Write data to csv file after each family rather than after entire set of families

This is so you can view progress if things crash midway
also saves family names alphabetically

Author: connie

testing/evaluateKinetics.py

@@ -233,20 +233,17 @@ def obtainKineticsFamilyStatistics(FullDatabase, trialDir):
     as it would add non-exact rules to the rule count)
     """
     allFamilyNames=FullDatabase.kinetics.families.keys()
-    allFamilyNames.sort()  # Perform test on families alphabetically
-
-    familyCount={}
+    allFamilyNames.sort(key=str.lower)  # Perform test on families alphabetically
 
-    for familyName in allFamilyNames: 
-        family=FullDatabase.kinetics.families[familyName]
-        print "Processing", familyName + '...', '(' + str(len(family.rules.entries)) + ' rate rules)'
-        familyCount[familyName]=countNodes(family)
-
     with open(os.path.join(trialDir, 'FamilyStatistics.csv'), 'wb') as csvfile:
         csvwriter=csv.writer(csvfile)
         csvwriter.writerow(['Family','Number of Rules', 'Top Node 1', 'Number of Groups', 'Top Node 2', 'Number of Groups', 'Top Node 3', 'Number of Groups'])
-        for key, value in familyCount.iteritems():
-            csvwriter.writerow(value)
+        
+        for familyName in allFamilyNames: 
+            family=FullDatabase.kinetics.families[familyName]
+            print "Processing", familyName + '...', '(' + str(len(family.rules.entries)) + ' rate rules)'
+            # Save to csv file
+            csvwriter.writerow(countNodes(family))
 
 
 def compareNIST(FullDatabase, trialDir):
@@ -262,50 +259,50 @@ def compareNIST(FullDatabase, trialDir):
 
 
     allFamilyNames=FullDatabase.kinetics.families.keys()
-    allFamilyNames.sort()  # Perform test on families alphabetically
+    allFamilyNames.sort(key=str.lower)  # Perform test on families alphabetically
 
     QDict={}
 
-    for familyName in allFamilyNames: 
-        family=FullDatabase.kinetics.families[familyName]
-        print "Processing", familyName + '...', '(' + str(len(family.rules.entries)) + ' rate rules)'
-        if len(family.rules.entries) < 2:
-            print '    Skipping', familyName, ': only has one rate rule...'
-        else:
-            exactKinetics, approxKinetics = getKineticsDepository(FullDatabase, family, 'NIST')
-            
-            #prune for exact matches only
-            keysToRemove=[]
-            for key, kinetics in approxKinetics.iteritems():
-                if not re.search('Exact', kinetics.comment):
-                    keysToRemove.append(key)
-            
-            for key in keysToRemove:
-                del approxKinetics[key]
-            
-            parityData=analyzeForParity(exactKinetics, approxKinetics, cutoff=8.0)
-
-            if len(parityData)<2:
-                print '    Skipping', familyName, ': {} reactions were compared...'.format(len(parityData))
-                continue
-            QDict[familyName]=calculateQ(parityData)
-            createParityPlot(parityData)
-            plt.title(familyName)
-            plt.savefig(os.path.join(trialDir, familyName +'.png'))
-            plt.clf()
-               
-            if not os.path.exists(os.path.join(os.path.join(trialDir, 'ParityData'))):
-                os.makedirs(os.path.join(trialDir, 'ParityData'))
-               
-            with open(os.path.join(trialDir, 'ParityData', familyName + '.csv'), 'wb') as csvfile:
-                csvwriter=csv.writer(csvfile)
-                for key, value in parityData.iteritems():
-                    csvwriter.writerow([key, value[0]/value[1], approxKinetics[key].comment]) 
- 
     with open(os.path.join(trialDir, 'QDict_LOO.csv'), 'wb') as csvfile:
         csvwriter=csv.writer(csvfile)
-        for key, value in QDict.iteritems():
-            csvwriter.writerow([key, value])
+        for familyName in allFamilyNames: 
+            family=FullDatabase.kinetics.families[familyName]
+            print "Processing", familyName + '...', '(' + str(len(family.rules.entries)) + ' rate rules)'
+            if len(family.rules.entries) < 2:
+                print '    Skipping', familyName, ': only has one rate rule...'
+            else:
+                exactKinetics, approxKinetics = getKineticsDepository(FullDatabase, family, 'NIST')
+                
+                #prune for exact matches only
+                keysToRemove=[]
+                for key, kinetics in approxKinetics.iteritems():
+                    if not re.search('Exact', kinetics.comment):
+                        keysToRemove.append(key)
+                
+                for key in keysToRemove:
+                    del approxKinetics[key]
+                
+                parityData=analyzeForParity(exactKinetics, approxKinetics, cutoff=8.0)
+
+                if len(parityData)<2:
+                    print '    Skipping', familyName, ': {} reactions were compared...'.format(len(parityData))
+                    continue
+                QDict[familyName]=calculateQ(parityData)
+                createParityPlot(parityData)
+                plt.title(familyName)
+                plt.savefig(os.path.join(trialDir, familyName +'.png'))
+                plt.clf()
+                   
+                if not os.path.exists(os.path.join(os.path.join(trialDir, 'ParityData'))):
+                    os.makedirs(os.path.join(trialDir, 'ParityData'))
+                   
+                with open(os.path.join(trialDir, 'ParityData', familyName + '.csv'), 'wb') as paritycsvfile:
+                    paritycsvwriter=csv.writer(paritycsvfile)
+                    for key, value in parityData.iteritems():
+                        paritycsvwriter.writerow([key, value[0]/value[1], approxKinetics[key].comment]) 
+            
+                # Save data to csv file
+                csvwriter.writerow([familyName, QDict[familyName]])
 
 
 
@@ -328,42 +325,44 @@ def leaveOneOut(FullDatabase, trialDir, averaging=True):
         os.makedirs(trialDir)
 
     allFamilyNames=FullDatabase.kinetics.families.keys()
-    allFamilyNames.sort()  # Perform test on families alphabetically
+    allFamilyNames.sort(key=str.lower)  # Perform test on families alphabetically
     QDict={}
 
-    for familyName in allFamilyNames: 
-        family=FullDatabase.kinetics.families[familyName]
-        print "Processing", familyName + '...', '(' + str(len(family.rules.entries)) + ' rate rules)'
-        if len(family.rules.entries) < 2:
-            print '    Skipping', familyName, ': only has one rate rule...'
-        else:
-            if not averaging:
-                # Pre-average the family if averaging is not turned on
-                family.fillKineticsRulesByAveragingUp()
-            exactKinetics, approxKinetics = getKineticsLeaveOneOut(family, averaging)
-            parityData=analyzeForParity(exactKinetics, approxKinetics, cutoff=8.0)
-
-            if len(parityData)<2:
-                print '    Skipping', familyName, ': {} rate rules were compared...'.format(len(parityData))
-                continue
-            QDict[familyName]=calculateQ(parityData)
-            createParityPlot(parityData)
-            plt.title(familyName)
-            plt.savefig(os.path.join(trialDir, familyName +'.png'))
-            plt.clf()
-              
-            if not os.path.exists(os.path.join(os.path.join(trialDir, 'ParityData'))):
-                os.makedirs(os.path.join(trialDir, 'ParityData'))
-              
-            with open(os.path.join(trialDir, 'ParityData', familyName + '.csv'), 'wb') as csvfile:
-                csvwriter=csv.writer(csvfile)
-                for key, value in parityData.iteritems():
-                    csvwriter.writerow([key, value[0]/value[1], approxKinetics[key].comment]) 
 
     with open(os.path.join(trialDir, 'QDict_LOO.csv'), 'wb') as csvfile:
         csvwriter=csv.writer(csvfile)
-        for key, value in QDict.iteritems():
-            csvwriter.writerow([key, value])
+
+        for familyName in allFamilyNames: 
+            family=FullDatabase.kinetics.families[familyName]
+            print "Processing", familyName + '...', '(' + str(len(family.rules.entries)) + ' rate rules)'
+            if len(family.rules.entries) < 2:
+                print '    Skipping', familyName, ': only has one rate rule...'
+            else:
+                if not averaging:
+                    # Pre-average the family if averaging is not turned on
+                    family.fillKineticsRulesByAveragingUp()
+                exactKinetics, approxKinetics = getKineticsLeaveOneOut(family, averaging)
+                parityData=analyzeForParity(exactKinetics, approxKinetics, cutoff=8.0)
+
+                if len(parityData)<2:
+                    print '    Skipping', familyName, ': {} rate rules were compared...'.format(len(parityData))
+                    continue
+                QDict[familyName]=calculateQ(parityData)
+                createParityPlot(parityData)
+                plt.title(familyName)
+                plt.savefig(os.path.join(trialDir, familyName +'.png'))
+                plt.clf()
+                  
+                if not os.path.exists(os.path.join(os.path.join(trialDir, 'ParityData'))):
+                    os.makedirs(os.path.join(trialDir, 'ParityData'))
+                  
+                with open(os.path.join(trialDir, 'ParityData', familyName + '.csv'), 'wb') as paritycsvfile:
+                    paritycsvwriter=csv.writer(paritycsvfile)
+                    for key, value in parityData.iteritems():
+                        paritycsvwriter.writerow([key, value[0]/value[1], approxKinetics[key].comment]) 
+
+                # Save to csv file
+                csvwriter.writerow([familyName, QDict[familyName]])
 
     return