def write_node_preview(pool, parent, chosen_idx): assert(chosen_idx == sorted(chosen_idx)) print "chosen_idx", chosen_idx trr_out_tmp_fn = mktemp(suffix='.trr') trr_out_tmp = open(trr_out_tmp_fn, "wb") trr_in = TrrFile(parent.trr_fn) curr_frame = trr_in.first_frame for i in chosen_idx: for dummy in range(i - curr_frame.number): curr_frame = curr_frame.next() assert(curr_frame.number == i) trr_out_tmp.write(curr_frame.raw_data) trr_in.close() trr_out_tmp.close() node_preview_fn = "node_preview_from_" + parent.name + ".pdb" cmd = ["trjconv", "-f", trr_out_tmp_fn, "-o", node_preview_fn, "-s", parent.pdb_fn, "-n", pool.ndx_fn] p = Popen(cmd, stdin=PIPE) p.communicate(input="MOI\n") assert(p.wait() == 0) os.remove(trr_out_tmp_fn) print "Node preview (MOI only) written to file: %s" % node_preview_fn
def get_merged_edr(node):
# get list of edr files
edr_fnames = sorted([ node.dir+"/"+fn for fn in os.listdir(node.dir) if re.match("[^#].+run\d+.edr", fn) ])
assert( len(edr_fnames) == node.extensions_max+1 )
# find out about trr time step
trr = TrrFile(node.trr_fn)
dt = trr.first_frame.next().t - trr.first_frame.t
trr.close()
# dt is sometimes noisy in the final digits (three digits is femtosecond step = enough)
dt = np.around(dt, decimals=3)
time_offset = node.sampling_length+dt
for edr_fn in edr_fnames[1:]:
# adapt edr starting times
cmd = ["eneconv", "-f", edr_fn, "-o", edr_fn, "-settime"]
print("Calling: "+(" ".join(cmd)))
p = Popen(cmd, stdin=PIPE)
p.communicate(input=(str(time_offset)+"\n"))
assert(p.wait() == 0)
time_offset += node.extensions_length+dt
# concatenate edr files with adapted starting times
cmd = ["eneconv", "-f"] + edr_fnames + ["-o", node.dir+"/ener.edr"]
print("Calling: "+(" ".join(cmd)))
p = Popen(cmd)
retcode = p.wait()
assert(retcode == 0)
def main():
options = options_desc.parse_args(sys.argv)[0]
pool = Pool()
needy_nodes = pool.where("state == 'merge-able'").multilock()
if(len(needy_nodes) == 0):
return
# find out about trr time step
dt = 0
nodeDir = needy_nodes[0].dir.split('/')[-1]
for fn in os.listdir(needy_nodes[0].dir):
if re.match("^"+nodeDir+".+run\d+\.trr", fn):
trr = TrrFile(needy_nodes[0].dir+"/"+fn)
dt = trr.first_frame.next().t - trr.first_frame.t
trr.close()
break
# dt is sometimes noisy in the final digits (three digits is femtosecond step = enough)
dt = np.around(dt, decimals=3)
for n in needy_nodes:
if(options.trr):
# merge sampling trajectories
trr_fns = sorted([ fn for fn in os.listdir(n.dir) if re.match("[^#].+run\d+.trr", fn) ])
cmd = ["trjcat", "-f"]
cmd += trr_fns
cmd += ["-o", "../../"+n.trr_fn, "-cat"]
print("Calling: %s"%" ".join(cmd))
check_call(cmd, cwd=n.dir)
if(options.edr):
# merge edr files
# get list of edr-files
edr_fnames = sorted([n.dir+"/"+fn for fn in os.listdir(n.dir) if re.match("[^#].+run\d+.edr", fn)])
assert( len(edr_fnames) == n.extensions_counter+1 )
assert( len(edr_fnames) == n.extensions_max+1 )
time_offset = n.sampling_length+dt
for edr_fn in edr_fnames[1:]:
# adapt edr starting times
cmd = ["eneconv", "-f", edr_fn, "-o", edr_fn, "-settime"]
print("Calling: "+(" ".join(cmd)))
p = Popen(cmd, stdin=PIPE)
p.communicate(input=(str(time_offset)+"\n"))
assert(p.wait() == 0)
time_offset += n.extensions_length+dt
# concatenate edr files with adapted starting times
cmd = ["eneconv", "-f"] + edr_fnames + ["-o", n.dir+"/ener.edr"]
print("Calling: "+(" ".join(cmd)))
p = Popen(cmd)
retcode = p.wait()
assert(retcode == 0)
needy_nodes.unlock()
def read_trajectory(self, fn):
"""
Reads a trr-trajectory, resolves periodic boundary conditions
with L{PbcResolver} and calculates internal coordinates.
@param fn: filename of a gromacs trr trajectory.
@return: L{InternalArray}
"""
required_atoms = set(sum([c.atoms for c in self], ()))
atoms_start = min(required_atoms)
atoms_end = max(required_atoms) + 1
f_trr = TrrFile(fn)
(frames_x, frames_box) = f_trr.read_frames(atoms_start,
atoms_end,
read_boxes=True)
f_trr.close()
pbc = PbcResolver(frames_box)
def dx_provider(atom1, atom2):
ai = frames_x[:, atom1 - atoms_start, :]
aj = frames_x[:, atom2 - atoms_start, :]
return (pbc.rvec_sub(ai, aj))
array = np.column_stack([c.from_externals(dx_provider) for c in self])
return (InternalArray(self, array))
def main():
if(len(sys.argv) != 3):
print("Takes a trr-trajectory and converts it into a mat-file for matlab.")
print("Usage: trr2mat.py <trr-input-file> <mat-output-file>")
sys.exit(1)
trr_fn = sys.argv[1]
mat_fn = sys.argv[2]
print("Opening: %s"%trr_fn)
f = TrrFile(trr_fn)
data = f.read_frames()
f.close()
print("Loaded trr-file with shape: "+str(data.shape))
print("Writing: %s"%mat_fn)
savemat(mat_fn, {"trr":data})
def extract_frames(pool):
needy_nodes = pool.where("state == 'created'")
# we want to scan through a parent-trr only once - saves time
parents = set([n.parent for n in needy_nodes])
for p in parents:
childs = [n for n in needy_nodes if n.parent == p]
childs.sort(key=lambda x: x.parent_frame_num)
trr_in = TrrFile(p.trr_fn)
frame = trr_in.first_frame
for n in childs:
for dummy in range(n.parent_frame_num - frame.number):
frame = frame.next()
assert (frame.number == n.parent_frame_num)
trr_tmp_fn = mktemp(suffix='.trr')
trr_tmp = open(trr_tmp_fn, "wb")
trr_tmp.write(frame.raw_data)
trr_tmp.close()
cmd = [
"trjconv", "-f", trr_tmp_fn, "-o", n.pdb_fn, "-s",
n.parent.pdb_fn
]
p = Popen(cmd, stdin=PIPE)
p.communicate(input="System\n")
assert (p.wait() == 0)
os.remove(trr_tmp_fn)
trr_in.close()
# Check if the right frames where extracted
# In principle, PDB coordinates should have a precision of 1e-4 nm
# beause they are given in Angström with three decimal places.
for n in needy_nodes:
a = pool.converter.read_pdb(n.pdb_fn)
d = np.max(np.abs(n.internals.array - a.array))
print n.name + ": pdb vs internals deviation: %.2e" % np.max(
np.abs(n.internals.array - a.array))
assert (1e-2 > d)
def read_trajectory(self, fn):
"""
Reads a trr-trajectory, resolves periodic boundary conditions
with L{PbcResolver} and calculates internal coordinates.
@param fn: filename of a gromacs trr trajectory.
@return: L{InternalArray}
"""
required_atoms = set(sum([c.atoms for c in self], () ))
atoms_start = min(required_atoms)
atoms_end = max(required_atoms) + 1
f_trr = TrrFile(fn)
(frames_x, frames_box) = f_trr.read_frames(atoms_start, atoms_end, read_boxes=True)
f_trr.close()
pbc = PbcResolver(frames_box)
def dx_provider(atom1, atom2):
ai = frames_x[:,atom1-atoms_start,:]
aj = frames_x[:,atom2-atoms_start,:]
return(pbc.rvec_sub(ai, aj))
array = np.column_stack([ c.from_externals(dx_provider) for c in self ])
return( InternalArray(self, array) )
def extract_frames(pool):
needy_nodes = pool.where("state == 'created'")
# we want to scan through a parent-trr only once - saves time
parents = set([n.parent for n in needy_nodes])
for p in parents:
childs = [n for n in needy_nodes if n.parent == p]
childs.sort(key=lambda x: x.parent_frame_num)
trr_in = TrrFile(p.trr_fn)
frame = trr_in.first_frame
for n in childs:
for dummy in range(n.parent_frame_num - frame.number):
frame = frame.next()
assert(frame.number == n.parent_frame_num)
trr_tmp_fn = mktemp(suffix='.trr')
trr_tmp = open(trr_tmp_fn, "wb")
trr_tmp.write(frame.raw_data)
trr_tmp.close()
cmd = ["trjconv", "-f", trr_tmp_fn, "-o", n.pdb_fn, "-s", n.parent.pdb_fn]
p = Popen(cmd, stdin=PIPE)
p.communicate(input="System\n")
assert(p.wait() == 0)
os.remove(trr_tmp_fn)
trr_in.close()
# Check if the right frames where extracted
# In principle, PDB coordinates should have a precision of 1e-4 nm
# beause they are given in Angström with three decimal places.
for n in needy_nodes:
a = pool.converter.read_pdb(n.pdb_fn)
d = np.max(np.abs(n.internals.array - a.array))
print n.name+": pdb vs internals deviation: %.2e"%np.max(np.abs(n.internals.array - a.array))
assert(1e-2 > d)
def main():
options = options_desc.parse_args(sys.argv)[0]
zgf_cleanup.main()
pool = Pool()
npz_file = np.load(pool.chi_mat_fn)
chi_matrix = npz_file['matrix']
node_names = npz_file['node_names']
n_clusters = npz_file['n_clusters']
active_nodes = [Node(nn) for nn in node_names]
# create and open dest_files, intialize counters for statistics
dest_filenames = [
pool.analysis_dir + "cluster%d.trr" % (c + 1)
for c in range(n_clusters)
]
dest_files = [open(fn, "wb") for fn in dest_filenames]
dest_frame_counters = np.zeros(n_clusters)
# For each active node...
for (i, n) in enumerate(active_nodes):
# ... find the clusters to which it belongs (might be more than one)...
belonging_clusters = np.argwhere(
chi_matrix[i] > options.node_threshold)
# ... and find all typical frames of this node.
#TODO not an optimal solution... discuss
# per default, we take every frame with above average weight
frame_threshold = options.frame_threshold * 2 * np.mean(n.frameweights)
typical_frame_nums = np.argwhere(n.frameweights > frame_threshold)
# Go through the node's trajectory ...
trr_in = TrrFile(n.trr_fn)
curr_frame = trr_in.first_frame
for i in typical_frame_nums:
# ...stop at each typical frame...
while (i != curr_frame.number):
curr_frame = curr_frame.next()
assert (curr_frame.number == i)
#... and copy it into the dest_file of each belonging cluster.
for c in belonging_clusters:
dest_files[c].write(curr_frame.raw_data)
dest_frame_counters[c] += 1
trr_in.close() # close source file
# close dest_files
for f in dest_files:
f.close()
del (dest_files)
# desolvate cluster-trajectories 'in-place'
if (not options.write_sol):
for dest_fn in dest_filenames:
tmp_fn = mktemp(suffix='.trr', dir=pool.analysis_dir)
os.rename(dest_fn, tmp_fn) # works as both files are in same dir
cmd = ["trjconv", "-f", tmp_fn, "-o", dest_fn, "-n", pool.ndx_fn]
p = Popen(cmd, stdin=PIPE)
p.communicate(input="MOI\n")
assert (p.wait() == 0)
os.remove(tmp_fn)
# register dependencies
for fn in dest_filenames:
register_file_dependency(fn, pool.chi_mat_fn)
# check number of written frames
sys.stdout.write("Checking lenghts of written trajectories... ")
for i in range(n_clusters):
f = TrrFile(dest_filenames[i])
assert (f.count_frames() == dest_frame_counters[i])
f.close()
print("done.")
#output statistics
print "\n### Extraction summary ###\nnode threshold: %1.1f, frame threshold: %1.1f" % (
options.node_threshold, options.frame_threshold)
print "Cluster trajectories were written to %s:" % pool.analysis_dir
for (c, f) in enumerate(dest_frame_counters):
print "cluster%d.trr [%d frames] from node(s):" % (c + 1, f)
print list(np.argwhere(chi_matrix[:, c] > options.node_threshold).flat)
def main():
options = options_desc.parse_args(sys.argv)[0]
if options.common_filename:
options.molecule = options.common_filename + ".pdb"
options.presampling = options.common_filename + ".trr"
options.internals = options.common_filename + ".int"
options.grompp = options.common_filename + ".mdp"
options.topology = options.common_filename + ".top"
options.index = options.common_filename + ".ndx"
print("Options:\n%s\n" % pformat(eval(str(options))))
assert path.exists(options.molecule)
assert path.exists(options.presampling)
assert path.exists(options.internals)
assert path.exists(options.grompp)
assert path.exists(options.topology)
# TODO: what if there is no index-file? (make_ndx)
assert path.exists(options.index)
assert "MOI" in gromacs.read_index_file(options.index), "group MOI should be defined in index file"
# checks e.g. if the mdp-file looks good
mdp_options = gromacs.read_mdp_file(options.grompp)
# options we cannot fix
for ref_t in re.findall("[0-9]+", mdp_options["ref_t"]):
assert int(ref_t) == options.temperature, "temperature in mdp file does not match ZIBgridfree temperature"
# TODO drop options.temperature and get temperature directly from mdp file... ask again if temperature is above 310K
# options we can fix
mdp_options_dirty = False # if set, a new mdp-file will be written
required_mdp_options = {"dihre": "yes", "dihre_fc": "1", "disre": "simple", "disre_fc": "1"}
for (k, v) in required_mdp_options.items():
if mdp_options.has_key(k):
assert mdp_options[k] == v # check, if we would overwrite something
else:
mdp_options[k] = v
mdp_options_dirty = True
if mdp_options.has_key("energygrps"):
assert "MOI" in [
str(egrp) for egrp in re.findall("[\S]+", mdp_options["energygrps"])
], "group MOI should be among energygrps in mdp file"
else:
mdp_options["energygrps"] = "MOI"
mdp_options_dirty = True
a, b = mdp_options.has_key("nstxout"), mdp_options.has_key("nstenergy")
if a and not b:
mdp_options["nstenergy"] = mdp_options["nstxout"]
mdp_options_dirty = True
elif b and not a:
mdp_options["nstxout"] = mdp_options["nstenergy"]
mdp_options_dirty = True
elif b and a:
assert mdp_options["nstxout"] == mdp_options["nstenergy"], "nstxout should equal nstenergy"
if int(mdp_options["nsteps"]) > 1e6:
msg = "Number of MD-steps?"
mdp_options["nsteps"] = str(userinput(msg, "int", default=int(mdp_options["nsteps"])))
# create a fixed mdp-file
if mdp_options_dirty:
print("Creating copy of mdp-file and adding missing options.")
out_fn = options.grompp.rsplit(".", 1)[0] + "_fixed.mdp"
f = open(out_fn, "w") # append
f.write("; Generated by zgf_create_pool\n")
for i in mdp_options.items():
f.write("%s = %s\n" % i)
f.write("; EOF\n")
f.close()
options.grompp = out_fn
# check if subsampling is reasonable
if os.path.getsize(options.presampling) > 100e6: # 100MB
print("Presampling trajectory is large")
trr = TrrFile(options.presampling)
dt = trr.first_frame.next().t - trr.first_frame.t
trr.close()
print("Presampling timestep is %.2f ps" % dt)
if dt < 10: # picoseconds
# TODO: maybe calculate subsampling factor individually, or ask?
msg = "Subsample presampling trajectory by a tenth?"
if userinput(msg, "bool"):
out_fn = options.presampling.rsplit(".", 1)[0] + "_tenth.trr"
cmd = ["trjconv", "-f", options.presampling, "-o", out_fn, "-skip", "10"]
check_call(cmd)
options.presampling = out_fn
# balance linears
if options.balance_linears:
print("Balance Linears")
old_converter = Converter(options.internals)
print("Loading presampling....")
frames = old_converter.read_trajectory(options.presampling)
new_coord_list = []
for c in old_converter:
if not isinstance(c, LinearCoordinate):
new_coord_list.append(c)
continue # we do not work on other Coordinate-Types
# TODO: is this a good way to determine new_weight and new_offset???
new_weight = c.weight / sqrt(2 * frames.var().getcoord(c))
new_offset = c.offset + frames.mean().getcoord(c)
new_coord = LinearCoordinate(*c.atoms, label=c.label, weight=new_weight, offset=new_offset)
new_coord_list.append(new_coord)
new_converter = Converter(coord_list=new_coord_list)
assert old_converter.filename.endswith(".int")
options.internals = old_converter.filename[:-4] + "_balanced.int"
print("Writing balanced Converter to: " + options.internals)
f = open(options.internals, "w")
f.write(new_converter.serialize())
f.close()
assert len(Converter(options.internals)) == len(new_coord_list) # try parsing
# Finally: Create root-node and pool
pool = Pool()
if len(pool) != 0:
print("ERROR: A pool already exists here.")
sys.exit(1)
pool.int_fn = options.internals
pool.mdp_fn = options.grompp
pool.top_fn = options.topology
pool.ndx_fn = options.index
pool.temperature = options.temperature
pool.gr_threshold = options.gr_threshold
pool.gr_chains = options.gr_chains
pool.alpha = None
pool.save() # save pool for the first time...
# ... then we can save the first node...
node0 = Node()
node0.state = "refined"
node0.save() # also creates the node directory ... needed for symlink
os.symlink(os.path.relpath(options.presampling, node0.dir), node0.trr_fn)
os.symlink(os.path.relpath(options.molecule, node0.dir), node0.pdb_fn)
pool.root_name = node0.name
pool.save() # ... now we have to save the pool again.
if not path.exists("analysis"):
os.mkdir("analysis")
def main():
options = options_desc.parse_args(sys.argv)[0]
if(options.common_filename):
options.molecule = options.common_filename+".pdb"
options.presampling = options.common_filename+".trr"
options.internals = options.common_filename+".int"
options.grompp = options.common_filename+".mdp"
options.topology = options.common_filename+".top"
options.index = options.common_filename+".ndx"
print("Options:\n%s\n"%pformat(eval(str(options))))
assert(path.exists(options.molecule))
assert(path.exists(options.presampling))
assert(path.exists(options.internals))
assert(path.exists(options.grompp))
assert(path.exists(options.topology))
#TODO: what if there is no index-file? (make_ndx)
assert(path.exists(options.index))
assert('moi' in gromacs.read_index_file(options.index)), "group 'MOI' should be defined in index file"
# checks e.g. if the mdp-file looks good
mdp_options = gromacs.read_mdp_file(options.grompp)
temperatures = [ref_t for ref_t in re.findall("[0-9]+", mdp_options["ref_t"])]
assert(len(set(temperatures)) == 1), "temperature definition in mdp file is ambiguous"
temperature = temperatures[0]
# get sampling temperature from mdp file
if(int(temperature) > 310):
if not(userinput("Your sampling temperature is set to %s K. Continue?"%temperature, "bool")):
sys.exit("Quit by user.")
# options we can fix
mdp_options_dirty = False #if set, a new mdp-file will be written
# the value of the following options need to be fixed
critical_mdp_options = {"dihre":"yes", "dihre_fc":"1", "disre":"simple", "disre_fc":"1", "gen_temp":temperature}
for (k,v) in critical_mdp_options.items():
if(mdp_options.has_key(k) and mdp_options[k].strip() != v):
print "Error. I do not want to use '%s' for option '%s' ('%s' required). Please fix your mdp file."%(mdp_options[k].strip(),k,v)
sys.exit("Quitting.")
else:
mdp_options[k] = v
mdp_options_dirty = True
# the value of the following options does not matter, but they should be there
noncritical_mdp_options = {"tcoupl":"no", "pcoupl":"no", "gen_vel":"no", "gen_seed":"-1"}
for (k,v) in noncritical_mdp_options.items():
if not(mdp_options.has_key(k)):
mdp_options[k] = v
mdp_options_dirty = True
a = mdp_options.has_key("energygrps") and "moi" not in [str(egrp) for egrp in re.findall('[\S]+', mdp_options["energygrps"])]
b = not(mdp_options.has_key("energygrps"))
if(a or b):
if not(userinput("'MOI' is not defined as an energy group in your mdp file. Maybe you have forgotten to define proper 'energygrps'. Continue?", "bool")):
sys.exit("Quit by user.")
a, b = mdp_options.has_key("nstxout"), mdp_options.has_key("nstenergy")
if(a and not b):
mdp_options["nstenergy"] = mdp_options["nstxout"]
mdp_options_dirty = True
elif(b and not a):
mdp_options["nstxout"] = mdp_options["nstenergy"]
mdp_options_dirty = True
elif(b and a):
assert(mdp_options["nstxout"] == mdp_options["nstenergy"]), "nstxout should equal nstenergy"
if(int(mdp_options["nsteps"]) > 1e6):
msg = "Number of MD-steps?"
mdp_options["nsteps"] = str( userinput(msg, "int", default=int(mdp_options["nsteps"])) )
# create a fixed mdp-file
if(mdp_options_dirty):
print("Creating copy of mdp-file and adding missing options.")
out_fn = options.grompp.rsplit(".", 1)[0] + "_fixed.mdp"
f = open(out_fn, "w") # append
f.write("; Generated by zgf_create_pool\n")
for i in sorted(mdp_options.items()):
f.write("%s = %s\n"%i)
f.write("; EOF\n")
f.close()
options.grompp = out_fn
# check if subsampling is reasonable
if(os.path.getsize(options.presampling) > 100e6): # 100MB
print("Presampling trajectory is large")
trr = TrrFile(options.presampling)
dt = trr.first_frame.next().t - trr.first_frame.t
trr.close()
print("Presampling timestep is %.2f ps"%dt)
if(dt < 10): # picoseconds
#TODO: maybe calculate subsampling factor individually, or ask?
msg = "Subsample presampling trajectory by a tenth?"
if(userinput(msg, "bool")):
out_fn = options.presampling.rsplit(".", 1)[0] + "_tenth.trr"
cmd = ["trjconv", "-f", options.presampling, "-o", out_fn, "-skip", "10"]
check_call(cmd)
options.presampling = out_fn
# balance linears
if(options.balance_linears):
print("Balance Linears")
old_converter = Converter(options.internals)
print("Loading presampling....")
frames = old_converter.read_trajectory(options.presampling)
new_coord_list = []
for c in old_converter:
if(not isinstance(c, LinearCoordinate)):
new_coord_list.append(c)
continue # we do not work on other Coordinate-Types
#TODO: is this a good way to determine new_weight and new_offset???
new_weight = c.weight / sqrt(2*frames.var().getcoord(c))
new_offset = c.offset + frames.mean().getcoord(c)
new_coord = LinearCoordinate(*c.atoms, label=c.label, weight=new_weight, offset=new_offset)
new_coord_list.append(new_coord)
new_converter = Converter(coord_list=new_coord_list)
assert(old_converter.filename.endswith(".int"))
options.internals = old_converter.filename[:-4] + "_balanced.int"
print("Writing balanced Converter to: "+options.internals)
f = open(options.internals, "w")
f.write(new_converter.serialize())
f.close()
assert(len(Converter(options.internals)) == len(new_coord_list)) #try parsing
# Finally: Create root-node and pool
pool = Pool()
if(len(pool) != 0):
print("ERROR: A pool already exists here.")
sys.exit(1)
pool.int_fn = options.internals
pool.mdp_fn = options.grompp
pool.top_fn = options.topology
pool.ndx_fn = options.index
pool.temperature = int(temperature)
pool.gr_threshold = options.gr_threshold
pool.gr_chains = options.gr_chains
pool.alpha = None
pool.save() # save pool for the first time...
# ... then we can save the first node...
node0 = Node()
node0.state = "refined"
node0.save() # also creates the node directory ... needed for symlink
os.symlink(os.path.relpath(options.presampling, node0.dir), node0.trr_fn)
os.symlink(os.path.relpath(options.molecule, node0.dir), node0.pdb_fn)
pool.root_name = node0.name
pool.save() #... now we have to save the pool again.
if(not path.exists("analysis")):
os.mkdir("analysis")
def main():
options = options_desc.parse_args(sys.argv)[0]
zgf_cleanup.main()
pool = Pool()
npz_file = np.load(pool.chi_mat_fn)
chi_matrix = npz_file['matrix']
node_names = npz_file['node_names']
n_clusters = npz_file['n_clusters']
active_nodes = [Node(nn) for nn in node_names]
# create and open dest_files, intialize counters for statistics
dest_filenames = [ pool.analysis_dir+"cluster%d.trr"%(c+1) for c in range(n_clusters) ]
dest_files = [ open(fn, "wb") for fn in dest_filenames ]
dest_frame_counters = np.zeros(n_clusters)
# For each active node...
for (i, n) in enumerate(active_nodes):
# ... find the clusters to which it belongs (might be more than one)...
belonging_clusters = np.argwhere(chi_matrix[i] > options.node_threshold)
# ... and find all typical frames of this node.
#TODO not an optimal solution... discuss
# per default, we take every frame with above average weight
frame_threshold = options.frame_threshold*2*np.mean(n.frameweights)
typical_frame_nums = np.argwhere(n.frameweights > frame_threshold)
# Go through the node's trajectory ...
trr_in = TrrFile(n.trr_fn)
curr_frame = trr_in.first_frame
for i in typical_frame_nums:
# ...stop at each typical frame...
while(i != curr_frame.number):
curr_frame = curr_frame.next()
assert(curr_frame.number == i)
#... and copy it into the dest_file of each belonging cluster.
for c in belonging_clusters:
dest_files[c].write(curr_frame.raw_data)
dest_frame_counters[c] += 1
trr_in.close() # close source file
# close dest_files
for f in dest_files:
f.close()
del(dest_files)
# desolvate cluster-trajectories 'in-place'
if(not options.write_sol):
for dest_fn in dest_filenames:
tmp_fn = mktemp(suffix='.trr', dir=pool.analysis_dir)
os.rename(dest_fn, tmp_fn) # works as both files are in same dir
cmd = ["trjconv", "-f", tmp_fn, "-o", dest_fn, "-n", pool.ndx_fn]
p = Popen(cmd, stdin=PIPE)
p.communicate(input="MOI\n")
assert(p.wait() == 0)
os.remove(tmp_fn)
# register dependencies
for fn in dest_filenames:
register_file_dependency(fn, pool.chi_mat_fn)
# check number of written frames
sys.stdout.write("Checking lenghts of written trajectories... ")
for i in range(n_clusters):
f = TrrFile(dest_filenames[i])
assert(f.count_frames() == dest_frame_counters[i])
f.close()
print("done.")
#output statistics
print "\n### Extraction summary ###\nnode threshold: %1.1f, frame threshold: %1.1f"%(options.node_threshold, options.frame_threshold)
print "Cluster trajectories were written to %s:"%pool.analysis_dir
for (c, f) in enumerate(dest_frame_counters):
print "cluster%d.trr [%d frames] from node(s):"%(c+1, f)
print list(np.argwhere(chi_matrix[:,c] > options.node_threshold).flat)
def main():
options = options_desc.parse_args(sys.argv)[0]
if (options.common_filename):
options.molecule = options.common_filename + ".pdb"
options.presampling = options.common_filename + ".trr"
options.internals = options.common_filename + ".int"
options.grompp = options.common_filename + ".mdp"
options.topology = options.common_filename + ".top"
options.index = options.common_filename + ".ndx"
print("Options:\n%s\n" % pformat(eval(str(options))))
assert (path.exists(options.molecule))
assert (path.exists(options.presampling))
assert (path.exists(options.internals))
assert (path.exists(options.grompp))
assert (path.exists(options.topology))
#TODO: what if there is no index-file? (make_ndx)
assert (path.exists(options.index))
assert ('moi' in gromacs.read_index_file(
options.index)), "group 'MOI' should be defined in index file"
# checks e.g. if the mdp-file looks good
mdp_options = gromacs.read_mdp_file(options.grompp)
temperatures = [
ref_t for ref_t in re.findall("[0-9]+", mdp_options["ref_t"])
]
assert (len(set(temperatures)) == 1
), "temperature definition in mdp file is ambiguous"
temperature = temperatures[0]
# get sampling temperature from mdp file
if (int(temperature) > 310):
if not (userinput(
"Your sampling temperature is set to %s K. Continue?" %
temperature, "bool")):
sys.exit("Quit by user.")
# options we can fix
mdp_options_dirty = False #if set, a new mdp-file will be written
# the value of the following options need to be fixed
critical_mdp_options = {
"dihre": "yes",
"dihre_fc": "1",
"disre": "simple",
"disre_fc": "1",
"gen_temp": temperature
}
for (k, v) in critical_mdp_options.items():
if (mdp_options.has_key(k) and mdp_options[k].strip() != v):
print "Error. I do not want to use '%s' for option '%s' ('%s' required). Please fix your mdp file." % (
mdp_options[k].strip(), k, v)
sys.exit("Quitting.")
else:
mdp_options[k] = v
mdp_options_dirty = True
# the value of the following options does not matter, but they should be there
noncritical_mdp_options = {
"tcoupl": "no",
"pcoupl": "no",
"gen_vel": "no",
"gen_seed": "-1"
}
for (k, v) in noncritical_mdp_options.items():
if not (mdp_options.has_key(k)):
mdp_options[k] = v
mdp_options_dirty = True
a = mdp_options.has_key("energygrps") and "moi" not in [
str(egrp) for egrp in re.findall('[\S]+', mdp_options["energygrps"])
]
b = not (mdp_options.has_key("energygrps"))
if (a or b):
if not (userinput(
"'MOI' is not defined as an energy group in your mdp file. Maybe you have forgotten to define proper 'energygrps'. Continue?",
"bool")):
sys.exit("Quit by user.")
a, b = mdp_options.has_key("nstxout"), mdp_options.has_key("nstenergy")
if (a and not b):
mdp_options["nstenergy"] = mdp_options["nstxout"]
mdp_options_dirty = True
elif (b and not a):
mdp_options["nstxout"] = mdp_options["nstenergy"]
mdp_options_dirty = True
elif (b and a):
assert (mdp_options["nstxout"] == mdp_options["nstenergy"]
), "nstxout should equal nstenergy"
if (int(mdp_options["nsteps"]) > 1e6):
msg = "Number of MD-steps?"
mdp_options["nsteps"] = str(
userinput(msg, "int", default=int(mdp_options["nsteps"])))
# create a fixed mdp-file
if (mdp_options_dirty):
print("Creating copy of mdp-file and adding missing options.")
out_fn = options.grompp.rsplit(".", 1)[0] + "_fixed.mdp"
f = open(out_fn, "w") # append
f.write("; Generated by zgf_create_pool\n")
for i in sorted(mdp_options.items()):
f.write("%s = %s\n" % i)
f.write("; EOF\n")
f.close()
options.grompp = out_fn
# check if subsampling is reasonable
if (os.path.getsize(options.presampling) > 100e6): # 100MB
print("Presampling trajectory is large")
trr = TrrFile(options.presampling)
dt = trr.first_frame.next().t - trr.first_frame.t
trr.close()
print("Presampling timestep is %.2f ps" % dt)
if (dt < 10): # picoseconds
#TODO: maybe calculate subsampling factor individually, or ask?
msg = "Subsample presampling trajectory by a tenth?"
if (userinput(msg, "bool")):
out_fn = options.presampling.rsplit(".", 1)[0] + "_tenth.trr"
cmd = [
"trjconv", "-f", options.presampling, "-o", out_fn,
"-skip", "10"
]
check_call(cmd)
options.presampling = out_fn
# balance linears
if (options.balance_linears):
print("Balance Linears")
old_converter = Converter(options.internals)
print("Loading presampling....")
frames = old_converter.read_trajectory(options.presampling)
new_coord_list = []
for c in old_converter:
if (not isinstance(c, LinearCoordinate)):
new_coord_list.append(c)
continue # we do not work on other Coordinate-Types
#TODO: is this a good way to determine new_weight and new_offset???
new_weight = c.weight / sqrt(2 * frames.var().getcoord(c))
new_offset = c.offset + frames.mean().getcoord(c)
new_coord = LinearCoordinate(*c.atoms,
label=c.label,
weight=new_weight,
offset=new_offset)
new_coord_list.append(new_coord)
new_converter = Converter(coord_list=new_coord_list)
assert (old_converter.filename.endswith(".int"))
options.internals = old_converter.filename[:-4] + "_balanced.int"
print("Writing balanced Converter to: " + options.internals)
f = open(options.internals, "w")
f.write(new_converter.serialize())
f.close()
assert (len(Converter(options.internals)) == len(new_coord_list)
) #try parsing
# Finally: Create root-node and pool
pool = Pool()
if (len(pool) != 0):
print("ERROR: A pool already exists here.")
sys.exit(1)
pool.int_fn = options.internals
pool.mdp_fn = options.grompp
pool.top_fn = options.topology
pool.ndx_fn = options.index
pool.temperature = int(temperature)
pool.gr_threshold = options.gr_threshold
pool.gr_chains = options.gr_chains
pool.alpha = None
pool.save() # save pool for the first time...
# ... then we can save the first node...
node0 = Node()
node0.state = "refined"
node0.save() # also creates the node directory ... needed for symlink
os.symlink(os.path.relpath(options.presampling, node0.dir), node0.trr_fn)
os.symlink(os.path.relpath(options.molecule, node0.dir), node0.pdb_fn)
pool.root_name = node0.name
pool.save() #... now we have to save the pool again.
if (not path.exists("analysis")):
os.mkdir("analysis")
def main():
options = options_desc.parse_args(sys.argv)[0]
pool = Pool()
needy_nodes = pool.where("state == 'merge-able'").multilock()
if (len(needy_nodes) == 0):
return
# find out about trr time step
dt = 0
nodeDir = needy_nodes[0].dir.split('/')[-1]
for fn in os.listdir(needy_nodes[0].dir):
if re.match("^" + nodeDir + ".+run\d+\.trr", fn):
trr = TrrFile(needy_nodes[0].dir + "/" + fn)
dt = trr.first_frame.next().t - trr.first_frame.t
trr.close()
break
# dt is sometimes noisy in the final digits (three digits is femtosecond step = enough)
dt = np.around(dt, decimals=3)
for n in needy_nodes:
if (options.trr):
# merge sampling trajectories
trr_fns = sorted([
fn for fn in os.listdir(n.dir)
if re.match("[^#].+run\d+.trr", fn)
])
cmd = ["trjcat", "-f"]
cmd += trr_fns
cmd += ["-o", "../../" + n.trr_fn, "-cat"]
print("Calling: %s" % " ".join(cmd))
check_call(cmd, cwd=n.dir)
if (options.edr):
# merge edr files
# get list of edr-files
edr_fnames = sorted([
n.dir + "/" + fn for fn in os.listdir(n.dir)
if re.match("[^#].+run\d+.edr", fn)
])
assert (len(edr_fnames) == n.extensions_counter + 1)
assert (len(edr_fnames) == n.extensions_max + 1)
time_offset = n.sampling_length + dt
for edr_fn in edr_fnames[1:]:
# adapt edr starting times
cmd = ["eneconv", "-f", edr_fn, "-o", edr_fn, "-settime"]
print("Calling: " + (" ".join(cmd)))
p = Popen(cmd, stdin=PIPE)
p.communicate(input=(str(time_offset) + "\n"))
assert (p.wait() == 0)
time_offset += n.extensions_length + dt
# concatenate edr files with adapted starting times
cmd = ["eneconv", "-f"] + edr_fnames + ["-o", n.dir + "/ener.edr"]
print("Calling: " + (" ".join(cmd)))
p = Popen(cmd)
retcode = p.wait()
assert (retcode == 0)
needy_nodes.unlock()
