def main_noninteractive():
"""The main routine for running non-interactively."""
global imcftpl
args = parse_arguments()
set_loglevel(args.verbose)
log.info('Running in non-interactive mode.')
log.debug('Python FluoView package file: %s' % fv.__file__)
base = dirname(args.mosaiclog)
fname = basename(args.mosaiclog)
mosaics = fv.FluoViewMosaic(join(base, fname))
log.warn(gen_mosaic_details(mosaics))
if args.templates is not None:
imcftpl = args.templates
code = imagej.gen_stitching_macro_code(mosaics, 'templates/stitching',
path=base, tplpath=imcftpl)
if not args.dryrun:
log.info('Writing stitching macro.')
imagej.write_stitching_macro(code, fname='stitch_all.ijm', dname=base)
log.info('Writing tile configuration files.')
imagej.write_all_tile_configs(mosaics, fixsep=True)
log.info('Launching stitching macro.')
IJ.runMacro(flatten(code))
else:
log.info('Dry-run was selected. Printing generated macro:')
log.warn(flatten(code))
def period_spread_constraint2(schedule, exams, periods, institute_con):
"""Returns penalty
This constraint allows an organisation to 'spread' an schedule's examinations over a specified number of periods.
This can be thought of an extension of the two constraints previously described. Within the �Institutional Model
Index', a figure is provided relating to how many periods the solution should be �optimised' over.
"""
period_spread_constraints = institute_con[InstitutionalEnum.PERIODSPREAD]
period_lengths = period_spread_constraints[0].values if len(period_spread_constraints) > 0 else []
period_to_exam = get_period_to_exam_mapping(schedule, exams, periods)
period_to_students = dict()
for period, exams in period_to_exam.items():
period_to_students[period] = set(flatten(map(lambda exam: exam.students, exams)))
periods = sorted(period_to_students.keys())
violations = 0
for period_length in period_lengths:
for period_start in periods:
for period_step in range(0, period_length - 1):
period = period_start + period_step
if period in period_to_students:
students_in_period = period_to_students[period]
other_periods = range(period + 1, period_start + period_length)
f_get_students = lambda p: period_to_students[p] if p in period_to_students else []
students_in_other_periods = set(flatten(map(f_get_students, other_periods)))
intersect = students_in_period & students_in_other_periods
violations += len(intersect)
return violations
def __call__(self, x):
result_continuations = [ResultContinuation.start(x, self.chunks)]
pending_ks, done_ks = self.partition_ks(result_continuations)
while pending_ks:
new_ks = flatten(
k.new_ks() for k in pending_ks
)
result_continuations = done_ks + new_ks
pending_ks, done_ks = self.partition_ks(result_continuations)
return Multiple(*
flatten(k.make_multiple() for k in result_continuations)
)
def run(self, project):
self.apply_preset(project.presets)
# Compile objects first
for obj in self.objects:
obj.run(project)
log(
1, 'compile {} -> {}'.format(
str(flatten([o.get_files() for o in self.objects])),
self.output))
self.compiler.create_executable(
untempl(self.output, project.props),
flatten([obj.get_files() for obj in self.objects]),
self.linked_libs, self.params)
def plot_2d_stats(stats, name):
for ck in stats:
if ck != "ID":
for ak in stats[ck]:
context_values_keys = sorted(stats[ck][ak].keys())
context_values = dict(zip(context_values_keys, range(len(context_values_keys))))
action_values_keys = sorted(set(flatten(list(map(lambda x: x.keys(), stats[ck][ak].values())))))
action_values = dict(zip(action_values_keys, range(len(action_values_keys))))
ck_ak_stats = np.zeros((len(context_values), len(action_values)))
maximum = 0
for cv in sorted(stats[ck][ak]):
for av in sorted(stats[ck][ak][cv]):
ck_ak_stats[context_values[cv], action_values[av]] = stats[ck][ak][cv][av]["rate"]
maximum = max(maximum, stats[ck][ak][cv][av]["rate"])
plt.imshow(ck_ak_stats.T, interpolation="none")
plt.clim([0, maximum])
if ck_ak_stats.shape[0] > ck_ak_stats.shape[1]:
plt.colorbar(orientation="horizontal")
else:
plt.colorbar(orientation="vertical")
plt.xticks(list(range(len(context_values))), list(context_values_keys), rotation='vertical')
plt.yticks(list(range(len(action_values))), list(action_values_keys))
plt.xlabel(ck)
plt.ylabel(ak)
plt.title("Revenue / show")
dir = "stats/" + name + "/rate_interaction/"
create_directory(dir)
plt.savefig(dir + ck + "-" + ak)
plt.close()
def gaussians(centers, sizes, covs):
r = []
for i, c in enumerate(centers):
r.extend(gaussian(c, sizes[i], covs[i]))
colors = [[i] * sizes[i] for i in range(len(sizes))]
colors = misc.flatten(colors)
r = np.array(r)
return r, colors
def getArgCount(args):
argcount = len(args)
if args:
for arg in args:
if isinstance(arg, TupleArg):
numNames = len(misc.flatten(arg.names))
argcount = argcount - numNames
return argcount
def make_substitution(elem, env):
# Encloses its answer in a list, no matter what. This enables a .subst()
# method to transform a single item into multiple items.
if hasattr(elem, 'subst'):
return elem.subst(env)
elif is_listlike(elem):
return [flatten(make_substitution(e, env) for e in elem)]
else:
return [elem]
def __init__(self, *raw_feet):
# Here, "raw" means "We don't yet know which syllables are stressed."
self.raw_feet = raw_feet
raw_syllable_map = SyllableMap(
self._word_instances(self.raw_feet),
flatten(foot.syllables for foot in raw_feet)
)
self.feet = [
foot.add_stresses(raw_syllable_map) for foot in self.raw_feet
]
def generateArgList(arglist):
"""Generate an arg list marking TupleArgs"""
args = []
extra = []
count = 0
for i in range(len(arglist)):
elt = arglist[i]
if type(elt) == types.StringType:
args.append(elt)
elif type(elt) == types.TupleType:
args.append(TupleArg(i * 2, elt))
extra.extend(misc.flatten(elt))
count = count + 1
else:
raise ValueError, "unexpect argument type:", elt
return args + extra, count
def plot_2d_stats(stats, name):
for ck in stats:
if ck != "ID":
for ak in stats[ck]:
context_values_keys = sorted(stats[ck][ak].keys())
context_values = dict(
zip(context_values_keys, range(len(context_values_keys))))
action_values_keys = sorted(
set(
flatten(
list(
map(lambda x: x.keys(),
stats[ck][ak].values())))))
action_values = dict(
zip(action_values_keys, range(len(action_values_keys))))
ck_ak_stats = np.zeros(
(len(context_values), len(action_values)))
maximum = 0
for cv in sorted(stats[ck][ak]):
for av in sorted(stats[ck][ak][cv]):
ck_ak_stats[
context_values[cv],
action_values[av]] = stats[ck][ak][cv][av]["rate"]
maximum = max(maximum, stats[ck][ak][cv][av]["rate"])
plt.imshow(ck_ak_stats.T, interpolation="none")
plt.clim([0, maximum])
if ck_ak_stats.shape[0] > ck_ak_stats.shape[1]:
plt.colorbar(orientation="horizontal")
else:
plt.colorbar(orientation="vertical")
plt.xticks(list(range(len(context_values))),
list(context_values_keys),
rotation='vertical')
plt.yticks(list(range(len(action_values))),
list(action_values_keys))
plt.xlabel(ck)
plt.ylabel(ak)
plt.title("Revenue / show")
dir = "stats/" + name + "/rate_interaction/"
create_directory(dir)
plt.savefig(dir + ck + "-" + ak)
plt.close()
def generateArgList(arglist):
# Generate an arg list marking TupleArgs
args = []
extra = []
count = 0
marg = False
for i, elt in enumerate(arglist):
if type(elt) is str:
args.append(elt)
if elt[0] == '$':
marg = True
elif type(elt) is tuple:
args.append(TupleArg(i * 2, elt))
extra.extend(misc.flatten(elt))
count = count + 1
else:
raise ValueError, "unexpect argument type:", elt
return args + extra, count, marg
def generateArgList(arglist): # Generate an arg list marking TupleArgs args = [] extra = [] count = 0 marg = False for i, elt in enumerate (arglist): if type(elt) is str: args.append(elt) if elt [0] == '$': marg = True elif type(elt) is tuple: args.append(TupleArg(i * 2, elt)) extra.extend(misc.flatten(elt)) count = count + 1 else: raise ValueError, "unexpect argument type:", elt return args + extra, count, marg
def _substitutions(self, replacement, env):
return flatten([
self._substitution(elem, env)
for elem in replacement
])
def fill_in_variables(args, env):
return flatten([
env.get(arg.name) if isinstance(arg, Var) else [arg]
for arg in args
])
def word_instances_to_letters(word_instances):
return [wordBreak] + flatten(intersperse([" "], [w.to_letters for w in word_instances])) + [wordBreak]
def subst(self, env):
return [flatten(
make_substitution(elem, env)
for elem in self.seq
)]
def match_and_replace(self, x, env):
return flatten([
chunk.match_and_replace(x, env)
for chunk in self.chunks
])
def letters(self):
return flatten(foot.letters for foot in self.feet)
def matches(self, x, env):
return flatten(
chunk.matches(x, env)
for chunk in self.chunks
)
def clusters(self):
return flatten(sy.clusters for sy in self.syllables)
def main_interactive():
"""The main routine for running interactively."""
log.info('Running in interactive mode.')
(base, fname) = ui_get_input_file()
if (base is None):
return
log.warn("Parsing project file: %s" % (base + fname))
IJ.showStatus("Parsing experiment file...")
mosaics = fv.FluoViewMosaic(join(base, fname), runparser=False)
IJ.showStatus("Parsing mosaics...")
progress = 0.0
count = len(mosaics.mosaictrees)
step = 1.0 / count
for subtree in mosaics.mosaictrees:
IJ.showProgress(progress)
mosaics.add_mosaic(subtree)
progress += step
IJ.showProgress(progress)
IJ.showStatus("Parsed %i mosaics." % len(mosaics))
dialog = GenericDialog('FluoView OIF / OIB Stitcher')
if len(mosaics) == 0:
msg = ("Couldn't find any (valid) mosaics in the project file.\n"
" \n"
"Please make sure to have all files available!\n"
" \n"
"Will stop now.\n")
log.warn(msg)
dialog.addMessage(msg)
dialog.showDialog()
return
msg = "------------------------ EXPORT OPTIONS ------------------------"
dialog.addMessage(msg)
formats = ["OME-TIFF", "ICS/IDS"]
dialog.addChoice("Export Format", formats, formats[0])
dialog.addCheckbox("separate files by Z slices (OME-TIFF only)?", False)
msg = "------------------------ EXPORT OPTIONS ------------------------"
dialog.addMessage(msg)
dialog.addMessage("")
dialog.addMessage("")
msg = gen_mosaic_details(mosaics)
log.warn(msg)
msg += "\n \nPress [OK] to write tile configuration files\n"
msg += "and continue with running the stitcher."
dialog.addMessage(msg)
dialog.showDialog()
opts = {}
if dialog.getNextChoice() == 'ICS/IDS':
opts['export_format'] = '".ids"'
else:
opts['export_format'] = '".ome.tif"'
if dialog.getNextBoolean() == True:
opts['split_z_slices'] = 'true'
code = imagej.gen_stitching_macro_code(mosaics, 'templates/stitching',
path=base, tplpath=imcftpl, opts=opts)
log.warn("============= generated macro code =============")
log.warn(flatten(code))
log.warn("============= end of generated macro code =============")
if dialog.wasOKed():
log.warn('Writing stitching macro.')
imagej.write_stitching_macro(code, fname='stitch_all.ijm', dname=base)
log.warn('Writing tile configuration files.')
imagej.write_all_tile_configs(mosaics, fixsep=True)
log.warn('Launching stitching macro.')
IJ.runMacro(flatten(code))
def make_flat_restraints(self):
"""
searches for rings in the graph G, splits it in 4-member chunks and tests if
they are flat: volume of tetrahedron of chunk < 0.1 A-3.
Additionally, the ring adjacent atoms are added and new chunks created.
returns list of flat chunks.
>>> from dbfile import ParseDB
>>> gdb = ParseDB('../dsr_db.txt')
>>> res = Restraints('benzene', gdb)
>>> sorted(res.make_flat_restraints())
[['C1', 'C2', 'C5', 'C6'], ['C3', 'C4', 'C5', 'C6']]
"""
import networkx as nx
list_of_rings = nx.cycle_basis(self._G)
if not list_of_rings:
return False
flats = []
neighbors = []
newflats = []
for ring in list_of_rings:
for atom in ring:
# lets see if there is a neighboring atom:
nb = self._G.neighbors(atom) # [1:]
for i in nb:
if not i in flatten(list_of_rings):
neighbors.append(i)
if len(ring) < 4:
continue # only proceed if ring is bigger than 3 atoms
chunks = get_overlapped_chunks(ring, 4)
for chunk in chunks:
if self.is_flat(chunk):
if not chunk in flats:
flats.append(chunk)
if not flats:
return False
newflats = []
# check for neighbours and add the to the flat list:
for chunk in flats:
newflats.append(chunk)
for atnum, chunkatom in enumerate(chunk[:]):
for nbatom in neighbors:
if self.binds_to(nbatom, chunkatom):
# add bound atoms near their partners:
ch = chunk[:]
ch.insert(atnum, nbatom)
ch = shift(ch, atnum)
H = self._G.subgraph(ch)
# Try to delete atoms in the subgraph and test if subgraph divides.
# If it not devides, remove the atom unless it is the just added neighbour.
for num, i in enumerate(reversed(ch), start=1):
# print(ch, nbatom, ch[-num], num, '###')
H.remove_node(ch[-num])
comp = list(nx.connected_components(H))
# check if graph is disconnected now:
if len(comp) > 1:
continue
else:
# do not delete the just added neighbour:
if ch[-num] in neighbors:
continue
del ch[-num]
break # finished, go to next flat
# only add if it really results in a flat composition:
ch.sort()
if self.is_flat(ch):
if ch in newflats:
pass
else:
newflats.append(ch)
return newflats
def letters(self):
return flatten(sy.letters for sy in self.syllables)
def _generateChecksForAllModels(self):
l = [m.members for m in self.models]
invocs = uniq(flatten(l))
self.contentProvider.generateChecksForInvocations(invocs)
def syllables(self):
return flatten(foot.syllables for foot in self.feet)
def letters(self):
return flatten(elem.letters for elem in self.elems)
def _generateChecksForAllModels(self):
l = [m.members for m in self.models]
invocs = uniq(flatten(l))
self.contentProvider.generateChecksForInvocations(invocs)
def dependents(self):
s = set(flatten(self.lhs))
for x in self.rhsset:
s = s | set(flatten(x))
return s
def __repr__(self):
return 'compile({} -> {})'.format(
str(flatten([o.get_files() for o in self.objects])), self.output)
def train(self, insts, rate, convergenceThreshold, maxIters):
'''Train this RBFNN - calculate beta values for each RBF node, and
perform gradient descent to learn weights for the weighted sum nodes.
The wtMean and wtStdDev parameters are the mean and standard deviation
of the gaussian distribution from which initial weights for the weighted
sum nodes will be randomly drawn.'''
protos, clusters = kMeans(self.numProtos, insts)
# Filter empty clusters
newProtos = []
newClusters = []
toRemove = [False if len(c) == 0 else True for c in clusters]
for idx, shouldKeep in enumerate(toRemove):
if shouldKeep:
newProtos.append(protos[idx])
newClusters.append(clusters[idx])
protos = newProtos
clusters = newClusters
# Calculate beta coefficients
betas = []
for cluster in clusters:
# If the cluster is empty, make the beta coefficient equal 1, which
# will cause the activation of this node decrease very sharply as
# the given instance gets further from the prototype, effectively
# rendering that prototype irrelevant.
if len(cluster) == 0:
betas.append(0)
else:
clusterMean = meanInst(cluster)
dists = [
euclideanDist(inst.data, clusterMean.data)
for inst in cluster
]
sigma = sum(dists) / len(cluster)
if sum(dists) == 0:
betas.append(1)
else:
betas.append(1.0 / (2 * math.pow(sigma, 2)))
# Create the RBF nodes from the prototype & beta coefficient
self.rbfNodes = [
RBFNode(proto, beta) for proto, beta in zip(protos, betas)
]
# Perform gradient descent to learn weights for the output nodes.
conv = ConvergenceTester(convergenceThreshold)
for x in range(maxIters):
rbfOutputs = [[1] + self.passRBFLayer(inst) for inst in insts]
predictions = [self.fwdPass(inst) for inst in insts]
for outputIndex, node in enumerate(self.wtSumNodes):
for wtIdx in range(len(node.wts)):
node.wts[wtIdx] -= (rate * (sum([( \
predictions[i][outputIndex] - \
inst.label[outputIndex]) * rbfOutputs[i][wtIdx] \
for i, inst in enumerate(insts)])/len(insts)))
if conv.test(flatten([node.wts for node in self.wtSumNodes])):
break
def dependents(self):
# What symbols are we dependent on?
return set(flatten(self.lhs)) | set(flatten(self.rhs))
def make_substitutions(self, env):
return CombinatoricAlternatives(*flatten(
flatten(make_substitution(alt, env)) for alt in self.alternatives
))
def can_push(self, defined):
return isinstance(self.rhs, str) and \
self.rhs not in defined and \
set(flatten(self.lhs)).issubset(set(defined))
def get_files(self):
return flatten(
[source.get_files() for source in self.sources]
)
