def main(argv=None):
if(argv==None):
argv = sys.argv
options = options_desc.parse_args(argv)[0]
assert(not(options.refine_all and options.extend_all))
pool = Pool()
needy_nodes = pool.where("isa_partition and is_sampled").multilock()
# 1. Trying to detect fake convergence
for n in pool.where("state == 'converged'"):
means = kmeans(n.trajectory, k=2)
d = (means[0] - means[1]).norm2()
if(d > 2.0 and (options.refine_all or userinput("%s has converged but appears to have a bimodal distribution.\nDo you want to refine?"%n.name, "bool"))): #TODO decide upon threshold (per coordinate?)
refine(n, options)
# 2. Dealing with not-converged nodes
for n in pool.where("state == 'not-converged'"):
if(not(options.refine_all or options.extend_all)):
choice = userchoice("%s has not converged. What do you want to do?"%n.name, ['_refine', '_extend', '_ignore'])
if(options.refine_all or choice=="r"):
refine(n, options)
elif(options.extend_all or choice=="e"):
extend(n)
elif(choice=="i"):
continue
for n in needy_nodes:
n.save()
n.unlock()
zgf_setup_nodes.main()
zgf_grompp.main()
zgf_cleanup.main()
def main(argv=None):
if argv == None:
argv = sys.argv
options = options_desc.parse_args(argv)[0]
pool = Pool()
found_parents = [n for n in pool if n.name == options.parent_node]
assert len(found_parents) == 1
parent = found_parents[0]
chosen_idx = np.linspace(start=0, stop=parent.trajectory.n_frames - 1, num=options.numnodes).astype(int)
print "choosen_idx: ", chosen_idx
for i in chosen_idx:
n = Node()
n.parent_frame_num = i
n.parent = parent
n.state = "created"
n.extensions_counter = 0
n.extensions_max = 0
n.extensions_length = 0
n.sampling_length = parent.sampling_length * 3
n.internals = parent.trajectory.getframe(i)
pool.append(n)
n.save()
def main(argv=None):
if (argv == None):
argv = sys.argv
options = options_desc.parse_args(argv)[0]
pool = Pool()
found_parents = [n for n in pool if n.name == options.parent_node]
assert (len(found_parents) == 1)
parent = found_parents[0]
chosen_idx = np.linspace(start=0,
stop=parent.trajectory.n_frames - 1,
num=options.numnodes).astype(int)
print "choosen_idx: ", chosen_idx
for i in chosen_idx:
n = Node()
n.parent_frame_num = i
n.parent = parent
n.state = "created"
n.extensions_counter = 0
n.extensions_max = 0
n.extensions_length = 0
n.sampling_length = parent.sampling_length * 3
n.internals = parent.trajectory.getframe(i)
pool.append(n)
n.save()
def main():
options = options_desc.parse_args(sys.argv)[0]
pool = Pool()
choice = "state in ('converged', 'refined')"
if(options.ignore_convergence):
choice = "state in ('converged', 'not-converged', 'refined')"
needy_nodes = NodeList([n for n in pool.where(choice) if not n == pool.root]) # we won't touch the root
if not(len(needy_nodes)):
sys.exit("Nothing to do.")
if not(userinput("Once the solvent has been removed, further refinement of the pool is not possible. This includes the generation of unrestrained transition nodes! Continue?", "bool")):
sys.exit("Quit by user.")
assert(len(needy_nodes) == len(needy_nodes.multilock())) # make sure we lock ALL nodes
try:
for n in needy_nodes:
discard_solvent(n, "pdb")
discard_solvent(n, "trr")
for n in needy_nodes:
n.unlock()
except:
traceback.print_exc()
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 main():
options = options_desc.parse_args(sys.argv)[0]
pool = Pool()
choice = "state in ('converged', 'refined')"
if (options.ignore_convergence):
choice = "state in ('converged', 'not-converged', 'refined')"
needy_nodes = NodeList([
n for n in pool.where(choice) if not n == pool.root
]) # we won't touch the root
if not (len(needy_nodes)):
sys.exit("Nothing to do.")
if not (userinput(
"Once the solvent has been removed, further refinement of the pool is not possible. This includes the generation of unrestrained transition nodes! Continue?",
"bool")):
sys.exit("Quit by user.")
assert (len(needy_nodes) == len(needy_nodes.multilock())
) # make sure we lock ALL nodes
try:
for n in needy_nodes:
discard_solvent(n, "pdb")
discard_solvent(n, "trr")
for n in needy_nodes:
n.unlock()
except:
traceback.print_exc()
def main():
pool = Pool()
if(len(pool.where("state == 'grompp-able'")) > 0):
call_grompp(pool)
if(len(pool.where("state == 'em-grompp-able'")) > 0):
assert(path.exists("em.mdp")) #TODO that's not super nice yet
call_grompp(pool, mode='em')
def main():
options = options_desc.parse_args(sys.argv)[0]
zgf_cleanup.main()
pool = Pool()
not_reweightable = "isa_partition and state not in ('converged'"
if (options.ignore_convergence):
not_reweightable += ",'not-converged'"
if (options.ignore_failed):
not_reweightable += ",'mdrun-failed'"
not_reweightable += ")"
if pool.where(not_reweightable):
print "Pool can not be reweighted due to the following nodes:"
for bad_guy in pool.where(not_reweightable):
print "Node %s with state %s." % (bad_guy.name, bad_guy.state)
sys.exit("Aborting.")
active_nodes = pool.where("isa_partition and state != 'mdrun-failed'")
assert (len(active_nodes) == len(active_nodes.multilock())
) # make sure we lock ALL nodes
if (options.check_restraint):
for n in active_nodes:
check_restraint_energy(n)
if (options.method == "direct"):
reweight_direct(active_nodes, options)
elif (options.method == "entropy"):
reweight_entropy(active_nodes, options)
elif (options.method == "presampling"):
reweight_presampling(active_nodes, options)
else:
raise (Exception("Method unkown: " + options.method))
weight_sum = np.sum([n.tmp['weight'] for n in active_nodes])
print "Thermodynamic weights calculated by method '%s':" % options.method
for n in active_nodes:
n.obs.weight_direct = n.tmp['weight'] / weight_sum
if (options.method == "direct"):
print(
" %s with mean_V: %f [kJ/mol], %d refpoints and weight: %f" %
(n.name, n.obs.mean_V, n.tmp['n_refpoints'],
n.obs.weight_direct))
else:
print(" %s with A: %f [kJ/mol] and weight: %f" %
(n.name, n.obs.A, n.obs.weight_direct))
print "The above weighting uses bonded energies='%s' and nonbonded energies='%s'." % (
options.e_bonded, options.e_nonbonded)
for n in active_nodes:
n.save()
active_nodes.unlock()
def main():
options = options_desc.parse_args(sys.argv)[0]
zgf_cleanup.main()
pool = Pool()
#not_reweightable = "state not in ('refined','converged')"
not_reweightable = "isa_partition and state!='converged'"
if options.ignore_convergence:
not_reweightable = "isa_partition and state not in ('converged','not-converged')"
if pool.where(not_reweightable):
print "Pool can not be reweighted due to the following nodes:"
for bad_guy in pool.where(not_reweightable):
print "Node %s with state %s."%(bad_guy.name, bad_guy.state)
sys.exit("Aborting.")
active_nodes = pool.where("isa_partition")
assert(len(active_nodes) == len(active_nodes.multilock())) # make sure we lock ALL nodes
for n in active_nodes:
check_restraint_energy(n)
# find out about number of energygrps
mdp_file = gromacs.read_mdp_file(pool.mdp_fn)
energygrps = [str(egrp) for egrp in re.findall('[\S]+', mdp_file["energygrps"])]
moi_energies = True
if len(energygrps) < 2:
moi_energies = False # Gromacs energies are named differently when there are less than two energygrps :(
if(options.method == "direct"):
reweight_direct(active_nodes, moi_energies, options.sol_energy, options.save_refpoints)
elif(options.method == "entropy"):
reweight_entropy(active_nodes, moi_energies, options.sol_energy, options.save_refpoints)
elif(options.method == "presampling"):
reweight_presampling(active_nodes, options.presamp_temp, moi_energies, options.sol_energy)
else:
raise(Exception("Method unkown: "+options.method))
weight_sum = np.sum([n.tmp['weight'] for n in active_nodes])
print "Thermodynamic weights calculated by method '%s' (sol-energy=%s):"%(options.method, options.sol_energy)
for n in active_nodes:
n.obs.weight_direct = n.tmp['weight'] / weight_sum
if(options.method == "direct"):
print(" %s with mean_V: %f [kJ/mol], %d refpoints and weight: %f" % (n.name, n.obs.mean_V, n.tmp['n_refpoints'], n.obs.weight_direct))
else:
print(" %s with A: %f [kJ/mol] and weight: %f" % (n.name, n.obs.A, n.obs.weight_direct))
for n in active_nodes:
n.save()
active_nodes.unlock()
def main():
pool = Pool()
needy_nodes = pool.where("state == 'grompp-able'").multilock()
try:
for n in needy_nodes:
call_grompp(n)
except:
traceback.print_exc()
for n in needy_nodes:
n.unlock()
def main():
options = options_desc.parse_args(sys.argv)[0]
pool = Pool()
needy_nodes = pool.where("state == 'grompp-able'")
assert (len(needy_nodes) == len(needy_nodes.multilock())
) # make sure we lock ALL nodes
if (options.solv_model == "tip3p"):
solv_box = "spc216.gro"
solv_fn = "tip3p.itp"
elif (options.solv_model == "tip4p"):
solv_box = "tip4p.gro"
solv_fn = "tip4p.itp"
elif (options.solv_model == "tip4pew"):
solv_box = "tip4p.gro"
solv_fn = "tip4pew.itp"
elif (options.solv_model == "tip5"):
solv_box = "tip5p.gro"
solv_fn = "tip5p.itp"
elif (options.solv_model == "spc"):
solv_box = "spc216.gro"
solv_fn = "spc.itp"
elif (options.solv_model == "spce"):
solv_box = "spc216.gro"
solv_fn = "spce.itp"
elif (
options.solv_model == "acetonitrile"
): # TODO one might change this one to "custom" and let user enter name of template box
solv_box = "acetonitrile.pdb"
msg = "Topology update for acetonitrile is not supported. Proceed?"
if not (userinput(msg, "bool")):
for n in needy_nodes:
n.unlock()
return ("Quit by user.")
# determine maximum length of linears, if any
max_linear = query_linear_length(pool)
# make box and fill with solvent
genbox(pool, max_linear, options.bt,
(options.box_x, options.box_y, options.box_z), solv_box)
# update topology files (add solvent model and ions includes)
if not (options.solv_model == "acetonitrile"):
update_tops(pool, solv_fn)
for n in needy_nodes:
n.state = "em-grompp-able"
zgf_grompp.call_grompp(
n, mdp_file=options.grompp, final_state="em-mdrun-able"
) # re-grompp to get a tpr for energy minimization
n.unlock()
def main():
pool = Pool()
needy_nodes = pool.where("state == 'grompp-able'").multilock()
try:
for n in needy_nodes:
call_grompp(n)
except:
traceback.print_exc()
for n in needy_nodes:
n.unlock()
def main():
options = options_desc.parse_args(sys.argv)[0]
zgf_cleanup.main()
pool = Pool()
not_reweightable = "isa_partition and state not in ('converged'"
if(options.ignore_convergence):
not_reweightable += ",'not-converged'"
if(options.ignore_failed):
not_reweightable += ",'mdrun-failed'"
not_reweightable += ")"
if pool.where(not_reweightable):
print "Pool can not be reweighted due to the following nodes:"
for bad_guy in pool.where(not_reweightable):
print "Node %s with state %s."%(bad_guy.name, bad_guy.state)
sys.exit("Aborting.")
active_nodes = pool.where("isa_partition and state != 'mdrun-failed'")
assert(len(active_nodes) == len(active_nodes.multilock())) # make sure we lock ALL nodes
if(options.check_restraint):
for n in active_nodes:
check_restraint_energy(n)
if(options.method == "direct"):
reweight_direct(active_nodes, options)
elif(options.method == "entropy"):
reweight_entropy(active_nodes, options)
elif(options.method == "presampling"):
reweight_presampling(active_nodes, options)
else:
raise(Exception("Method unkown: "+options.method))
weight_sum = np.sum([n.tmp['weight'] for n in active_nodes])
print "Thermodynamic weights calculated by method '%s':"%options.method
for n in active_nodes:
n.obs.weight_direct = n.tmp['weight'] / weight_sum
if(options.method == "direct"):
print(" %s with mean_V: %f [kJ/mol], %d refpoints and weight: %f" % (n.name, n.obs.mean_V, n.tmp['n_refpoints'], n.obs.weight_direct))
else:
print(" %s with A: %f [kJ/mol] and weight: %f" % (n.name, n.obs.A, n.obs.weight_direct))
print "The above weighting uses bonded energies='%s' and nonbonded energies='%s'."%(options.e_bonded, options.e_nonbonded)
for n in active_nodes:
n.save()
active_nodes.unlock()
def main():
options = options_desc.parse_args(sys.argv)[0]
#TODO put somehow into Options, e.g. min_value=1 or required=True
if(not options.doomed_nodes):
sys.exit("Option --doomed_nodes is required.")
pool = Pool()
old_pool_size = len(pool)
old_alpha = pool.alpha
doomed_nodes = NodeList()
#TODO: maybe this code should go into ZIBMolPy.ui
for name in options.doomed_nodes.split(","):
found = [n for n in pool if n.name == name]
if(len(found) != 1):
sys.exit("Coult not find node '%s'"%(name))
doomed_nodes.append(found[0])
for n in doomed_nodes:
if(n == pool.root):
sys.exit("Node %s is the root. Removal not allowed."%(n.name))
#if(len(n.children) > 0):
# sys.exit("Node %s has children. Removal not allowed."%(n.name)) #TODO why should we forbid this?
if not(userinput("The selected node(s) will be removed permanently. Continue?", "bool")):
sys.exit("Quit by user.")
assert(len(doomed_nodes) == len(doomed_nodes.multilock()))
for n in doomed_nodes:
print("Removing directory: "+n.dir)
shutil.rmtree(n.dir)
pool.reload_nodes()
#TODO: this code-block also exists in zgf_create_node
if(len(pool.where("isa_partition")) < 2):
pool.alpha = None
elif(options.methodalphas == "theta"):
pool.alpha = zgf_create_nodes.calc_alpha_theta(pool)
elif(options.methodalphas == "user"):
pool.alpha = userinput("Please enter a value for alpha", "float")
else:
raise(Exception("Method unkown: "+options.methodalphas))
pool.history.append({'removed_nodes': [(n.name, n.state) for n in doomed_nodes], 'size':old_pool_size, 'alpha':old_alpha, 'timestamp':datetime.now()})
pool.save()
#TODO: deal with analysis dir and dependencies
zgf_cleanup.main()
def main():
pool = Pool()
needy_nodes = pool.where("state == 'created'")
assert(len(needy_nodes) == len(needy_nodes.multilock())) # make sure we lock ALL nodes
extract_frames(pool)
generate_topology(pool)
generate_mdp(pool)
for n in needy_nodes:
n.state = "grompp-able"
n.save()
n.unlock()
def main():
options = options_desc.parse_args(sys.argv)[0]
zgf_cleanup.main()
pool = Pool()
needy_nodes = pool.where("state == '%s'"%options.current_state).multilock()
for n in needy_nodes:
print "Recovering node %s with state %s to state %s ..."%(n.name, n.state, options.recover_state)
n.state = options.recover_state
n.save()
n.unlock()
def main():
pool = Pool()
needy_nodes = pool.where("state == 'created'")
assert (len(needy_nodes) == len(needy_nodes.multilock())
) # make sure we lock ALL nodes
extract_frames(pool)
generate_topology(pool)
generate_mdp(pool)
for n in needy_nodes:
n.state = "grompp-able"
n.save()
n.unlock()
def main():
print("Cleaning up.")
#check locks
pool = Pool()
for n in pool:
sys.stdout.write("Node %s is %s and " % (n.name, n.state))
if (not n.is_locked):
print("not locked.")
elif (n.is_lock_valid):
print("is locked and valid.")
else:
print("its lock is stale - removing it.")
os.remove(n.lock_fn)
#check zgf-dep files
files = []
for root, _, names in os.walk(os.getcwd()):
for n in names:
if (not n.endswith(".zgf-dep")):
continue
dep = path.join(root, n)
fn = dep2fn(dep)
if (not path.exists(fn)):
print("Removing lonely dep-file: " + dep)
os.remove(dep)
continue
files.append(
fn
) #in extra loop - otherwise check_files and remove(dep) collide
for fn in files:
check_file(fn)
def main():
options = options_desc.parse_args(sys.argv)[0]
pool = Pool()
needy_nodes = pool.where("state == 'em-mdrun-able'")
assert(len(needy_nodes) == len(needy_nodes.multilock())) # make sure we lock ALL nodes
# add ions to simulation boxes
call_genion(pool, options.np, options.pname, options.nn, options.nname, options.random_seed)
for n in needy_nodes:
n.state = "em-grompp-able"
zgf_grompp.call_grompp(n, mdp_file=options.grompp, final_state="em-mdrun-able") # re-grompp to get a tpr for energy minimization
n.unlock()
def __new__(cls, name=None):
""" Instanciates a node from a file, and (name=None) creates a new node
Caution: Filenames are always given relative to the root-dir
When no name is given, a new node is created. """
if (name != None and cls._instances.has_key(name)):
return (cls._instances[name])
if (name == None): # a new node, lets find a name
for i in itertools.count(0):
name = "node%.4d" % i
if (cls._instances.has_key(name)):
continue # new nodes might not been saved, yet
if (path.exists("./nodes/" + name)): continue
break
self = object.__new__(cls)
cls._instances[name] = self
#actuall init-code
from ZIBMolPy.pool import Pool #avoids circular imports
self._pool = Pool() #Pool is a singleton
self._name = name
self._tmp = Store() #for thing that need to be stored temporarly
self._obs = Store()
self.parent = None
if (path.exists(self.dir)):
self.reload()
#self.pool.append(self) #register with pool
return (self)
def main():
options = options_desc.parse_args(sys.argv)[0]
outfile = open(options.outfile,"w")
pool = Pool()
needy_nodes = pool.where("isa_partition and state not in ('refined','mdrun-failed')")
for n in needy_nodes:
outfile.write("%s, state: '%s':\n"%(n.name,n.state))
outfile.write(str(n.internals.array)+"\n")
outfile.write("mean pot.: %f, std pot.: %f, free energy estimate: %f\n"%(n.obs.mean_V,n.obs.std_V,n.obs.A))
outfile.write("#========================================================================#\n")
outfile.close()
print "Pool info was written to %s."%options.outfile
def main():
options = options_desc.parse_args(sys.argv)[0]
zgf_cleanup.main()
pool = Pool()
needy_nodes = pool.where("state == '%s'" %
options.current_state).multilock()
for n in needy_nodes:
print "Recovering node %s with state %s to state %s ..." % (
n.name, n.state, options.recover_state)
n.state = options.recover_state
n.save()
n.unlock()
def main():
options = options_desc.parse_args(sys.argv)[0]
pool = Pool()
needy_nodes = pool.where("state == 'grompp-able'")
assert(len(needy_nodes) == len(needy_nodes.multilock())) # make sure we lock ALL nodes
if(options.solv_model == "tip3p"):
solv_box = "spc216.gro"
solv_fn = "tip3p.itp"
elif(options.solv_model == "tip4p"):
solv_box = "tip4p.gro"
solv_fn = "tip4p.itp"
elif(options.solv_model == "tip4pew"):
solv_box = "tip4p.gro"
solv_fn = "tip4pew.itp"
elif(options.solv_model == "tip5"):
solv_box = "tip5p.gro"
solv_fn = "tip5p.itp"
elif(options.solv_model == "spc"):
solv_box = "spc216.gro"
solv_fn = "spc.itp"
elif(options.solv_model == "spce"):
solv_box = "spc216.gro"
solv_fn = "spce.itp"
elif(options.solv_model == "acetonitrile"): # TODO one might change this one to "custom" and let user enter name of template box
solv_box = "acetonitrile.pdb"
msg = "Topology update for acetonitrile is not supported. Proceed?"
if not(userinput(msg, "bool")):
for n in needy_nodes:
n.unlock()
return("Quit by user.")
# determine maximum length of linears, if any
max_linear = query_linear_length(pool)
# make box and fill with solvent
genbox(pool, max_linear, options.bt, (options.box_x, options.box_y, options.box_z), solv_box)
# update topology files (add solvent model and ions includes)
if not(options.solv_model == "acetonitrile"):
update_tops(pool, solv_fn)
for n in needy_nodes:
n.state = "em-grompp-able"
zgf_grompp.call_grompp(n, mdp_file=options.grompp, final_state="em-mdrun-able") # re-grompp to get a tpr for energy minimization
n.unlock()
def main():
options = options_desc.parse_args(sys.argv)[0]
#TODO put somehow into Options, e.g. min_value=1 or required=True
if(not options.doomed_nodes):
sys.exit("Option --doomed_nodes is required.")
pool = Pool()
old_pool_size = len(pool)
old_alpha = pool.alpha
doomed_nodes = NodeList()
#TODO: maybe this code should go into ZIBMolPy.ui
for name in options.doomed_nodes.split(","):
found = [n for n in pool if n.name == name]
if(len(found) != 1):
sys.exit("Coult not find node '%s'"%(name))
doomed_nodes.append(found[0])
for n in doomed_nodes:
if(n == pool.root):
sys.exit("Node %s is the root. Removal not allowed."%(n.name))
#if(len(n.children) > 0):
# sys.exit("Node %s has children. Removal not allowed."%(n.name)) #TODO why should we forbid this?
if not(userinput("The selected node(s) will be removed permanently. Continue?", "bool")):
sys.exit("Quit by user.")
assert(len(doomed_nodes) == len(doomed_nodes.multilock()))
for n in doomed_nodes:
print("Removing directory: "+n.dir)
shutil.rmtree(n.dir)
pool.reload_nodes()
#TODO: this code-block also exists in zgf_create_node
if(len(pool.where("isa_partition")) < 2):
pool.alpha = None
elif(options.methodalphas == "theta"):
pool.alpha = zgf_create_nodes.calc_alpha_theta(pool)
elif(options.methodalphas == "user"):
pool.alpha = userinput("Please enter a value for alpha", "float")
else:
raise(Exception("Method unkown: "+options.methodalphas))
pool.history.append({'removed_nodes': [(n.name, n.state) for n in doomed_nodes], 'size':old_pool_size, 'alpha':old_alpha, 'timestamp':datetime.now()})
pool.save()
#TODO: deal with analysis dir and dependencies
zgf_cleanup.main()
def main():
options = options_desc.parse_args(sys.argv)[0]
pool = Pool()
needy_nodes = pool.where("state == 'em-mdrun-able'")
assert (len(needy_nodes) == len(needy_nodes.multilock())
) # make sure we lock ALL nodes
# add ions to simulation boxes
call_genion(pool, options.np, options.pname, options.nn, options.nname,
options.random_seed)
for n in needy_nodes:
n.state = "em-grompp-able"
zgf_grompp.call_grompp(
n, mdp_file=options.grompp, final_state="em-mdrun-able"
) # re-grompp to get a tpr for energy minimization
n.unlock()
def main():
options = options_desc.parse_args(sys.argv)[0]
pool = Pool()
needy_nodes = pool.where("state == 'em-mdrun-able'")
assert(len(needy_nodes) == len(needy_nodes.multilock())) # make sure we lock ALL nodes
# add ions to simulation boxes
call_genion(pool, options.np, options.pname, options.nn, options.nname, options.random_seed)
for n in needy_nodes:
n.state = "em-grompp-able"
n.save()
n.unlock()
zgf_grompp.main()
def main():
options = options_desc.parse_args(sys.argv)[0]
outfile = open(options.outfile, "w")
pool = Pool()
needy_nodes = pool.where(
"isa_partition and state not in ('refined','mdrun-failed')")
for n in needy_nodes:
outfile.write("%s, state: '%s':\n" % (n.name, n.state))
outfile.write(str(n.internals.array) + "\n")
outfile.write(
"mean pot.: %f, std pot.: %f, free energy estimate: %f\n" %
(n.obs.mean_V, n.obs.std_V, n.obs.A))
outfile.write(
"#========================================================================#\n"
)
outfile.close()
print "Pool info was written to %s." % options.outfile
def main(argv=None):
if (argv == None):
argv = sys.argv
options = options_desc.parse_args(argv)[0]
assert (not (options.refine_all and options.extend_all))
pool = Pool()
needy_nodes = pool.where("isa_partition and is_sampled").multilock()
# 1. Trying to detect fake convergence
for n in pool.where("state == 'converged'"):
means = kmeans(n.trajectory, k=2)
d = (means[0] - means[1]).norm2()
if (d > 2.0 and (options.refine_all or userinput(
"%s has converged but appears to have a bimodal distribution.\nDo you want to refine?"
% n.name,
"bool"))): #TODO decide upon threshold (per coordinate?)
refine(n, options)
# 2. Dealing with not-converged nodes
for n in pool.where("state == 'not-converged'"):
if (not (options.refine_all or options.extend_all)):
choice = userchoice(
"%s has not converged. What do you want to do?" % n.name,
['_refine', '_extend', '_ignore'])
if (options.refine_all or choice == "r"):
refine(n, options)
elif (options.extend_all or choice == "e"):
extend(n)
elif (choice == "i"):
continue
for n in needy_nodes:
n.save()
n.unlock()
zgf_setup_nodes.main()
zgf_grompp.main()
zgf_cleanup.main()
def main():
options = options_desc.parse_args(sys.argv)[0]
pool = Pool()
needy_nodes = pool.where("state == 'grompp-able'")
assert(len(needy_nodes) == len(needy_nodes.multilock())) # make sure we lock ALL nodes
if(options.solv_model == "tip3p"):
solv_box = "spc216.gro"
solv_fn = "tip3p.itp"
elif(options.solv_model == "tip4p"):
solv_box = "tip4p.gro"
solv_fn = "tip4p.itp"
elif(options.solv_model == "tip4pew"):
solv_box = "tip4p.gro"
solv_fn = "tip4pew.itp"
elif(options.solv_model == "tip5"):
solv_box = "tip5p.gro"
solv_fn = "tip5p.itp"
elif(options.solv_model == "spc"):
solv_box = "spc216.gro"
solv_fn = "spc.itp"
elif(options.solv_model == "spce"):
solv_box = "spc216.gro"
solv_fn = "spce.itp"
# determine maximum length of linears, if any
max_linear = query_linear_length(pool)
# make box and fill with solvent
genbox(pool, max_linear, options.bt, (options.box_x, options.box_y, options.box_z), solv_box)
# update topology files (add solvent model and ions includes)
update_tops(pool, solv_fn)
for n in needy_nodes:
n.state = "em-grompp-able"
n.save()
n.unlock()
def main():
options = options_desc.parse_args(sys.argv)[0]
pool = Pool()
if(options.convtest):
for n in pool.where("state in ('converged', 'not-converged')"):
print("\n\nRunning Gelman-Rubin on %s"%n)
conv_check_gelman_rubin(n)
return # exit
auto_refines_counter = 0
while(True):
pool.reload()
pool.reload_nodes()
for n in pool:
n.reload()
active_node = None
for n in pool.where("state in ('em-mdrun-able', 'mdrun-able', 'rerun-able-converged', 'rerun-able-not-converged')"):
if(n.lock()):
active_node = n
break
if(active_node == None):
if(auto_refines_counter < options.auto_refines):
auto_refines_counter += 1
print("\n\nRunning 'zgf_refine --refine-all' for the %d time..."%auto_refines_counter)
zgf_refine.main(["--refine-all"])
continue
else:
break # we're done - exit
try:
process(active_node, options)
active_node.save()
active_node.unlock()
except:
print "MDRUN FAILED"
active_node.state = "mdrun-failed"
active_node.save()
active_node.unlock()
traceback.print_exc()
continue
def main():
pool = Pool()
for n in pool.where("isa_partition"):
for cn in n.children.where("is_sampled"):
print cn.trajectory
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 is_applicable():
pool = Pool()
return(len(pool.where("state == 'created'")) > 0)
def main(argv=None):
if(argv==None):
argv = sys.argv
options = options_desc.parse_args(argv)[0]
print("Options:\n%s\n"%pformat(eval(str(options))))
if(options.random_seed):
# using numpy-random because python-random differs beetween 32 and 64 bit
np.random.seed(hash(options.random_seed))
pool = Pool()
old_pool_size = len(pool)
print "pool", pool
if(options.parent_node == "root"):
parent = pool.root
else:
found = [n for n in pool if n.name == options.parent_node]
assert(len(found) == 1)
parent = found[0]
print "### Generate nodes: %s ###" % options.methodnodes
if(options.methodnodes == "kmeans"):
chosen_idx = mknodes_kmeans(parent, options.numnodes)
elif(options.methodnodes == "equidist"):
chosen_idx = mknodes_equidist(parent, options.numnodes)
elif(options.methodnodes == "maxdist"):
chosen_idx = mknodes_maxdist(parent, options.numnodes)
elif(options.methodnodes == "all"):
chosen_idx = mknodes_all(parent)
else:
raise(Exception("Method unknown: "+options.methodnodes))
chosen_idx.sort() # makes preview-trajectory easier to understand
if(options.write_preview):
write_node_preview(pool, parent, chosen_idx)
for i in chosen_idx:
n = Node()
n.parent_frame_num = i
n.parent = parent
n.state = "creating-a-partition" # will be set to "created" at end of script
n.extensions_counter = 0
n.extensions_max = options.ext_max
n.extensions_length = options.ext_length
n.sampling_length = options.sampling_length
n.internals = parent.trajectory.getframe(i)
pool.append(n)
print "\n### Obtain alpha: %s ###" % options.methodalphas
old_alpha = pool.alpha
if(options.methodalphas == "theta"):
pool.alpha = calc_alpha_theta(pool)
elif(options.methodalphas == "user"):
pool.alpha = userinput("Please enter a value for alpha", "float")
else:
raise(Exception("Method unknown: "+options.methodalphas))
pool.history.append({'refined_node': (parent.name, parent.state), 'size':old_pool_size, 'alpha':old_alpha, 'timestamp':datetime.now()})
pool.save() # alpha might have changed
print "\n### Obtain phi fit: %s ###" % options.methodphifit
if(options.methodphifit == "harmonic"):
do_phifit_harmonic(pool)
elif(options.methodphifit == "switch"):
do_phifit_switch(pool)
elif(options.methodphifit == "leastsq"):
do_phifit_leastsq(pool)
else:
raise(Exception("Method unkown: "+options.methodphifit))
for n in pool.where("state == 'creating-a-partition'"):
n.state = "created"
n.save()
print "saving " +str(n)
zgf_cleanup.main()
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 is_applicable():
pool = Pool()
return( len(pool.where("'weight_direct' in obs")) > 0 and len(pool.where("isa_partition and 'weight_direct' not in obs")) == 0 )
def is_applicable():
pool = Pool()
return( len(pool.where("state in ('converged', 'not-converged', 'refined')")) > 1 )
def is_applicable():
pool = Pool()
return(len(pool.where("state=='mdrun-able'")) > 0)
def is_applicable():
pool = Pool()
return(len(pool.where("state in ('em-mdrun-able', 'mdrun-able', 'converged', 'not-converged', 'rerun-able-converged', 'rerun-able-not-converged')")) > 0)
def main():
options = options_desc.parse_args(sys.argv)[0]
pool = Pool()
if options.ignore_convergence:
needy_nodes = pool.where("state in ('converged','not-converged')")
else:
needy_nodes = pool.where("state == 'converged'")
assert(len(needy_nodes) == len(needy_nodes.multilock())) # make sure we lock ALL nodes
for node in needy_nodes:
if ( path.exists(node.dir+"/rerun_me.trr") and path.exists(node.dir+"/rerun_me.pdb") and path.exists(node.dir+"/rerun_me.top") and path.exists(node.dir+"/rerun_me.tpr") ):
print("All four rerun files (rerun_me.*) already existing in " + node.dir + ".")
print("Be sure you want to keep them!")
continue
# if "none", assume that sim is implicit or in vacuum. thus, trjconv not required.
if options.pbc_removal != "none":
# desolvate trr
if not( path.exists(node.dir+"/rerun_me.trr")):
cmd = ["trjconv", "-f", node.trr_fn, "-o", node.dir+"/rerun_me.trr", "-s", node.tpr_fn, "-n", node.pool.ndx_fn, "-pbc", options.pbc_removal]
print("Calling: "+(" ".join(cmd)))
p = Popen(cmd, stdin=PIPE)
p.communicate(input=("MOI\n"))
assert(p.wait() == 0)
else:
print("Rerun trajectory file (rerun_me.trr) already existing in " + node.dir + ".")
print("Be sure you want to keep it!")
#sys.exit(0)
#continue
# desolvate pdb
if not( path.exists(node.dir+"/rerun_me.pdb") ):
cmd = ["trjconv", "-f", node.pdb_fn, "-o", node.dir+"/rerun_me.pdb", "-s", node.tpr_fn, "-n", node.pool.ndx_fn, "-pbc", options.pbc_removal]
print("Calling: "+(" ".join(cmd)))
p = Popen(cmd, stdin=PIPE)
p.communicate(input=("MOI\n"))
assert(p.wait() == 0)
# desolvate topology
infile = open(node.top_fn, "r").readlines()
mol_section = False
out_top = []
for line in infile:
if( re.match("\s*\[\s*(molecules)\s*\]\s*", line.lower()) ):
# we are past the "molecules" section
mol_section = True
if(mol_section):
# comment out lines that belong to solvent (SOL, CL, NA)... add more if necessary
if( re.match("\s*(sol|cl|na|tsl|tcm|mth)\s*\d+", line.lower()) ):
line = ";"+line
out_top.append(line)
outfile = open(node.dir+"/rerun_me.top","w").writelines(out_top)
else:
if not( path.exists(node.dir+"/rerun_me.trr") ):
symlink(node.name+".trr", node.dir+"/rerun_me.trr")
else:
print("Rerun trajectory file (rerun_me.trr) already existing in " + node.dir + ".")
print("Be sure you want to keep it!")
#continue
if not( path.exists(node.dir+"/rerun_me.pdb") ):
symlink(node.name+"_conf.pdb", node.dir+"/rerun_me.pdb")
if not( path.exists(node.dir+"/rerun_me.top") ):
symlink(node.name+".top", node.dir+"/rerun_me.top")
grompp2state = "rerun-able-"+node.state
# get rid of old checkpoint file (it might mess up the rerun)
if( path.exists(node.dir+"/state.cpt") ):
remove(node.dir+"/state.cpt")
#zgf_grompp.call_grompp(node, mdp_file=options.grompp, final_state=grompp2state)
#TODO code borrowed from zgf_grompp
#TODO make the original method fit for grompping reruns
if not( path.exists(node.dir+"/rerun_me.trr") ):
cmd = ["grompp"]
cmd += ["-f", "../../"+options.grompp]
cmd += ["-n", "../../"+node.pool.ndx_fn]
cmd += ["-c", "../../"+node.dir+"/rerun_me.pdb"]
cmd += ["-p", "../../"+node.dir+"/rerun_me.top"]
cmd += ["-o", "../../"+node.dir+"/rerun_me.tpr"]
print("Calling: %s"%" ".join(cmd))
p = Popen(cmd, cwd=node.dir)
retcode = p.wait()
assert(retcode == 0) # grompp should never fail
node.state = grompp2state
node.save()
node.unlock()
def is_applicable():
pool = Pool()
return (len(pool.where("state == 'created'")) > 0)
def is_applicable():
pool = Pool()
return (len(pool) > 1 and len(
pool.where(
"isa_partition and state in ('converged','not-converged','mdrun-failed')"
)) == len(pool.where("isa_partition")))
def main():
options = options_desc.parse_args(sys.argv)[0]
pool = Pool()
if options.ignore_convergence:
needy_nodes = pool.where("state in ('converged','not-converged')")
else:
needy_nodes = pool.where("state == 'converged'")
assert (len(needy_nodes) == len(needy_nodes.multilock())
) # make sure we lock ALL nodes
for node in needy_nodes:
if (path.exists(node.dir + "/rerun_me.trr")
and path.exists(node.dir + "/rerun_me.pdb")
and path.exists(node.dir + "/rerun_me.top")
and path.exists(node.dir + "/rerun_me.tpr")):
print("All four rerun files (rerun_me.*) already existing in " +
node.dir + ".")
print("Be sure you want to keep them!")
continue
# if "none", assume that sim is implicit or in vacuum. thus, trjconv not required.
if options.pbc_removal != "none":
# desolvate trr
if not (path.exists(node.dir + "/rerun_me.trr")):
cmd = [
"trjconv", "-f", node.trr_fn, "-o",
node.dir + "/rerun_me.trr", "-s", node.tpr_fn, "-n",
node.pool.ndx_fn, "-pbc", options.pbc_removal
]
print("Calling: " + (" ".join(cmd)))
p = Popen(cmd, stdin=PIPE)
p.communicate(input=("MOI\n"))
assert (p.wait() == 0)
else:
print(
"Rerun trajectory file (rerun_me.trr) already existing in "
+ node.dir + ".")
print("Be sure you want to keep it!")
#sys.exit(0)
#continue
# desolvate pdb
if not (path.exists(node.dir + "/rerun_me.pdb")):
cmd = [
"trjconv", "-f", node.pdb_fn, "-o",
node.dir + "/rerun_me.pdb", "-s", node.tpr_fn, "-n",
node.pool.ndx_fn, "-pbc", options.pbc_removal
]
print("Calling: " + (" ".join(cmd)))
p = Popen(cmd, stdin=PIPE)
p.communicate(input=("MOI\n"))
assert (p.wait() == 0)
# desolvate topology
infile = open(node.top_fn, "r").readlines()
mol_section = False
out_top = []
for line in infile:
if (re.match("\s*\[\s*(molecules)\s*\]\s*", line.lower())):
# we are past the "molecules" section
mol_section = True
if (mol_section):
# comment out lines that belong to solvent (SOL, CL, NA)... add more if necessary
if (re.match("\s*(sol|cl|na|tsl|tcm|mth)\s*\d+",
line.lower())):
line = ";" + line
out_top.append(line)
outfile = open(node.dir + "/rerun_me.top", "w").writelines(out_top)
else:
if not (path.exists(node.dir + "/rerun_me.trr")):
symlink(node.name + ".trr", node.dir + "/rerun_me.trr")
else:
print(
"Rerun trajectory file (rerun_me.trr) already existing in "
+ node.dir + ".")
print("Be sure you want to keep it!")
#continue
if not (path.exists(node.dir + "/rerun_me.pdb")):
symlink(node.name + "_conf.pdb", node.dir + "/rerun_me.pdb")
if not (path.exists(node.dir + "/rerun_me.top")):
symlink(node.name + ".top", node.dir + "/rerun_me.top")
grompp2state = "rerun-able-" + node.state
# get rid of old checkpoint file (it might mess up the rerun)
if (path.exists(node.dir + "/state.cpt")):
remove(node.dir + "/state.cpt")
#zgf_grompp.call_grompp(node, mdp_file=options.grompp, final_state=grompp2state)
#TODO code borrowed from zgf_grompp
#TODO make the original method fit for grompping reruns
if not (path.exists(node.dir + "/rerun_me.trr")):
cmd = ["grompp"]
cmd += ["-f", "../../" + options.grompp]
cmd += ["-n", "../../" + node.pool.ndx_fn]
cmd += ["-c", "../../" + node.dir + "/rerun_me.pdb"]
cmd += ["-p", "../../" + node.dir + "/rerun_me.top"]
cmd += ["-o", "../../" + node.dir + "/rerun_me.tpr"]
print("Calling: %s" % " ".join(cmd))
p = Popen(cmd, cwd=node.dir)
retcode = p.wait()
assert (retcode == 0) # grompp should never fail
node.state = grompp2state
node.save()
node.unlock()
def main():
options = options_desc.parse_args(sys.argv)[0]
zgf_cleanup.main()
print("Options:\n%s\n"%pformat(eval(str(options))))
pool = Pool()
parent = pool.root
active_nodes = pool.where("isa_partition")
assert(len(active_nodes) == len(active_nodes.multilock())) # make sure we lock ALL nodes
if active_nodes.where("'weight_direct' not in obs"):
sys.exit("Q-Matrix calculation not possible: Not all of the nodes have been reweighted.")
node_weights = np.array([node.obs.weight_direct for node in active_nodes])
print "### Generate bins: equidist ###"
result = q_equidist(parent, options.numnodes)
chosen_idx=result['chosen_idx']
frames_chosen=result['frames_chosen']
theta=result['theta']
chosen_idx.sort() # makes preview-trajectory easier to understand
dimension=len(chosen_idx)
print "chosen_idx"
print chosen_idx
print "### Generate bin weights ###"
bin_weights=np.zeros(dimension)
for (i,n) in enumerate(active_nodes):
w_denom = np.sum(n.frameweights)
for t in range(len(n.trajectory)):
diffs = (frames_chosen - n.trajectory.getframe(t)).norm()
j = np.argmin(diffs)
bin_weights[j] = bin_weights[j] + node_weights[i] * n.frameweights[t] / w_denom
print "bin_weights"
print bin_weights
print "### Generate q_all (entries only for neighboring bins) ###"
q_all = np.empty((dimension, dimension), dtype=np.float)
for i in range(dimension):
sum_row = 0.0
diffs = (frames_chosen - frames_chosen.getframe(i)).norm()
print "diffs"
print diffs
for j in range(dimension):
if (diffs[j] < 2.0 * theta) and (bin_weights[i] > 0.0):
q_all[i,j] = np.sqrt(bin_weights[j]) / np.sqrt(bin_weights[i])
sum_row = sum_row + q_all[i , j]
else:
q_all[i,j] = 0
q_all[i, i] = q_all[i, i]- sum_row
print "Q_All"
print q_all
if options.export_matlab:
savemat(pool.analysis_dir+"q_all.mat", {"q_all":q_all})
active_nodes.unlock()
zgf_cleanup.main()
def is_applicable():
pool = Pool()
return (len(pool.where("state == 'grompp-able'")) > 0)
def is_applicable():
pool = Pool()
return (len(pool.where("state == 'em-mdrun-able'")) > 0)
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 is_applicable():
pool = Pool()
return (path.exists(pool.chi_mat_fn))
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():
pool = Pool()
for n in pool.where("isa_partition"):
for cn in n.children.where("is_sampled"):
print cn.trajectory
def is_applicable():
pool = Pool()
return (pool.where("isa_partition and state != 'mdrun-failed'") > 0)
def main():
options = options_desc.parse_args(sys.argv)[0]
zgf_cleanup.main()
print("Options:\n%s\n"%pformat(eval(str(options))))
pool = Pool()
parent = pool.root
active_nodes = pool.where("isa_partition")
assert(len(active_nodes) == len(active_nodes.multilock())) # make sure we lock ALL nodes
if active_nodes.where("'weight_direct' not in obs"):
sys.exit("Q-Matrix calculation not possible: Not all of the nodes have been reweighted.")
node_weights = np.array([node.obs.weight_direct for node in active_nodes])
print "### Generate bins: equidist ###"
result = q_equidist(parent, options.numnodes)
chosen_idx=result['chosen_idx']
frames_chosen=result['frames_chosen']
theta=result['theta']
chosen_idx.sort() # makes preview-trajectory easier to understand
dimension=len(chosen_idx)
print "chosen_idx"
print chosen_idx
print "### Generate bin weights ###"
bin_weights=np.zeros(dimension)
for (i,n) in enumerate(active_nodes):
w_denom = np.sum(n.frameweights)
for t in range(len(n.trajectory)):
diffs = (frames_chosen - n.trajectory.getframe(t)).norm()
j = np.argmin(diffs)
bin_weights[j] = bin_weights[j] + node_weights[i] * n.frameweights[t] / w_denom
print "bin_weights"
print bin_weights
print "### Generate q_all (entries only for neighboring bins) ###"
q_all = np.empty((dimension, dimension), dtype=np.float)
for i in range(dimension):
sum_row = 0.0
diffs = (frames_chosen - frames_chosen.getframe(i)).norm()
print "diffs"
print diffs
for j in range(dimension):
if (diffs[j] < 2.0 * theta) and (bin_weights[i] > 0.0):
q_all[i,j] = np.sqrt(bin_weights[j]) / np.sqrt(bin_weights[i])
sum_row = sum_row + q_all[i , j]
else:
q_all[i,j] = 0
q_all[i, i] = q_all[i, i]- sum_row
print "Q_All"
print q_all
if options.export_matlab:
savemat(pool.analysis_dir+"q_all.mat", {"q_all":q_all})
active_nodes.unlock()
zgf_cleanup.main()
def is_applicable():
pool = Pool()
return (len(pool) > 0)
def is_applicable():
pool = Pool()
return( len(pool.where("'weight_direct' in obs")) > 0 and len(pool.where("isa_partition and 'weight_direct' not in obs")) == 0 )
def is_applicable():
pool = Pool()
return(len(pool.where("isa_partition and is_sampled")) > 0)
def is_applicable():
pool = Pool()
return( len(pool) > 1 and len(pool.where("isa_partition and state in ('converged','not-converged','mdrun-failed')")) == len(pool.where("isa_partition")) )
def is_applicable():
pool = Pool()
return (len(pool.where("isa_partition and is_sampled")) > 0)
