Git-Mirrors/brozzler
1
0
Fork 0
mirror of https://github.com/internetarchive/brozzler.git synced 2025-02-23 16:19:49 -05:00
Code Issues Actions Packages Projects Releases Wiki Activity

Merge pull request #28 from internetarchive/ari-3940

Browse Source 
Ari 3940 - prioritize scrolling all the way to the bottom
This commit is contained in:
vonrosen 2014-10-09 21:21:02 +00:00
commit 01ed5a7d4d
3 changed files with 25 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

30
umbra/behaviors.d/facebook.js
View File

@ -8,8 +8,14 @@ var umbraAboveBelowOrOnScreen = function(e) {
if (eTop < window.scrollY) {
return -1; // above
} else if (eTop > window.scrollY + window.innerHeight) {
// if (e.clientWidth != 0) {
// console.warn("e.clientWidth=" + e.clientWidth + " though it appears to be below the screen? e.getBoundingClientRect().top=" + eTop + " window.scrollY=" + window.scrollY + " window.innerHeight=" + window.innerHeight + " e=" + e);
// }
return 1; // below
} else {
// if (e.clientWidth != 0) {
// console.warn("e.clientWidth=" + e.clientWidth + " though it appears to be on screen? e.getBoundingClientRect().top=" + eTop + " window.scrollY=" + window.scrollY + " window.innerHeight=" + window.innerHeight + " e=" + e);
// }
return 0; // on screen
}
}
@ -17,7 +23,7 @@ var umbraAboveBelowOrOnScreen = function(e) {
// comments - 'a.UFIPagerLink > span, a.UFIPagerLink, span.UFIReplySocialSentenceLinkText'
var UMBRA_THINGS_TO_CLICK_SELECTOR = 'a[href^="/browse/likes"], *[rel="theater"]';
var umbraAlreadyClicked = {};
var umbraState = {'idleSince':null,'expectingSomething':null};
var umbraState = {'idleSince':null,'expectingSomething':null,'bottomReachedScrollY':0};
var umbraIntervalFunc = function() {
var closeButtons = document.querySelectorAll('a[title="Close"], a.closeTheater');
@ -70,18 +76,22 @@ var umbraIntervalFunc = function() {
}
}
if (window.scrollY > umbraState.bottomReachedScrollY) {
umbraState.bottomReachedScrollY = window.scrollY;
}
if (!clickedSomething) {
if (somethingLeftAbove) {
console.log("scrolling UP because everything on this screen has been clicked but we missed something above");
window.scrollBy(0, -500);
if (somethingLeftBelow) {
// console.log("scrolling down because everything on this screen has been clicked but there's more below document.body.clientHeight=" + document.body.clientHeight);
window.scrollBy(0, 300);
umbraState.idleSince = null;
} else if (somethingLeftBelow) {
console.log("scrolling because everything on this screen has been clicked but there's more below document.body.clientHeight=" + document.body.clientHeight);
window.scrollBy(0, 200);
} else if (umbraState.bottomReachedScrollY + window.innerHeight < document.documentElement.scrollHeight) {
// console.log("scrolling down because we haven't reached the bottom yet document.body.clientHeight=" + document.body.clientHeight);
window.scrollBy(0, 300);
umbraState.idleSince = null;
} else if (window.scrollY + window.innerHeight < document.documentElement.scrollHeight) {
console.log("scrolling because we're not to the bottom yet document.body.clientHeight=" + document.body.clientHeight);
window.scrollBy(0, 200);
} else if (somethingLeftAbove) {
// console.log("scrolling UP because we've already been to the bottom, everything on or below this screen has been clicked, but we missed something above");
window.scrollBy(0, -600);
umbraState.idleSince = null;
} else if (umbraState.idleSince == null) {
umbraState.idleSince = Date.now();

3
umbra/behaviors.py
View File

@ -68,7 +68,8 @@ class Behavior:
self.notify_of_activity()
def is_finished(self):
msg_id = self.umbra_worker.send_to_chrome(method="Runtime.evaluate", params={"expression": "umbraBehaviorFinished()"})
msg_id = self.umbra_worker.send_to_chrome(method="Runtime.evaluate",
suppress_logging=True, params={"expression":"umbraBehaviorFinished()"})
self.waiting_result_msg_ids.append(msg_id)
request_idle_timeout_sec = 30

5
umbra/browser.py
View File

@ -144,11 +144,12 @@ class Browser:
self._behavior = None
def send_to_chrome(self, **kwargs):
def send_to_chrome(self, suppress_logging=False, **kwargs):
msg_id = next(self.command_id)
kwargs['id'] = msg_id
msg = json.dumps(kwargs)
self.logger.debug('sending message to {}: {}'.format(self._websock, msg))
if not suppress_logging:
self.logger.debug('sending message to {}: {}'.format(self._websock, msg))
self._websock.send(msg)
return msg_id