit_cell_histograms()
if (work_params.rescale_with_average_cell):
average_cell_abc = scaler.uc_values.get_average_cell_dimensions()
average_cell = uctbx.unit_cell(
list(average_cell_abc) +
list(work_params.target_unit_cell.parameters()[3:]))
work_params.target_unit_cell = average_cell
print >> out, ""
print >> out, "#" * 80
print >> out, "RESCALING WITH NEW TARGET CELL"
print >> out, " average cell: %g %g %g %g %g %g" % \
work_params.target_unit_cell.parameters()
print >> out, ""
assert False, "must do this step again with MPI"
scaler.reset()
scaler.scale_all(frame_files)
scaler.show_unit_cell_histograms()
print >> out, "\n"
# Sum the observations of I and I/sig(I) for each reflection.
sum_I = flex.double(miller_set.size(), 0.)
sum_I_SIGI = flex.double(miller_set.size(), 0.)
for i in xrange(miller_set.size()):
index = miller_set.indices()[i]
if index in scaler.ISIGI:
for t in scaler.ISIGI[index]:
sum_I[i] += t[0]
sum_I_SIGI[i] += t[1]
miller_set_avg = miller_set.customized_copy(
unit_cell=work_params.target_unit_cell)
table1 = cxi_merge.show_overall_observations(
obs=miller_set_avg,
redundancy=scaler.completeness,
redundancy_to_edge=scaler.completeness_predictions,
summed_wt_I=scaler.summed_wt_I,
summed_weight=scaler.summed_weight,
ISIGI=scaler.ISIGI,
n_bins=work_params.output.n_bins,
title="Statistics for all reflections",
out=out,
work_params=work_params)
print >> out, ""
if work_params.model is not None:
n_refl, corr = scaler.get_overall_correlation(sum_I)
else:
n_refl, corr = ((scaler.completeness > 0).count(True), 0)
print >> out, "\n"
table2 = cxi_merge.show_overall_observations(
obs=miller_set_avg,
redundancy=scaler.summed_N,
redundancy_to_edge=scaler.completeness_predictions,
summed_wt_I=scaler.summed_wt_I,
summed_weight=scaler.summed_weight,
ISIGI=scaler.ISIGI,
n_bins=work_params.output.n_bins,
title="Statistics for reflections where I > 0",
out=out,
work_params=work_params)
#from libtbx import easy_pickle
#easy_pickle.dump(file_name="stats.pickle", obj=stats)
#stats.report(plot=work_params.plot)
#miller_counts = miller_set_p1.array(data=stats.counts.as_double()).select(
# stats.counts != 0)
#miller_counts.as_mtz_dataset(column_root_label="NOBS").mtz_object().write(
# file_name="nobs.mtz")
if work_params.data_subsubsets.subsubset is not None and work_params.data_subsubsets.subsubset_total is not None:
easy_pickle.dump(
"scaler_%d.pickle" % work_params.data_subsubsets.subsubset,
scaler)
explanation = """
Explanation:
Completeness = # unique Miller indices present in data / # Miller indices theoretical in asymmetric unit
Asu. Multiplicity = # measurements / # Miller indices theoretical in asymmetric unit
Obs. Multiplicity = # measurements / # unique Miller indices present in data
Pred. Multiplicity = # predictions on all accepted images / # Miller indices theoretical in asymmetric unit"""
print >> out, explanation
mtz_file, miller_array = scaler.finalize_and_save_data()
#table_pickle_file = "%s_graphs.pkl" % work_params.output.prefix
#easy_pickle.dump(table_pickle_file, [table1, table2])
loggraph_file = os.path.abspath("%s_graphs.log" %
work_params.output.prefix)
f = open(loggraph_file, "w")
f.write(table1.format_loggraph())
f.write("\n")
f.write(table2.format_loggraph())
f.close()
result = cxi_merge.scaling_result(miller_array=miller_array,
plots=scaler.get_plot_statistics(),
mtz_file=mtz_file,
loggraph_file=loggraph_file,
obs_table=table1,
all_obs_table=table2,
n_reflections=n_refl,
overall_correlation=corr)
easy_pickle.dump("%s.pkl" % work_params.output.prefix, result)
return result
def finalize(self, scaler):
scaler.show_unit_cell_histograms()
if (self.params.rescale_with_average_cell) :
average_cell_abc = scaler.uc_values.get_average_cell_dimensions()
average_cell = uctbx.unit_cell(list(average_cell_abc) +
list(self.params.target_unit_cell.parameters()[3:]))
self.params.target_unit_cell = average_cell
print >> out, ""
print >> out, "#" * 80
print >> out, "RESCALING WITH NEW TARGET CELL"
print >> out, " average cell: %g %g %g %g %g %g" % \
self.params.target_unit_cell.parameters()
print >> out, ""
scaler.reset()
scaler.scale_all(frame_files)
scaler.show_unit_cell_histograms()
if False : #(self.params.output.show_plots) :
try :
plot_overall_completeness(completeness)
except Exception as e :
print "ERROR: can't show plots"
print " %s" % str(e)
print >> self.out, "\n"
sum_I, sum_I_SIGI = scaler.sum_intensities()
miller_set_avg = self.miller_set.customized_copy(
unit_cell=self.params.target_unit_cell)
table1 = show_overall_observations(
obs=miller_set_avg,
redundancy=scaler.completeness,
redundancy_to_edge=scaler.completeness_predictions,
summed_wt_I=scaler.summed_wt_I,
summed_weight=scaler.summed_weight,
ISIGI=scaler.ISIGI,
n_bins=self.params.output.n_bins,
title="Statistics for all reflections",
out=self.out,
work_params=self.params)
print >> self.out, ""
if self.params.model is not None:
n_refl, corr = scaler.get_overall_correlation(sum_I)
else:
n_refl, corr = ((scaler.completeness > 0).count(True), 0)
print >> self.out, "\n"
table2 = show_overall_observations(
obs=miller_set_avg,
redundancy=scaler.summed_N,
redundancy_to_edge=scaler.completeness_predictions,
summed_wt_I=scaler.summed_wt_I,
summed_weight=scaler.summed_weight,
ISIGI=scaler.ISIGI,
n_bins=self.params.output.n_bins,
title="Statistics for reflections where I > 0",
out=self.out,
work_params=self.params)
#from libtbx import easy_pickle
#easy_pickle.dump(file_name="stats.pickle", obj=stats)
#stats.report(plot=self.params.plot)
#miller_counts = miller_set_p1.array(data=stats.counts.as_double()).select(
# stats.counts != 0)
#miller_counts.as_mtz_dataset(column_root_label="NOBS").mtz_object().write(
# file_name="nobs.mtz")
if self.params.data_subsubsets.subsubset is not None and self.params.data_subsubsets.subsubset_total is not None:
easy_pickle.dump("scaler_%d.pickle"%self.params.data_subsubsets.subsubset, scaler)
explanation = """
Explanation:
Completeness = # unique Miller indices present in data / # Miller indices theoretical in asymmetric unit
Asu. Multiplicity = # measurements / # Miller indices theoretical in asymmetric unit
Obs. Multiplicity = # measurements / # unique Miller indices present in data
Pred. Multiplicity = # predictions on all accepted images / # Miller indices theoretical in asymmetric unit"""
print >> self.out, explanation
mtz_file, miller_array = scaler.finalize_and_save_data()
#table_pickle_file = "%s_graphs.pkl" % self.params.output.prefix
#easy_pickle.dump(table_pickle_file, [table1, table2])
loggraph_file = os.path.abspath("%s_graphs.log" % self.params.output.prefix)
f = open(loggraph_file, "w")
f.write(table1.format_loggraph())
f.write("\n")
f.write(table2.format_loggraph())
f.close()
result = scaling_result(
miller_array=miller_array,
plots=scaler.get_plot_statistics(),
mtz_file=mtz_file,
loggraph_file=loggraph_file,
obs_table=table1,
all_obs_table=table2,
n_reflections=n_refl,
overall_correlation=corr)
easy_pickle.dump("%s.pkl" % self.params.output.prefix, result)
return result
class Script(object):
'''A class for running the script.'''
def __init__(self, scaler_class):
# The script usage
import libtbx.load_env
self.usage = "usage: %s [options] [param.phil] " % libtbx.env.dispatcher_name
self.parser = None
self.scaler_class = scaler_class
def initialize(self):
'''Initialise the script.'''
from dials.util.options import OptionParser
from iotbx.phil import parse
phil_scope = parse(master_phil)
# Create the parser
self.parser = OptionParser(usage=self.usage,
phil=phil_scope,
epilog=help_message)
self.parser.add_option('--plots',
action='store_true',
default=False,
dest='show_plots',
help='Show some plots.')
# Parse the command line. quick_parse is required for MPI compatibility
params, options = self.parser.parse_args(show_diff_phil=True,
quick_parse=True)
self.params = params
self.options = options
def validate(self):
from xfel.merging.phil_validation import application
application(self.params)
if ((self.params.d_min is None) or (self.params.data is None)
or ((self.params.model is None)
and self.params.scaling.algorithm != "mark1")):
command_name = os.environ["LIBTBX_DISPATCHER_NAME"]
raise Usage(command_name + " "
"d_min=4.0 "
"data=~/scratch/r0220/006/strong/ "
"model=3bz1_3bz2_core.pdb")
if ((self.params.rescale_with_average_cell)
and (not self.params.set_average_unit_cell)):
raise Usage(
"If rescale_with_average_cell=True, you must also specify " +
"set_average_unit_cell=True.")
if [
self.params.raw_data.sdfac_auto,
self.params.raw_data.sdfac_refine,
self.params.raw_data.errors_from_sample_residuals
].count(True) > 1:
raise Usage(
"Specify only one of sdfac_auto, sdfac_refine or errors_from_sample_residuals."
)
def read_models(self):
# Read Nat's reference model from an MTZ file. XXX The observation
# type is given as F, not I--should they be squared? Check with Nat!
log = open("%s.log" % self.params.output.prefix, "w")
out = multi_out()
out.register("log", log, atexit_send_to=None)
out.register("stdout", sys.stdout)
print >> out, "I model"
if self.params.model is not None:
from xfel.merging.general_fcalc import run as run_fmodel
i_model = run_fmodel(self.params)
self.params.target_unit_cell = i_model.unit_cell()
self.params.target_space_group = i_model.space_group_info()
i_model.show_summary()
else:
i_model = None
print >> out, "Target unit cell and space group:"
print >> out, " ", self.params.target_unit_cell
print >> out, " ", self.params.target_space_group
from xfel.command_line.cxi_merge import consistent_set_and_model
self.miller_set, self.i_model = consistent_set_and_model(
self.params, i_model)
self.frame_files = get_observations(self.params)
self.out = out
def scale_all(self):
scaler = self.scaler_class(miller_set=self.miller_set,
i_model=self.i_model,
params=self.params,
log=self.out)
scaler.scale_all(self.frame_files)
return scaler
def finalize(self, scaler):
scaler.show_unit_cell_histograms()
if (self.params.rescale_with_average_cell):
average_cell_abc = scaler.uc_values.get_average_cell_dimensions()
average_cell = uctbx.unit_cell(
list(average_cell_abc) +
list(self.params.target_unit_cell.parameters()[3:]))
self.params.target_unit_cell = average_cell
print >> out, ""
print >> out, "#" * 80
print >> out, "RESCALING WITH NEW TARGET CELL"
print >> out, " average cell: %g %g %g %g %g %g" % \
self.params.target_unit_cell.parameters()
print >> out, ""
scaler.reset()
scaler.scale_all(frame_files)
scaler.show_unit_cell_histograms()
if False: #(self.params.output.show_plots) :
try:
plot_overall_completeness(completeness)
except Exception, e:
print "ERROR: can't show plots"
print " %s" % str(e)
print >> self.out, "\n"
sum_I, sum_I_SIGI = scaler.sum_intensities()
miller_set_avg = self.miller_set.customized_copy(
unit_cell=self.params.target_unit_cell)
table1 = show_overall_observations(
obs=miller_set_avg,
redundancy=scaler.completeness,
redundancy_to_edge=scaler.completeness_predictions,
summed_wt_I=scaler.summed_wt_I,
summed_weight=scaler.summed_weight,
ISIGI=scaler.ISIGI,
n_bins=self.params.output.n_bins,
title="Statistics for all reflections",
out=self.out,
work_params=self.params)
print >> self.out, ""
if self.params.model is not None:
n_refl, corr = scaler.get_overall_correlation(sum_I)
else:
n_refl, corr = ((scaler.completeness > 0).count(True), 0)
print >> self.out, "\n"
table2 = show_overall_observations(
obs=miller_set_avg,
redundancy=scaler.summed_N,
redundancy_to_edge=scaler.completeness_predictions,
summed_wt_I=scaler.summed_wt_I,
summed_weight=scaler.summed_weight,
ISIGI=scaler.ISIGI,
n_bins=self.params.output.n_bins,
title="Statistics for reflections where I > 0",
out=self.out,
work_params=self.params)
#from libtbx import easy_pickle
#easy_pickle.dump(file_name="stats.pickle", obj=stats)
#stats.report(plot=self.params.plot)
#miller_counts = miller_set_p1.array(data=stats.counts.as_double()).select(
# stats.counts != 0)
#miller_counts.as_mtz_dataset(column_root_label="NOBS").mtz_object().write(
# file_name="nobs.mtz")
if self.params.data_subsubsets.subsubset is not None and self.params.data_subsubsets.subsubset_total is not None:
easy_pickle.dump(
"scaler_%d.pickle" % self.params.data_subsubsets.subsubset,
scaler)
explanation = """
Explanation:
Completeness = # unique Miller indices present in data / # Miller indices theoretical in asymmetric unit
Asu. Multiplicity = # measurements / # Miller indices theoretical in asymmetric unit
Obs. Multiplicity = # measurements / # unique Miller indices present in data
Pred. Multiplicity = # predictions on all accepted images / # Miller indices theoretical in asymmetric unit"""
print >> self.out, explanation
mtz_file, miller_array = scaler.finalize_and_save_data()
#table_pickle_file = "%s_graphs.pkl" % self.params.output.prefix
#easy_pickle.dump(table_pickle_file, [table1, table2])
loggraph_file = os.path.abspath("%s_graphs.log" %
self.params.output.prefix)
f = open(loggraph_file, "w")
f.write(table1.format_loggraph())
f.write("\n")
f.write(table2.format_loggraph())
f.close()
result = scaling_result(miller_array=miller_array,
plots=scaler.get_plot_statistics(),
mtz_file=mtz_file,
loggraph_file=loggraph_file,
obs_table=table1,
all_obs_table=table2,
n_reflections=n_refl,
overall_correlation=corr)
easy_pickle.dump("%s.pkl" % self.params.output.prefix, result)
return result
