merged unescapeHTML branch; removed lxml dependency

devscripts/wine-py2exe.sh

@@ -18,7 +18,6 @@ if [ ! -d wine-py2exe ]; then
 
     axel -a "http://www.python.org/ftp/python/2.7/python-2.7.msi"
     axel -a "http://downloads.sourceforge.net/project/py2exe/py2exe/0.6.9/py2exe-0.6.9.win32-py2.7.exe"
-    axel -a "http://pypi.python.org/packages/2.7/l/lxml/lxml-2.3.win32-py2.7.exe"
     #axel -a "http://winetricks.org/winetricks"
 
     # http://appdb.winehq.org/objectManager.php?sClass=version&iId=21957
@@ -28,13 +27,9 @@ if [ ! -d wine-py2exe ]; then
     echo "Follow py2exe setup on screen"
     wine py2exe-0.6.9.win32-py2.7.exe
     
-    echo "Follow lxml setup on screen"
-    wine lxml-2.3.win32-py2.7.exe
-    
     #echo "Follow Microsoft Visual C++ 2008 Redistributable Package setup on screen"
     #bash winetricks vcrun2008
 
-    rm lxml-2.3.win32-py2.7.exe
     rm py2exe-0.6.9.win32-py2.7.exe
     rm python-2.7.msi
     #rm winetricks

youtube_dl/InfoExtractors.py

@@ -24,11 +24,6 @@
 except ImportError:
 	from cgi import parse_qs
 
-try:
-	import lxml.etree
-except ImportError:
-	pass # Handled below
-
 try:
 	import xml.etree.ElementTree
 except ImportError: # Python<2.5: Not officially supported, but let it slip
@@ -193,8 +188,8 @@ def _closed_captions_xml_to_srt(self, xml_string):
 			end = start + float(dur)
 			start = "%02i:%02i:%02i,%03i" %(start/(60*60), start/60%60, start%60, start%1*1000)
 			end = "%02i:%02i:%02i,%03i" %(end/(60*60), end/60%60, end%60, end%1*1000)
-			caption = re.sub(ur'(?u)&(.+?);', htmlentity_transform, caption)
+			caption = unescapeHTML(caption)
-			caption = re.sub(ur'(?u)&(.+?);', htmlentity_transform, caption) # double cycle, inentional
+			caption = unescapeHTML(caption) # double cycle, inentional
 			srt += str(n) + '\n'
 			srt += start + ' --> ' + end + '\n'
 			srt += caption + '\n\n'
@@ -364,18 +359,9 @@ def _real_extract(self, url):
 					pass
 
 		# description
-		try:
+		video_description = get_element_by_id("eow-description", video_webpage)
-			lxml.etree
+		if video_description: video_description = clean_html(video_description.decode('utf8'))
-		except NameError:
+		else: video_description = ''
-			video_description = u'No description available.'
-			mobj = re.search(r'<meta name="description" content="(.*?)">', video_webpage)
-			if mobj is not None:
-				video_description = mobj.group(1).decode('utf-8')
-		else:
-			html_parser = lxml.etree.HTMLParser(encoding='utf-8')
-			vwebpage_doc = lxml.etree.parse(StringIO.StringIO(video_webpage), html_parser)
-			video_description = u''.join(vwebpage_doc.xpath('id("eow-description")//text()'))
-			# TODO use another parser
 			
 		# closed captions
 		video_subtitles = None
@@ -992,7 +978,7 @@ def _real_extract(self, url, new_video=True):
 			self._downloader.trouble(u'ERROR: Unable to extract media URL')
 			return
 		video_url = urllib.unquote(mobj.group(1) + mobj.group(2)).decode('utf-8')
-		video_url = re.sub(r'(?u)&(.+?);', htmlentity_transform, video_url)
+		video_url = unescapeHTML(video_url)
 
 		return [{
 			'id':		video_id.decode('utf-8'),
@@ -1069,18 +1055,9 @@ def _real_extract(self, url, new_video=True):
 		video_thumbnail = config["video"]["thumbnail"]
 
 		# Extract video description
-		try:
+		video_description = get_element_by_id("description", webpage)
-			lxml.etree
+		if video_description: video_description = clean_html(video_description.decode('utf8'))
-		except NameError:
+		else: video_description = ''
-			video_description = u'No description available.'
-			mobj = re.search(r'<meta name="description" content="(.*?)" />', webpage, re.MULTILINE)
-			if mobj is not None:
-				video_description = mobj.group(1)
-		else:
-			html_parser = lxml.etree.HTMLParser()
-			vwebpage_doc = lxml.etree.parse(StringIO.StringIO(webpage), html_parser)
-			video_description = u''.join(vwebpage_doc.xpath('id("description")//text()')).strip()
-			# TODO use another parser
 
 		# Extract upload date
 		video_upload_date = u'NA'
@@ -2248,8 +2225,6 @@ def report_config_download(self, showName):
 		self._downloader.to_screen(u'[escapist] %s: Downloading configuration' % showName)
 
 	def _real_extract(self, url):
-		htmlParser = HTMLParser.HTMLParser()
-
 		mobj = re.match(self._VALID_URL, url)
 		if mobj is None:
 			self._downloader.trouble(u'ERROR: invalid URL: %s' % url)
@@ -2265,11 +2240,11 @@ def _real_extract(self, url):
 			return
 
 		descMatch = re.search('<meta name="description" content="([^"]*)"', webPage)
-		description = htmlParser.unescape(descMatch.group(1))
+		description = unescapeHTML(descMatch.group(1))
 		imgMatch = re.search('<meta property="og:image" content="([^"]*)"', webPage)
-		imgUrl = htmlParser.unescape(imgMatch.group(1))
+		imgUrl = unescapeHTML(imgMatch.group(1))
 		playerUrlMatch = re.search('<meta property="og:video" content="([^"]*)"', webPage)
-		playerUrl = htmlParser.unescape(playerUrlMatch.group(1))
+		playerUrl = unescapeHTML(playerUrlMatch.group(1))
 		configUrlMatch = re.search('config=(.*)$', playerUrl)
 		configUrl = urllib2.unquote(configUrlMatch.group(1))
 
@@ -2324,8 +2299,6 @@ def report_extraction(self, video_id):
 		self._downloader.to_screen(u'[%s] %s: Extracting information' % (self.IE_NAME, video_id))
 
 	def _real_extract(self, url):
-		htmlParser = HTMLParser.HTMLParser()
-
 		mobj = re.match(self._VALID_URL, url)
 		if mobj is None:
 			self._downloader.trouble(u'ERROR: invalid URL: %s' % url)
@@ -2391,8 +2364,6 @@ def report_extraction(self, video_id):
 		self._downloader.to_screen(u'[%s] %s: Extracting information' % (self.IE_NAME, video_id))
 
 	def _real_extract(self, url):
-		htmlParser = HTMLParser.HTMLParser()
-
 		mobj = re.match(self._VALID_URL, url)
 		if mobj is None:
 			self._downloader.trouble(u'ERROR: invalid URL: %s' % url)
@@ -2475,8 +2446,6 @@ def report_extraction(self, video_id):
 		self._downloader.to_screen(u'[%s] %s: Extracting information' % (self.IE_NAME, video_id))
 
 	def _real_extract(self, url):
-		htmlParser = HTMLParser.HTMLParser()
-
 		mobj = re.match(self._VALID_URL, url)
 		if mobj is None:
 			self._downloader.trouble(u'ERROR: invalid URL: %s' % url)
@@ -2561,8 +2530,6 @@ def report_extraction(self, video_id):
 		self._downloader.to_screen(u'[%s] %s: Extracting information' % (self.IE_NAME, video_id))
 
 	def _real_extract(self, url):
-		htmlParser = HTMLParser.HTMLParser()
-
 		mobj = re.match(self._VALID_URL, url)
 		if mobj is None:
 			self._downloader.trouble(u'ERROR: invalid URL: %s' % url)
@@ -2782,8 +2749,6 @@ def _real_extract(self, url):
 			info['format'] = info['ext']
 			return [info]
 		elif mobj.group('course'): # A course page
-			unescapeHTML = HTMLParser.HTMLParser().unescape
-
 			course = mobj.group('course')
 			info = {
 				'id': simplify_title(course),
@@ -2822,8 +2787,6 @@ def _real_extract(self, url):
 			return results
 			
 		else: # Root page
-			unescapeHTML = HTMLParser.HTMLParser().unescape
-
 			info = {
 				'id': 'Stanford OpenClassroom',
 				'type': 'playlist',

youtube_dl/utils.py

@@ -74,9 +74,86 @@ def htmlentity_transform(matchobj):
 	return (u'&%s;' % entity)
 
 
+class IDParser(HTMLParser.HTMLParser):
+	"""Modified HTMLParser that isolates a tag with the specified id"""
+	def __init__(self, id):
+		self.id = id
+		self.result = None
+		self.started = False
+		self.depth = {}
+		self.html = None
+		self.watch_startpos = False
+		HTMLParser.HTMLParser.__init__(self)
+
+	def loads(self, html):
+		self.html = html
+		self.feed(html)
+		self.close()
+
+	def handle_starttag(self, tag, attrs):
+		attrs = dict(attrs)
+		if self.started:
+			self.find_startpos(None)
+		if 'id' in attrs and attrs['id'] == self.id:
+			self.result = [tag]
+			self.started = True
+			self.watch_startpos = True
+		if self.started:
+			if not tag in self.depth: self.depth[tag] = 0
+			self.depth[tag] += 1
+
+	def handle_endtag(self, tag):
+		if self.started:
+			if tag in self.depth: self.depth[tag] -= 1
+			if self.depth[self.result[0]] == 0:
+				self.started = False
+				self.result.append(self.getpos())
+
+	def find_startpos(self, x):
+		"""Needed to put the start position of the result (self.result[1])
+		after the opening tag with the requested id"""
+		if self.watch_startpos:
+			self.watch_startpos = False
+			self.result.append(self.getpos())
+	handle_entityref = handle_charref = handle_data = handle_comment = \
+	handle_decl = handle_pi = unknown_decl = find_startpos
+
+	def get_result(self):
+		if self.result == None: return None
+		if len(self.result) != 3: return None
+		lines = self.html.split('\n')
+		lines = lines[self.result[1][0]-1:self.result[2][0]]
+		lines[0] = lines[0][self.result[1][1]:]
+		if len(lines) == 1:
+			lines[-1] = lines[-1][:self.result[2][1]-self.result[1][1]]
+		lines[-1] = lines[-1][:self.result[2][1]]
+		return '\n'.join(lines).strip()
+
+def get_element_by_id(id, html):
+	"""Return the content of the tag with the specified id in the passed HTML document"""
+	parser = IDParser(id)
+	try:
+		parser.loads(html)
+	except HTMLParser.HTMLParseError:
+		pass
+	return parser.get_result()
+
+
+def clean_html(html):
+	"""Clean an HTML snippet into a readable string"""
+	# Newline vs <br />
+	html = html.replace('\n', ' ')
+	html = re.sub('\s*<\s*br\s*/?\s*>\s*', '\n', html)
+	# Strip html tags
+	html = re.sub('<.*?>', '', html)
+	# Replace html entities
+	html = unescapeHTML(html)
+	return html
+
+
 def sanitize_title(utitle):
 	"""Sanitizes a video title so it could be used as part of a filename."""
-	utitle = re.sub(ur'(?u)&(.+?);', htmlentity_transform, utitle)
+	utitle = unescapeHTML(utitle)
 	return utitle.replace(unicode(os.sep), u'%')
 
 
@@ -133,8 +210,8 @@ def unescapeHTML(s):
 	"""
 	assert type(s) == type(u'')
 
-	htmlParser = HTMLParser.HTMLParser()
+	result = re.sub(ur'(?u)&(.+?);', htmlentity_transform, s)
-	return htmlParser.unescape(s)
+	return result
 
 def encodeFilename(s):
 	"""