Fix #3373 (I have problems with characters in title and book author name)

src/calibre/ebooks/metadata/rtf.py

@@ -3,7 +3,7 @@ __copyright__ = '2008, Kovid Goyal <kovid at kovidgoyal.net>'
 """
 Edit metadata in RTF files.
 """
-import re, cStringIO, sys
+import re, cStringIO, codecs
 
 from calibre.ebooks.metadata import MetaInformation, string_to_authors
 
@@ -52,6 +52,32 @@ def get_document_info(stream):
             break
     return data.getvalue(), pos
 
+def detect_codepage(stream):
+    pat = re.compile(r'\\ansicpg(\d+)')
+    match = pat.search(stream.read(512))
+    if match is not None:
+        num = match.group(1)
+        if num == '0':
+            num = '1250'
+        codec = 'cp'+num
+        try:
+            codecs.lookup(codec)
+            return codec
+        except:
+            pass
+
+def decode(raw, codec):
+    if codec is not None:
+        def codepage(match):
+            return chr(int(match.group(1), 16))
+        raw = re.sub(r"\\'([a-fA-F0-9]{2})", codepage, raw)
+        raw = raw.decode(codec)
+
+    def uni(match):
+        return unichr(int(match.group(1)))
+    raw = re.sub(r'\\u([0-9]{4}).', uni, raw)
+    return raw
+
 def get_metadata(stream):
     """ Return metadata as a L{MetaInfo} object """
     title, author, comment, category = None, None, None, None
@@ -62,18 +88,22 @@ def get_metadata(stream):
     if not block:
         return MetaInformation(None, None)
 
+    stream.seek(0)
+    cpg = detect_codepage(stream)
+    stream.seek(0)
+
     title_match = title_pat.search(block)
     if title_match:
-        title = title_match.group(1).strip()
+        title = decode(title_match.group(1).strip(), cpg)
     author_match = author_pat.search(block)
     if author_match:
-        author = author_match.group(1).strip()
+        author = decode(author_match.group(1).strip(), cpg)
     comment_match = comment_pat.search(block)
     if comment_match:
-        comment = comment_match.group(1).strip()
+        comment = decode(comment_match.group(1).strip(), cpg)
     category_match = category_pat.search(block)
     if category_match:
-        category = category_match.group(1).strip()
+        category = decode(category_match.group(1).strip(), cpg)
     mi = MetaInformation(title, author)
     if author:
         mi.authors = string_to_authors(author)