Git-Mirrors/brozzler
1
0
Fork 0
mirror of https://github.com/internetarchive/brozzler.git synced 2025-08-09 15:02:28 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'master' into qa

Browse source 
* master:
  add --yaml option to brozzler-list-* commands
  take screenshot before running behavior (but after login) - thanks danielbicho
This commit is contained in:
Noah Levitt 2017-02-15 23:13:32 +00:00
commit f6fdb91d57
3 changed files with 42 additions and 31 deletions
Download patch file Download diff file Expand all files Collapse all files

18
brozzler/browser.py
View file

@ -426,13 +426,12 @@ class Browser:
user_agent=user_agent, timeout=300)
if password:
self.try_login(username, password, timeout=300)
if on_screenshot:
jpeg_bytes = self.screenshot()
on_screenshot(jpeg_bytes)
behavior_script = brozzler.behavior_script(
page_url, behavior_parameters)
self.run_behavior(behavior_script, timeout=900)
if on_screenshot:
self.scroll_to_top()
jpeg_bytes = self.screenshot()
on_screenshot(jpeg_bytes)
outlinks = self.extract_outlinks()
## for each hashtag not already visited:
## navigate_to_hashtag (nothing to wait for so no timeout?)
@ -502,17 +501,6 @@ class Browser:
jpeg_bytes = base64.b64decode(message['result']['data'])
return jpeg_bytes
def scroll_to_top(self, timeout=30):
self.logger.info('scrolling to top')
self.websock_thread.expect_result(self._command_id.peek())
msg_id = self.send_to_chrome(
method='Runtime.evaluate',
params={'expression': 'window.scrollTo(0, 0);'})
self._wait_for(
lambda: self.websock_thread.received_result(msg_id),
timeout=timeout)
self.websock_thread.pop_result(msg_id)
def url(self, timeout=30):
'''
Returns value of document.URL from the browser.