Git-Mirrors/brozzler
1
0
Fork 0
mirror of https://github.com/internetarchive/brozzler.git synced 2025-08-06 21:44:29 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'master' into qa

Browse source 
* master:
  bump dev version number after some PR merges
  bugfix for BrozzlerWorker._needs_browsing
  Remove redundant method parameter.
  bugfix
  Make youtube-dl optional in BrozzlerWorker.brozzle_page
This commit is contained in:
Noah Levitt 2017-08-01 12:05:07 -07:00
commit 5be7dd4407
3 changed files with 43 additions and 35 deletions
Download patch file Download diff file Expand all files Collapse all files

2
brozzler/browser.py
View file

@ -376,7 +376,7 @@ class Browser:
return self.websock_url is not None
def browse_page(
self, page_url, ignore_cert_errors=False, extra_headers=None,
self, page_url, extra_headers=None,
user_agent=None, behavior_parameters=None,
on_request=None, on_response=None, on_screenshot=None,
username=None, password=None, hashtags=None,

74
brozzler/worker.py
View file

@ -316,30 +316,34 @@ class BrozzlerWorker:
return full_jpeg, thumb_jpeg
def brozzle_page(self, browser, site, page, on_screenshot=None):
def brozzle_page(self, browser, site, page, on_screenshot=None,
enable_youtube_dl=True):
self.logger.info("brozzling {}".format(page))
try:
with tempfile.TemporaryDirectory(prefix='brzl-ydl-') as tempdir:
ydl = self._youtube_dl(tempdir, site)
ydl_spy = ydl.brozzler_spy # remember for later
self._try_youtube_dl(ydl, site, page)
except brozzler.ReachedLimit as e:
raise
except brozzler.ShutdownRequested:
raise
except brozzler.ProxyError:
raise
except Exception as e:
if (hasattr(e, 'exc_info') and len(e.exc_info) >= 2
and hasattr(e.exc_info[1], 'code')
and e.exc_info[1].code == 430):
self.logger.info(
'youtube-dl got %s %s processing %s',
e.exc_info[1].code, e.exc_info[1].msg, page.url)
else:
self.logger.error(
'youtube_dl raised exception on %s', page,
exc_info=True)
if enable_youtube_dl:
try:
with tempfile.TemporaryDirectory(prefix='brzl-ydl-') as tempdir:
ydl = self._youtube_dl(tempdir, site)
ydl_spy = ydl.brozzler_spy # remember for later
self._try_youtube_dl(ydl, site, page)
except brozzler.ReachedLimit as e:
raise
except brozzler.ShutdownRequested:
raise
except brozzler.ProxyError:
raise
except Exception as e:
if (hasattr(e, 'exc_info') and len(e.exc_info) >= 2
and hasattr(e.exc_info[1], 'code')
and e.exc_info[1].code == 430):
self.logger.info(
'youtube-dl got %s %s processing %s',
e.exc_info[1].code, e.exc_info[1].msg, page.url)
else:
self.logger.error(
'youtube_dl raised exception on %s', page,
exc_info=True)
else:
ydl_spy = False
if self._needs_browsing(page, ydl_spy):
self.logger.info('needs browsing: %s', page)
@ -435,19 +439,23 @@ class BrozzlerWorker:
'proxy error fetching %s' % page.url) from e
def _needs_browsing(self, page, brozzler_spy):
final_bounces = brozzler_spy.final_bounces(page.url)
if not final_bounces:
return True
for txn in final_bounces:
if txn['response_headers'].get_content_type() in [
'text/html', 'application/xhtml+xml']:
if brozzler_spy:
final_bounces = brozzler_spy.final_bounces(page.url)
if not final_bounces:
return True
return False
for txn in final_bounces:
if txn['response_headers'].get_content_type() in [
'text/html', 'application/xhtml+xml']:
return True
return False
else:
return True
def _already_fetched(self, page, brozzler_spy):
for txn in brozzler_spy.final_bounces(page.url):
if (txn['method'] == 'GET' and txn['status_code'] == 200):
return True
if brozzler_spy:
for txn in brozzler_spy.final_bounces(page.url):
if (txn['method'] == 'GET' and txn['status_code'] == 200):
return True
return False
def brozzle_site(self, browser, site):

2
setup.py
View file

@ -32,7 +32,7 @@ def find_package_data(package):
setuptools.setup(
name='brozzler',
version='1.1b12.dev265',
version='1.1b12.dev266',
description='Distributed web crawling with browsers',
url='https://github.com/internetarchive/brozzler',
author='Noah Levitt',