HTML parser based on BeautifulSoup4

diff --git a/doc/TODO b/doc/TODO
index e4fe65203b33b6bb9f1f6a9652b3e97312e13f82..deb79cab111efc8b36cdb6fa0fe12560c1c461ab 100644 (file)
--- a/doc/TODO
+++ b/doc/TODO
@@ -1,5 +1,3 @@
-HTML parser based on BeautifulSoup4. Bs3 for Python 2, bs4 for Py3.
-
 Replace subproc.py with some IPC. Or update for Python 3.
 
 Python 3.

diff --git a/parse_html/bkmk_parse_html.py b/parse_html/bkmk_parse_html.py
index 7bc4640d31ccad2a902335c8a2aad978b02bb09c..fa72e1f2aa14f5296343434895bcf6ccb22e088d 100644 (file)
--- a/parse_html/bkmk_parse_html.py
+++ b/parse_html/bkmk_parse_html.py
@@ -18,6 +18,14 @@ DEFAULT_CHARSET = "cp1251" # Stupid default for Russian Cyrillic
 
 parsers = []
 
+try:
+    from . import bkmk_ph_beautifulsoup4
+except ImportError:
+    pass
+else:
+    bkmk_ph_beautifulsoup4.DEFAULT_CHARSET = DEFAULT_CHARSET
+    parsers.append(bkmk_ph_beautifulsoup4.parse_html)
+
 try:
     from . import bkmk_ph_beautifulsoup
 except ImportError:
@@ -136,6 +144,17 @@ def parse_html(html_text, charset=None, log=None):
         p, parser = _parsers[0]
     if log: log("   Using %s" % p.__module__)
 
+    title = parser.title
+    if isinstance(title, unicode):
+        if parser.charset:
+            parser.title = title.encode(parser.charset)
+        else:
+            try:
+                parser.title = title.encode('ascii')
+            except UnicodeEncodeError:
+                parser.charset = DEFAULT_CHARSET
+                parser.title = title.encode(DEFAULT_CHARSET)
+
     converted_title = title = parser.title
     if title and (not parser.charset):
         try:

diff --git a/parse_html/bkmk_ph_beautifulsoup4.py b/parse_html/bkmk_ph_beautifulsoup4.py
new file mode 100644 (file)
index 0000000..aad3f8f
--- /dev/null
+++ b/parse_html/bkmk_ph_beautifulsoup4.py
@@ -0,0 +1,114 @@
+"""HTML Parser using BeautifulSoup4
+
+This file is a part of Bookmarks database and Internet robot.
+
+"""
+
+__author__ = "Oleg Broytman <phd@phdru.name>"
+__copyright__ = "Copyright (C) 2017 PhiloSoft Design"
+__license__ = "GNU GPL"
+
+__all__ = ['parse_html']
+
+
+import re
+from bs4 import BeautifulSoup
+from .bkmk_ph_util import HTMLParser
+
+DEFAULT_CHARSET = "cp1251" # Stupid default for Russian Cyrillic
+
+def _parse_html(html_text, charset):
+    try:
+        return BeautifulSoup(html_text, from_encoding=charset)
+    except TypeError:
+        return None
+
+def parse_html(html_text, charset=None, log=None):
+    root = _parse_html(html_text, charset)
+    if root is None:
+        return None
+
+    _charset = root.originalEncoding
+    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
+
+    title = head.title
+    if (title is None) and (html is not head):
+        # Some sites put TITLE in HTML outside of HEAD
+        title = html.title
+
+    if title is None:
+        # Lookup TITLE in the root
+        title = root.title
+
+    if title is not None:
+        if title.string:
+            title = title.string
+        else:
+            parts = []
+            for part in title:
+                if not isinstance(part, basestring):
+                    part = unicode(part)
+                parts.append(part.strip())
+            title = ''.join(parts)
+
+    meta = head.find(_find_contenttype, recursive=False)
+    if meta:
+        try:
+            meta_content = meta.get("content")
+            if meta_content:
+                __charset = meta_content.lower().split('charset=')[1].split(';')[0]
+            else:
+                __charset = False
+        except IndexError: # No charset in the META Content-Type
+            meta_charset = False
+        else:
+            meta_charset = _charset = __charset
+    else:
+        meta_charset = False
+
+    if not meta_charset:
+        meta = head.find(_find_charset, recursive=False)
+        if meta:
+            meta_content = meta.get("charset")
+            if meta_content:
+                meta_charset = _charset = meta_content.lower()
+
+    if title and (_charset or meta_charset):
+        title = title.encode(_charset or meta_charset)
+
+    meta = head.find(_find_refresh, recursive=False)
+    if meta:
+        refresh = meta.get("content")
+    else:
+        refresh = None
+
+    meta = head.find(_find_icon, recursive=False)
+    if meta:
+        icon = meta.get("href")
+    else:
+        icon = None
+
+    if (title is None) and (refresh is None) and (icon is None):
+        return None
+    return HTMLParser(_charset, meta_charset, title, refresh, icon)
+
+def _find_contenttype(Tag):
+    return (Tag.name == "meta") and \
+       (Tag.get_attribute_list("http-equiv", '')[0].lower() == "content-type")
+
+def _find_charset(Tag):
+    return (Tag.name == "meta") and Tag.get("charset", '')
+
+def _find_refresh(Tag):
+    return (Tag.name == "meta") and \
+       (Tag.get_attribute_list("http-equiv", '')[0].lower() == "refresh")
+
+def _find_icon(Tag):
+    return (Tag.name == "link") and \
+       (Tag.get_attribute_list("rel", '')[0].lower() in ('icon', 'shortcut icon'))