monkey-patch youtube-dl to short-circuit

video extraction using generic extractor in case of very large url (more
than 20 mb) that youtube-dl interprets as html, to avoid spinning
forever here:

Traceback (most recent call first):
  File "/opt/brozzler-ve3/lib/python3.5/re.py", line 213, in findall
    return _compile(pattern, flags).findall(string)
  File "/opt/brozzler-ve3/lib/python3.5/site-packages/youtube_dl/extractor/generic.py", line 2878, in _real_extract
    'uploader': video_uploader,
  File "/opt/brozzler-ve3/lib/python3.5/site-packages/youtube_dl/extractor/common.py", line 503, in extract
    ie_result = self._real_extract(url)
  File "/opt/brozzler-ve3/lib/python3.5/site-packages/youtube_dl/YoutubeDL.py", line 792, in extract_info
    ie_result = ie.extract(url)
  File "/opt/brozzler-ve3/lib/python3.5/site-packages/brozzler/worker.py", line 302, in _try_youtube_dl
    info = ydl.extract_info(str(urlcanon.whatwg(page.url)))
  File "/opt/brozzler-ve3/lib/python3.5/site-packages/brozzler/worker.py", line 361, in brozzle_page
    self._try_youtube_dl(ydl, site, page)

brozzler/worker.py

@@ -39,6 +39,17 @@ import rethinkdb as r
 import datetime
 import urllib.parse
 
+_orig_webpage_read_content = youtube_dl.extractor.generic.GenericIE._webpage_read_content
+def _webpage_read_content(self, *args, **kwargs):
+    content = _orig_webpage_read_content(self, *args, **kwargs)
+    if len(content) > 20000000:
+        logging.warn(
+                'bypassing youtube-dl extraction because content is '
+                'too large (%s characters)', len(content))
+        return ''
+    return content
+youtube_dl.extractor.generic.GenericIE._webpage_read_content = _webpage_read_content
+
 class ExtraHeaderAdder(urllib.request.BaseHandler):
     def __init__(self, extra_headers):
         self.extra_headers = extra_headers

setup.py

@@ -32,7 +32,7 @@ def find_package_data(package):
 
 setuptools.setup(
         name='brozzler',
-        version='1.1b13.dev290',
+        version='1.1b13.dev291',
         description='Distributed web crawling with browsers',
         url='https://github.com/internetarchive/brozzler',
         author='Noah Levitt',