def postAnalyzeFile(outputFile, bngLocation, database):
"""
Performs a postcreation file analysis based on context information
"""
#print('Transforming generated BNG file to BNG-XML representation for analysis')
postAnalysisHelper(outputFile, bngLocation, database)
# recreate file using information from the post analysis
returnArray = analyzeHelper(database.document, database.reactionDefinitions, database.useID,
outputFile, database.speciesEquivalence, database.atomize, database.translator, database)
with open(outputFile, 'w') as f:
f.write(returnArray.finalString)
# recompute bng-xml file
consoleCommands.bngl2xml(outputFile)
bngxmlFile = '.'.join(outputFile.split('.')[:-1]) + '.xml'
# recompute context information
contextAnalysis = postAnalysis.ModelLearning(bngxmlFile)
# get those species patterns that follow uncommon motifs
motifSpecies, motifDefinitions = contextAnalysis.processContextMotifInformation(database.assumptions['lexicalVsstoch'], database)
#motifSpecies, motifDefinitions = contextAnalysis.processAllContextInformation()
if len(motifDefinitions) > 0:
logMess('INFO:CTX003', 'Species with suspect context information were found. Information is being dumped to {0}_context.log'.format(outputFile))
with open('{0}_context.log'.format(outputFile), 'w') as f:
pprint.pprint(dict(motifSpecies), stream=f)
pprint.pprint(motifDefinitions, stream=f)
def main():
fileName = 'complex/output19.bngl'
console.bngl2xml(fileName)
species, rules, par = readBNGXML.parseXML('output19.xml')
#print rules
transformationCenter = []
transformationContext = []
#extract the context of such reactions
for idx, rule in enumerate(rules):
tatomicArray, ttransformationCenter, ttransformationContext, \
tproductElements,tactionNames,tlabelArray = extractAtomic.extractTransformations([rule])
#atomicArray.append(tatomicArray)
transformationCenter.append(ttransformationCenter)
transformationContext.append(ttransformationContext)
redundantDict, patternDictList = extractRedundantContext(
rules, transformationCenter, transformationContext)
#print redundantDict
#construct rule based patterns based on the reaction patterns they have to match and
#the ones they have to discriminate
for center in patternDictList:
for rate in patternDictList[center]:
match = patternDictList[center][rate]
notContext = []
for cRate in [x for x in patternDictList[center] if x != rate]:
notContext.append([
transformationContext[x]
for x in redundantDict[center][cRate]
])
ruleSet = [rules[x] for x in redundantDict[center][rate]]
createMetaRule(ruleSet, match)
newRules = range(0, len(rules))
for center in redundantDict:
for context in redundantDict[center]:
for element in range(1, len(redundantDict[center][context])):
newRules.remove(redundantDict[center][context][element])
#for element in newRules:
# print str(rules[element][0])
newRulesArray = []
for element in newRules:
newRulesArray.append('{0}\n'.format(str(rules[element][0])))
lines = readFile(fileName)
startR = lines.index('begin reaction rules\n')
endR = lines.index('end reaction rules\n')
startP = lines.index('begin parameters\n')
endP = lines.index('end parameters\n')
newPar = findNewParameters(lines[startP + 1:endP], par)
newLines = lines[0:endP] + newPar + lines[
endP:startR + 1] + newRulesArray + lines[endR:len(lines)]
f = open(fileName + 'reduced.bngl', 'w')
f.writelines(newLines)
'''
def main():
fileName = 'complex/output19.bngl'
console.bngl2xml(fileName)
species,rules,par= readBNGXML.parseXML('output19.xml')
#print rules
transformationCenter = []
transformationContext = []
#extract the context of such reactions
for idx,rule in enumerate(rules):
tatomicArray, ttransformationCenter, ttransformationContext, \
tproductElements,tactionNames,tlabelArray = extractAtomic.extractTransformations([rule])
#atomicArray.append(tatomicArray)
transformationCenter.append(ttransformationCenter)
transformationContext.append(ttransformationContext)
redundantDict,patternDictList = extractRedundantContext(rules,transformationCenter,transformationContext)
#print redundantDict
#construct rule based patterns based on the reaction patterns they have to match and
#the ones they have to discriminate
for center in patternDictList:
for rate in patternDictList[center]:
match = patternDictList[center][rate]
notContext = []
for cRate in [x for x in patternDictList[center] if x != rate]:
notContext.append([transformationContext[x] for x in redundantDict[center][cRate]])
ruleSet = [rules[x] for x in redundantDict[center][rate]]
createMetaRule(ruleSet,match)
newRules = range(0,len(rules))
for center in redundantDict:
for context in redundantDict[center]:
for element in range(1,len(redundantDict[center][context])):
newRules.remove(redundantDict[center][context][element])
#for element in newRules:
# print str(rules[element][0])
newRulesArray = []
for element in newRules:
newRulesArray.append('{0}\n'.format(str(rules[element][0])))
lines = readFile(fileName)
startR = lines.index('begin reaction rules\n')
endR = lines.index('end reaction rules\n')
startP = lines.index('begin parameters\n')
endP = lines.index('end parameters\n')
newPar = findNewParameters(lines[startP+1:endP],par)
newLines = lines[0:endP] + newPar + lines[endP:startR+1] + newRulesArray + lines[endR:len(lines)]
f = open(fileName + 'reduced.bngl','w')
f.writelines(newLines)
'''
def postAnalyzeFile(outputFile, bngLocation, database):
"""
Performs a postcreation file analysis based on context information
"""
#print('Transforming generated BNG file to BNG-XML representation for analysis')
consoleCommands.setBngExecutable(bngLocation)
outputDir = os.sep.join(outputFile.split(os.sep)[:-1])
if outputDir != '':
retval = os.getcwd()
os.chdir(outputDir)
consoleCommands.bngl2xml(outputFile.split(os.sep)[-1])
if outputDir != '':
os.chdir(retval)
bngxmlFile = '.'.join(outputFile.split('.')[:-1]) + '.xml'
#print('Sending BNG-XML file to context analysis engine')
contextAnalysis = postAnalysis.ModelLearning(bngxmlFile)
# analysis of redundant bonds
deleteBonds = contextAnalysis.analyzeRedundantBonds(database.assumptions['redundantBonds'])
modificationFlag = True
for molecule in database.assumptions['redundantBondsMolecules']:
if molecule[0] in deleteBonds:
for bond in deleteBonds[molecule[0]]:
database.translator[molecule[1]].deleteBond(bond)
logMess('INFO:Atomization', 'Used context information to determine that the bond {0} in species {1} is not likely'.format(bond,molecule[1]))
returnArray = analyzeHelper(database.document, database.reactionDefinitions, database.useID,
outputFile, database.speciesEquivalence, database.atomize, database.translator)
with open(outputFile, 'w') as f:
f.write(returnArray.finalString)
# recompute bng-xml file
consoleCommands.bngl2xml(outputFile)
bngxmlFile = '.'.join(outputFile.split('.')[:-1]) + '.xml'
# recompute context information
contextAnalysis = postAnalysis.ModelLearning(bngxmlFile)
# get those species patterns that follow uncommon motifs
motifSpecies, motifDefinitions = contextAnalysis.processContextMotifInformation(database.assumptions['lexicalVsstoch'], database)
#motifSpecies, motifDefinitions = contextAnalysis.processAllContextInformation()
if len(motifDefinitions) > 0:
logMess('WARNING:ContextAnalysis', 'Species with suspect context information were found. Information is being dumped to {0}_context.log'.format(outputFile))
with open('{0}_context.log'.format(outputFile), 'w') as f:
pprint.pprint(dict(motifSpecies), stream=f)
pprint.pprint(motifDefinitions, stream=f)
def postAnalysisHelper(outputFile, bngLocation, database):
consoleCommands.setBngExecutable(bngLocation)
outputDir = os.sep.join(outputFile.split(os.sep)[:-1])
if outputDir != '':
retval = os.getcwd()
os.chdir(outputDir)
consoleCommands.bngl2xml(outputFile.split(os.sep)[-1])
if outputDir != '':
os.chdir(retval)
bngxmlFile = '.'.join(outputFile.split('.')[:-1]) + '.xml'
#print('Sending BNG-XML file to context analysis engine')
contextAnalysis = postAnalysis.ModelLearning(bngxmlFile)
# analysis of redundant bonds
deleteBonds = contextAnalysis.analyzeRedundantBonds(database.assumptions['redundantBonds'])
for molecule in database.assumptions['redundantBondsMolecules']:
if molecule[0] in deleteBonds:
for bond in deleteBonds[molecule[0]]:
database.translator[molecule[1]].deleteBond(bond)
logMess('INFO:CTX002', 'Used context information to determine that the bond {0} in species {1} is not likely'.format(bond,molecule[1]))
def extractStatistics():
number = 151
console.bngl2xml('complex/output{0}.bngl'.format(number))
species, rules, parameterDict = readBNGXML.parseXML(
'output{0}.xml'.format(number))
#print rules
transformationCenter = []
transformationContext = []
k = []
actions = Counter()
actionSet = Counter()
for idx, rule in enumerate(rules):
tatomicArray, ttransformationCenter, ttransformationContext, \
tproductElements,tactionNames,tlabelArray = extractAtomic.extractTransformations([rule])
#atomicArray.append(tatomicArray)
transformationCenter.append(ttransformationCenter)
transformationContext.append(ttransformationContext)
k.append(len(rule[0].actions))
#if len(rule[0].actions) > 3:
# print rule[0].reactants
actions.update([x.action for x in rule[0].actions])
tmp = [x.action for x in rule[0].actions]
tmp.sort()
actionSet.update([tuple(tmp)])
#print actions
#print actionSet
print 'number of species', len(species)
print 'avg number o actions', np.average(k), np.std(k)
centerDict = groupByReactionCenter(transformationCenter)
print 'singletons', len(
{x: centerDict[x]
for x in centerDict if len(centerDict[x]) == 1})
tmp = [[tuple(set(x)), len(centerDict[x])] for x in centerDict]
#reactionCenterGraph(species,tmp)
#tmp.sort(key=lambda x:x[1],reverse=True)
print 'number of reaction centers', len(centerDict.keys())
print 'number of rules', len(rules)
#print 'unique',[x for x in centerDict[element] if len(centerDict[element]) == 1]
redundantDict = groupByReactionCenterAndRateAndActions(rules, centerDict)
#print redundantDict
tmp2 = [('$\\tt{{{0}}}$'.format(tuple(set(x))), tuple(set(y[:-1])), y[-1],
len(redundantDict[x][y])) for x in redundantDict
for y in redundantDict[x] if 'kr' not in y[-1]]
tmp2 = set(tmp2)
tmp2 = list(tmp2)
tmp2.sort(key=lambda x: x[3], reverse=True)
tmp2.sort(key=lambda x: x[0], reverse=True)
tmp2 = [
'{0} & {1} & {2} & {3}\\\\\n'.format(element[0], element[1],
element[2], element[3])
for element in tmp2
]
with open('table.tex', 'w') as f:
f.write('\\begin{tabular}{|cccc|}\n')
f.write('\\hline\n')
f.write('Reaction Centers & Action & Score\\\\\\hline\n')
for element in tmp2:
f.write(element)
f.write('\\hline\n')
f.write('\\end{tabular}\n')
#print redundantDict
x = [len(centerDict[x]) for x in centerDict]
#122,138
print 'average number of reactions with same rate and reaction cneter', np.average(
x), np.std(x)
print 'total number of clusters', len(x)
print x
def extractStatistics():
number = 151
console.bngl2xml('complex/output{0}.bngl'.format(number))
species,rules,parameterDict= readBNGXML.parseXML('output{0}.xml'.format(number))
#print rules
transformationCenter = []
transformationContext = []
k = []
actions = Counter()
actionSet = Counter()
for idx,rule in enumerate(rules):
tatomicArray, ttransformationCenter, ttransformationContext, \
tproductElements,tactionNames,tlabelArray = extractAtomic.extractTransformations([rule])
#atomicArray.append(tatomicArray)
transformationCenter.append(ttransformationCenter)
transformationContext.append(ttransformationContext)
k.append(len(rule[0].actions))
#if len(rule[0].actions) > 3:
# print rule[0].reactants
actions.update([x.action for x in rule[0].actions])
tmp = [x.action for x in rule[0].actions]
tmp.sort()
actionSet.update([tuple(tmp)])
#print actions
#print actionSet
print 'number of species',len(species)
print 'avg number o actions',np.average(k),np.std(k)
centerDict = groupByReactionCenter(transformationCenter)
print 'singletons',len({x:centerDict[x] for x in centerDict if len(centerDict[x]) == 1})
tmp = [[tuple(set(x)),len(centerDict[x])] for x in centerDict]
#reactionCenterGraph(species,tmp)
#tmp.sort(key=lambda x:x[1],reverse=True)
print 'number of reaction centers',len(centerDict.keys())
print 'number of rules',len(rules)
#print 'unique',[x for x in centerDict[element] if len(centerDict[element]) == 1]
redundantDict = groupByReactionCenterAndRateAndActions(rules,centerDict)
#print redundantDict
tmp2 = [('$\\tt{{{0}}}$'.format(tuple(set(x))),tuple(set(y[:-1])),y[-1],len (redundantDict[x][y])) for x in redundantDict for y in redundantDict[x] if 'kr' not in y[-1]]
tmp2 = set(tmp2)
tmp2 = list(tmp2)
tmp2.sort(key=lambda x:x[3],reverse=True)
tmp2.sort(key=lambda x:x[0],reverse=True)
tmp2 = ['{0} & {1} & {2} & {3}\\\\\n'.format(element[0],element[1],element[2],element[3]) for element in tmp2]
with open('table.tex','w') as f:
f.write('\\begin{tabular}{|cccc|}\n')
f.write('\\hline\n')
f.write('Reaction Centers & Action & Score\\\\\\hline\n')
for element in tmp2:
f.write(element)
f.write('\\hline\n')
f.write('\\end{tabular}\n')
#print redundantDict
x = [len(centerDict[x]) for x in centerDict]
#122,138
print 'average number of reactions with same rate and reaction cneter',np.average(x),np.std(x)
print 'total number of clusters',len(x)
print x
def getInformationContent(fileName):
console.bngl2xml(fileName,timeout=30)
xmlFileName = '.'.join(fileName.split('.')[:-1])+'.xml'
xmlFileName =xmlFileName.split('/')[-1]
species,reactions,_= readBNGXML.parseXML(xmlFileName)
calculateSpeciesInformationContent(species)
def main():
with open('linkArray.dump', 'rb') as f:
linkArray = pickle.load(f)
with open('xmlAnnotationsExtended.dump', 'rb') as f:
annotations = pickle.load(f)
speciesEquivalence = {}
onlyDicts = [x for x in listdir('./complex')]
onlyDicts = [x for x in onlyDicts if '.bngl.dict' in x]
for x in onlyDicts:
with open('complex/{0}'.format(x), 'rb') as f:
speciesEquivalence[int(x.split('.')[0][6:])] = pickle.load(f)
for cidx, cluster in enumerate(linkArray):
#FIXME:only do the first cluster
cidx = 0
cluster = linkArray[0]
if len(cluster) == 1:
continue
annotationsDict = {idx: x for idx, x in enumerate(annotations)}
annotationsIncidence = {}
#get the model intersection points
for element in annotationsDict:
for species in annotationsDict[element]:
key = annotationsDict[element][species][1]
if key not in annotationsIncidence:
annotationsIncidence[key] = set([])
annotationsIncidence[key].add(element)
#number = 19
graph = nx.Graph()
nodeDict = {}
superNodeDict = {}
for element in annotationsIncidence:
if len(annotationsIncidence[element]) > 1:
tmpEdgeList = []
names = []
for idx in annotationsIncidence[element]:
for species in [
x for x in annotationsDict[idx]
if annotationsDict[idx][x][1] == element
]:
try:
print(speciesEquivalence[idx])
if isinstance(speciesEquivalence[idx][species],
str):
tmpEdgeList.append('{0}_{1}'.format(
idx,
speciesEquivalence[idx][species][0:-2]))
names.append(
speciesEquivalence[idx][species][0:-2])
else:
tmpEdgeList.append('{0}_{1}'.format(
idx, speciesEquivalence[idx]
[species].molecules[0].name))
names.append(speciesEquivalence[idx]
[species].molecules[0].name)
except:
print('----', 'error', idx, species)
pass
names = [x.lower() for x in names]
name = Counter(names).most_common(5)
name = [x for x in name if len(x[0]) > 1]
if len(name) == 0:
name = Counter(names).most_common(1)
for edge in tmpEdgeList:
superNodeDict[edge] = name[0][0]
if len(name) > 0:
superNodeDict[name[0][0]] = len(tmpEdgeList)
#pairs = [(x, y, 1) for x in tmpEdgeList for y in tmpEdgeList if x!= y]
#graph.add_weighted_edges_from(pairs)
print superNodeDict
for x in cluster:
number = x - 1
try:
console.bngl2xml('complex/output{0}.bngl'.format(number))
species, rules, parameterDict, observableList = readBNGXML.parseXML(
'output{0}.xml'.format(number))
except:
continue
nodeDict[number] = simpleGraph(graph, species, observableList,
number, superNodeDict)
nx.write_gml(graph, 'cgraph_{0}.gml'.format(cidx))
break
def main():
with open('linkArray.dump', 'rb') as f:
linkArray = pickle.load(f)
with open('xmlAnnotationsExtended.dump','rb') as f:
annotations = pickle.load(f)
speciesEquivalence = {}
onlyDicts = [ x for x in listdir('./complex')]
onlyDicts = [x for x in onlyDicts if '.bngl.dict' in x]
for x in onlyDicts:
with open('complex/{0}'.format(x),'rb') as f:
speciesEquivalence[int(x.split('.')[0][6:])] = pickle.load(f)
for cidx,cluster in enumerate(linkArray):
#FIXME:only do the first cluster
cidx = 0
cluster = linkArray[0]
if len(cluster) == 1:
continue
annotationsDict = {idx:x for idx,x in enumerate(annotations)}
annotationsIncidence = {}
#get the model intersection points
for element in annotationsDict:
for species in annotationsDict[element]:
key = annotationsDict[element][species][1]
if key not in annotationsIncidence:
annotationsIncidence[key] = set([])
annotationsIncidence[key].add(element)
#number = 19
graph = nx.Graph()
nodeDict = {}
superNodeDict = {}
for element in annotationsIncidence:
if len(annotationsIncidence[element]) > 1:
tmpEdgeList = []
names= []
for idx in annotationsIncidence[element]:
for species in [x for x in annotationsDict[idx] if annotationsDict[idx][x][1] == element]:
try:
print speciesEquivalence[idx]
if isinstance(speciesEquivalence[idx][species],str):
tmpEdgeList.append('{0}_{1}'.format(idx,speciesEquivalence[idx][species][0:-2]))
names.append(speciesEquivalence[idx][species][0:-2])
else:
tmpEdgeList.append('{0}_{1}'.format(idx,speciesEquivalence[idx][species].molecules[0].name))
names.append(speciesEquivalence[idx][species].molecules[0].name)
except:
print '----','error',idx,species
pass
names = [x.lower() for x in names]
name = Counter(names).most_common(5)
name = [x for x in name if len(x[0])> 1]
if len(name) == 0:
name = Counter(names).most_common(1)
for edge in tmpEdgeList:
superNodeDict[edge] = name[0][0]
if len(name) > 0:
superNodeDict[name[0][0]] = len(tmpEdgeList)
#pairs = [(x,y,1) for x in tmpEdgeList for y in tmpEdgeList if x!= y]
#graph.add_weighted_edges_from(pairs)
print superNodeDict
for x in cluster:
number = x-1
try:
console.bngl2xml('complex/output{0}.bngl'.format(number))
species,rules,parameterDict,observableList= readBNGXML.parseXML('output{0}.xml'.format(number))
except:
continue
nodeDict[number] = simpleGraph(graph,species,observableList,number,superNodeDict)
nx.write_gml(graph,'cgraph_{0}.gml'.format(cidx))
break