various bug fixes

lib/msg.py

@@ -1,11 +1,13 @@
 #!/usr/bin/env python2
 # -*- coding: UTF-8 -*-
 # File: msg.py
-# Date: Wed Dec 24 22:00:42 2014 +0800
+# Date: Thu Dec 25 09:56:24 2014 +0800
 # Author: Yuxin Wu <[email protected]>
 
+import re
 from datetime import datetime
 from pyquery import PyQuery
+
 from .utils import ensure_bin_str, ensure_unicode
 
 TYPE_MSG = 1
@@ -66,7 +68,14 @@ class WeChatMsg(object):
         elif self.type == TYPE_VIDEO:
             return "VIDEO FILE"
         elif self.type == TYPE_NAMECARD:
-            pq = PyQuery(self.content)
+            try:
+                pq = PyQuery(self.content)
+            except ValueError:
+                # pq doesn't support xml
+                pat = re.compile('<msg.*<\/msg>', re.DOTALL)
+                msg = pat.search(self.content).group()
+                pq = PyQuery(msg)
+
             msg = pq('msg').attr
             name = msg['nickname']
             if not name:

lib/parser.py

@@ -1,7 +1,7 @@
 #!/usr/bin/env python2
 # -*- coding: UTF-8 -*-
 # File: parser.py
-# Date: Mon Dec 22 22:14:18 2014 +0800
+# Date: Thu Dec 25 00:34:35 2014 +0800
 # Author: Yuxin Wu <[email protected]>
 
 import sqlite3
@@ -32,7 +32,7 @@ class WeChatDBParser(object):
         self.msgs_by_talker = defaultdict(list)
         self.emojis = {}
         self.internal_emojis = {}
-        self.parse()
+        self._parse()
 
     def _parse_contact(self):
         contacts = self.cc.execute(
@@ -105,7 +105,7 @@ SELECT {} FROM message
             self.internal_emojis[md5] = name
 
 
-    def parse(self):
+    def _parse(self):
         self._parse_userinfo()
         self._parse_contact()
         self._parse_msg()

lib/render.py

@@ -1,7 +1,7 @@
 #!/usr/bin/env python2
 # -*- coding: UTF-8 -*-
 # File: render.py
-# Date: Tue Dec 23 00:01:11 2014 +0800
+# Date: Thu Dec 25 10:01:58 2014 +0800
 # Author: Yuxin Wu <[email protected]>
 
 import os
@@ -92,8 +92,12 @@ class HTMLRender(object):
         elif msg.type == TYPE_IMG:
             # imgPath was original THUMBNAIL_DIRPATH://th_xxxxxxxxx
             imgpath = msg.imgPath.split('_')[-1]
+            if not imgpath:
+                logger.warn('No imgpath in an image message. Perhaps a bug in wechat.')
+                return fallback()
             bigimgpath = self.parser.imginfo.get(msg.msgSvrId)
-            fnames = [k for k in [imgpath, bigimgpath] if k is not None]
+            fnames = [k for k in [imgpath, bigimgpath] if k]
+            assert len(fnames) > 0, msg.msg_str()
             bigimg, smallimg = self.res.get_img(fnames)
             if not smallimg:
                 logger.warn("No image thumbnail found for {}".format(imgpath))

lib/res.py

@@ -1,7 +1,7 @@
 #!/usr/bin/env python2
 # -*- coding: UTF-8 -*-
 # File: res.py
-# Date: Mon Dec 22 16:52:56 2014 +0800
+# Date: Thu Dec 25 10:04:29 2014 +0800
 # Author: Yuxin Wu <[email protected]>
 
 import glob
@@ -127,6 +127,7 @@ class Resource(object):
 
     def get_img(self, fnames):
         """ return two base64 jpg string"""
+        fnames = [k for k in fnames if k]   # filter out empty string
         big_file, small_file = self._get_img_file(fnames)
 
         def get_jpg_b64(img_file):