Split the title into subparts, reassemble the subparts and recode.

[bookmarks_db.git] / Robots / parse_html_beautifulsoup.py

diff --git a/Robots/parse_html_beautifulsoup.py b/Robots/parse_html_beautifulsoup.py
index c45af2b0221a61f441cf5165fc2a25f6fe978bb5..db291f9fd46df8607709333dbaa304eaba00007e 100644 (file)
--- a/Robots/parse_html_beautifulsoup.py
+++ b/Robots/parse_html_beautifulsoup.py
@@ -39,7 +39,7 @@ class BadDeclParser(BeautifulSoup):
              except SGMLParseError:
                  # Could not parse the DOCTYPE declaration
                  # Try to just skip the actual declaration
-                 match = re.search(r'<!DOCTYPE([^>]*?)>', self.rawdata[i:], re.MULTILINE)
+                 match = re.search(r'<!DOCTYPE([^>]*?)>', self.rawdata[i:], re.MULTILINE|re.IGNORECASE)
                  if match:
                      toHandle = self.rawdata[i:match.end()]
                  else:
@@ -49,44 +49,64 @@ class BadDeclParser(BeautifulSoup):
          return j
 
 
-def parse_html(filename, charset=None):
+def _parse_html(filename, charset):
    infile = open(filename, 'r')
    try:
-      root = BadDeclParser(infile, fromEncoding=charset)
+      return BadDeclParser(infile, fromEncoding=charset)
    except TypeError:
       return None
    finally:
       infile.close()
 
-   try:
-      head = root.html.head
-   except AttributeError:
+def parse_html(filename, charset=None, log=None):
+   root = _parse_html(filename, charset)
+   if root is None:
       return None
 
-   if head is None:
-      head = root.html # Some sites put TITLE in HTML without HEAD
-
    _charset = root.originalEncoding
-   if _charset == "windows-1252": # Replace default
+   if _charset in ("ISO-8859-2", "windows-1252", "MacCyrillic"): # Replace default
       _charset = DEFAULT_CHARSET
+      root = _parse_html(filename, _charset)
+      if root is None:
+         return None
 
-   try:
-      title = head.title.string.encode(_charset)
-   except AttributeError:
-      title = '' # HEAD but no TITLE
+   html = root.html
+   if html is None:
+      html = root
+
+   head = html.head
+   if head is None:
+      head = html # Some sites put TITLE in HTML without HEAD
 
-   if (not title) and (head is not root.html):
+   title = head.title
+   if (title is None) and (html is not head):
       # Some sites put TITLE in HTML outside of HEAD
+      title = html.title
 
-      try:
-         title = root.html.title.string.encode(_charset)
-      except AttributeError:
-         title = '' # no TITLE in HTML too
+   if title is None:
+      # Lookup TITLE in the root
+      title = root.title
+
+   if title is None:
+      title = ''
+   elif title.string:
+      title = title.string.encode(_charset)
+   else:
+      parts = []
+      for part in title:
+         if not isinstance(part, basestring):
+            part = unicode(part)
+         parts.append(part.strip())
+      title = ''.join(parts).encode(_charset)
 
    meta = head.find(_find_contenttype, recursive=False)
    if meta:
-      __charset = meta.get("content").lower().split('charset=')[1].split(';')[0]
-      meta_charset = _charset == __charset
+      try:
+         __charset = meta.get("content").lower().split('charset=')[1].split(';')[0]
+      except IndexError: # No charset in the META Content-Type
+         meta_charset = False
+      else:
+         meta_charset = _charset == __charset
    else:
       meta_charset = False