add readme and driver

dump_msg.py

@@ -0,0 +1,30 @@
+#!/usr/bin/env python2
+# -*- coding: UTF-8 -*-
+# File: dump_msg.py
+# Date: Fri Nov 21 14:36:23 2014 +0800
+# Author: Yuxin Wu <[email protected]>
+
+from lib.Parser import WeChatDBParser
+
+import sys, os
+
+if len(sys.argv) != 3:
+    sys.exit("Usage: {0} <path to decoded_database.db> <output_dir>".format(sys.argv[0]))
+
+db_file = sys.argv[1]
+output_dir = sys.argv[2]
+try:
+    os.mkdir(output_dir)
+except:
+    pass
+if not os.path.isdir(output_dir):
+    sys.exit("Error creating directory {}".format(output_dir))
+
+parser = WeChatDBParser(db_file)
+parser.parse()
+
+for name, msgs in parser.msgs_by_talker.iteritems():
+    print u"Writing msgs for {}".format(name)
+    with open(os.path.join(output_dir, name + '.txt'), 'w') as f:
+        for m in msgs:
+            print >> f, m

lib/Msg.py

@@ -1,7 +1,7 @@
 #!/usr/bin/env python2
 # -*- coding: UTF-8 -*-
 # File: Msg.py
-# Date: Fri Nov 21 14:08:33 2014 +0800
+# Date: Fri Nov 21 14:10:21 2014 +0800
 # Author: Yuxin Wu <[email protected]>
 
 from datetime import datetime
@@ -62,7 +62,7 @@ class WeChatMsg(object):
             self.createTime,
             ensure_unicode(self.msg_str())).encode('utf-8')
         if self.imgPath:
-            ret = u"{}|img:{}".format(ensure_unicode(ret), self.imgPath)
+            ret = u"{}|img:{}".format(ensure_unicode(ret.strip()), self.imgPath)
             return ret.encode('utf-8')
         else:
             return ret

lib/Parser.py

@@ -1,16 +1,13 @@
 #!/usr/bin/env python2
 # -*- coding: UTF-8 -*-
-# File: dump.py
-# Date: Fri Nov 21 14:08:45 2014 +0800
+# File: Parser.py
+# Date: Fri Nov 21 14:36:18 2014 +0800
 # Author: Yuxin Wu <[email protected]>
 
 import sqlite3
 from collections import defaultdict
-from pprint import PrettyPrinter
-pp = PrettyPrinter()
-def log(x): print repr(x).decode('unicode-escape')
-from lib.Msg import WeChatMsg
-from lib.utils import ensure_unicode
+from .Msg import WeChatMsg
+from .utils import ensure_unicode
 
 """ tables in concern:
 emojiinfo
@@ -71,7 +68,3 @@ SELECT {} FROM message
     def parse(self):
         self._parse_contact()
         self._parse_msg()
-
-if __name__ == '__main__':
-    parser = WeChatDBParser('./decoded_database.db')
-    parser.parse()