Add month mode

export_data-v2.py

@@ -3,6 +3,7 @@
 # ------------------------------------------------
 
 from matplotlib import pyplot as plt
+from datetime import datetime
 import os
 import json
 import re
@@ -15,6 +16,9 @@ import sys, getopt
 
 DATA_PATH = './data/'
 
+MONTH_MODE = False
+MONTH = None
+
 OTHER_LABEL = 'Les Autres'
 
 # JSON tags
@@ -28,19 +32,26 @@ SENDER = 'sender_name'
 HELP = """General options :
     -h, --help          Consulter l'aide
     --path=<path>       Redéfinir le chemin d'accès aux données (par défaut ./data)
+    --month <mm/yyyy>
     """
 
 # ------------------------------------------------
 #                   Functions
 # ------------------------------------------------
 
+def printHelp():
+    print('Usage:\n '+os.path.basename(__file__)+' <command> [option]\n')
+    print(HELP)
+    sys.exit(2)
+
 def handleArguments(argv):
+    global MONTH_MODE
+    global MONTH
+    
     try:
-        opts, args = getopt.getopt(argv, 'h',['help','path='])
+        opts, args = getopt.getopt(argv, 'h',['help','path=', 'month='])
     except getopt.GetoptError:
-        print('Usage:\n '+os.path.basename(__file__)+' <command> [option]\n')
-        print(HELP)
-        sys.exit(2)
+        printHelp()
 
     for opt, arg in opts:
         if opt in ('-h', '--help'):
@@ -48,6 +59,15 @@ def handleArguments(argv):
             sys.exit()
         elif opt in ('--path'):
             DATA_PATH = arg
+        elif opt in ('--month'):
+            MONTH_MODE = True
+            try:
+                t = arg.split("/")
+                int(t[0])
+                int(t[1])
+                MONTH = t 
+            except:
+                printHelp()
 
 def readBrokenFbJson(datafile_path):
     # ntm facebook
@@ -65,7 +85,6 @@ def computeData():
     datafiles_path = [DATA_PATH + filename for filename in os.listdir(DATA_PATH)]
     messages, participants = [], []
 
-    print(datafiles_path)
     
     for datafile_path in datafiles_path:
         datacontent = readBrokenFbJson(datafile_path)
@@ -82,7 +101,15 @@ def cleanParticipants(rawParticipants):
     return set([participant[NAME] for participant in rawParticipants])
 
 def cleanMessages(rawMessages):
-    cleanMessages = [message for message in rawMessages if CONTENT in message]
+    if MONTH_MODE:
+        cleanMessages = [
+            message for message in rawMessages if \
+            CONTENT in message and \
+            datetime.fromtimestamp(message[TIMESTAMP]/1000).month == int(MONTH[0]) and \
+            datetime.fromtimestamp(message[TIMESTAMP]/1000).year == int(MONTH[1])
+        ]
+    else:
+        cleanMessages = [message for message in rawMessages if CONTENT in message]
     return sorted(cleanMessages, key = lambda x: x[TIMESTAMP])
 
 # TODO tester l'approche en recherche incrémentale
@@ -96,16 +123,19 @@ def computeParticipation(messages):
 
     for message in messages:
         sender = message[SENDER]
-        result[sender] = result[sender]+1 if sender in result else 1
+        countParticipation(result, sender, message)
 
     return sorted(result.items(), key = lambda x: x[1])
 
+def countParticipation(participations, sender, message):
+    participations[sender] = participations[sender] + 1 if sender in participations else 1
+
 def mergeSmallParticipation(rawParticipation, threshold = 1):
     values = [e[1] for e in rawParticipation]
     labels = [e[0] for e in rawParticipation]
 
     totalValues = sum(values)
-
+    idx = 0
     for idx, value in enumerate(values):
         if 100 * value / totalValues >= threshold: break