Git-Mirrors/brozzler
1
0
Fork 0
mirror of https://github.com/internetarchive/brozzler.git synced 2025-02-24 00:29:53 -05:00
Code Issues Actions Packages Projects Releases Wiki Activity

convert mouseovers and simpleclicks to jinja2

Browse Source
This commit is contained in:
Noah Levitt 2016-12-20 17:34:29 -08:00
parent e2dbf68ccd
commit edf0a3a50d
4 changed files with 12 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
brozzler/behaviors.yaml
View File

@ -42,36 +42,42 @@
behavior_js_template: simpleclicks.js.j2
default_parameters:
click_css_selector: img.img-responsive
click_until_hard_timeout: False
request_idle_timeout_sec: 10
- # acalog https://webarchive.jira.com/browse/ARI-3775
url_regex: '^https?://.*[?&]catoid=[^?]*$'
behavior_js_template: simpleclicks.js.j2
default_parameters:
click_css_selector: a[onclick]
click_until_hard_timeout: False
request_idle_timeout_sec: 10
- # https://webarchive.jira.com/browse/ARI-3956
url_regex: '^https?://(?:www\.)?usask.ca/.*$'
behavior_js_template: simpleclicks.js.j2
default_parameters:
click_css_selector: a[id='feature-next']
click_until_hard_timeout: False
request_idle_timeout_sec: 10
- # https://webarchive.jira.com/browse/AITFIVE-451
url_regex: '^https?://(?:www\.)?soundcloud.com/.*$'
behavior_js_template: simpleclicks.js.j2
default_parameters:
click_css_selector: button.sc-button-play, button.playButton
click_until_hard_timeout: False
request_idle_timeout_sec: 10
- # https://webarchive.jira.com/browse/AITFIVE-463
url_regex: '^https?://(?:www\.)?christophercerrone.com/.*$'
behavior_js_template: simpleclicks.js.j2
default_parameters:
click_css_selector: button.playButton.medium
click_until_hard_timeout: False
request_idle_timeout_sec: 10
- # https://webarchive.jira.com/browse/ARI-4690
url_regex: '^https?://(?:www\.)?youtube.com/.*$'
behavior_js_template: simpleclicks.js.j2
default_parameters:
click_css_selector: span.load-more-text
click_until_hard_timeout: False
request_idle_timeout_sec: 10
- # https://webarchive.jira.com/browse/ARI-4725
url_regex: '^https?://(?:www\.)?moma.org/.*$'
@ -88,6 +94,7 @@
behavior_js_template: mouseovers.js.j2
default_parameters:
mouseover_css_selector: .menu-item a
mouseover_until_hard_timeout: False
request_idle_timeout_sec: 10
- # default fallback behavior
url_regex: '^.*$'

7
brozzler/js-templates/mouseovers.js.j2
View File

@ -26,11 +26,8 @@ var umbraBehavior = {
var mouseoveredSomething = false;
var somethingLeftBelow = false;
var somethingLeftAbove = false;
var cssSelector = "${mouseover_css_selector}";
var mouseoverUntilTimeout = "${mouseover_until_hard_timeout}";
//handle Python to JavaScript boolean conversion
mouseoverUntilTimeout == "True" ? mouseoverUntilTimeout = true : mouseoverUntilTimeout = false;
var cssSelector = {{mouseover_css_selector|json}};
var mouseoverUntilTimeout = {{mouseover_until_hard_timeout|json}};
var iframes = document.querySelectorAll("iframe");
var documents = Array(iframes.length + 1);

7
brozzler/js-templates/simpleclicks.js.j2
View File

@ -26,11 +26,8 @@ var umbraBehavior = {
var clickedSomething = false;
var somethingLeftBelow = false;
var somethingLeftAbove = false;
var cssSelector = "${click_css_selector}";
var clickUntilTimeout = "${click_until_hard_timeout}";
//handle Python to JavaScript boolean conversion
clickUntilTimeout == "True" ? clickUntilTimeout = true : clickUntilTimeout = false;
var cssSelector = {{click_css_selector|json}};
var clickUntilTimeout = {{click_until_hard_timeout|json}};
var iframes = document.querySelectorAll("iframe");
var documents = Array(iframes.length + 1);

2
setup.py
View File

@ -32,7 +32,7 @@ def find_package_data(package):
setuptools.setup(
name='brozzler',
version='1.1b9.dev150',
version='1.1b9.dev151',
description='Distributed web crawling with browsers',
url='https://github.com/internetarchive/brozzler',
author='Noah Levitt',