new test of frontier.seed_page

This commit is contained in:
Noah Levitt 2017-03-24 15:45:40 -07:00
parent 0e35de43b6
commit a826fdc7ef
2 changed files with 25 additions and 1 deletions

View File

@ -32,7 +32,7 @@ def find_package_data(package):
setuptools.setup( setuptools.setup(
name='brozzler', name='brozzler',
version='1.1b10.dev220', version='1.1b10.dev222',
description='Distributed web crawling with browsers', description='Distributed web crawling with browsers',
url='https://github.com/internetarchive/brozzler', url='https://github.com/internetarchive/brozzler',
author='Noah Levitt', author='Noah Levitt',

View File

@ -567,3 +567,27 @@ def test_completed_page():
assert page.brozzle_count == 1 assert page.brozzle_count == 1
assert page.claimed == False assert page.claimed == False
def test_seed_page():
rr = doublethink.Rethinker('localhost', db='ignoreme')
frontier = brozzler.RethinkDbFrontier(rr)
site = brozzler.Site(rr, {'seed':'http://example.com/a/'})
site.save()
assert frontier.seed_page(site.id) is None
page1 = brozzler.Page(rr, {
'site_id': site.id,
'url': 'http://example.com/a/b/',
'hops_from_seed': 1})
page1.save()
assert frontier.seed_page(site.id) is None
page0 = brozzler.Page(rr, {
'site_id': site.id,
'url': 'http://example.com/a/',
'hops_from_seed': 0})
page0.save()
assert frontier.seed_page(site.id) == page0