Add infrastructure for paged lists

This commit allows to download pages in playlists as needed instead of all at once.
Before this commit,
    youtube-dl http://www.youtube.com/user/ANNnewsCH/videos --playlist-end 2 --skip-download
took quite some time - now it's almost instantaneous.
As an example, the youtube:user extractor has been converted.
Fixes #2175
This commit is contained in:
Philipp Hagemeister 2014-01-20 11:36:47 +01:00
parent c91778f8c0
commit b7ab059084
4 changed files with 92 additions and 25 deletions

View file

@ -18,6 +18,7 @@ from youtube_dl.utils import (
find_xpath_attr,
get_meta_content,
orderedSet,
PagedList,
parse_duration,
sanitize_filename,
shell_quote,
@ -200,5 +201,26 @@ class TestUtil(unittest.TestCase):
self.assertEqual(parse_duration('9:12:43'), 33163)
self.assertEqual(parse_duration('x:y'), None)
def test_paged_list(self):
def testPL(size, pagesize, sliceargs, expected):
def get_page(pagenum):
firstid = pagenum * pagesize
upto = min(size, pagenum * pagesize + pagesize)
for i in range(firstid, upto):
yield i
pl = PagedList(get_page, pagesize)
got = pl.getslice(*sliceargs)
self.assertEqual(got, expected)
testPL(5, 2, (), [0, 1, 2, 3, 4])
testPL(5, 2, (1,), [1, 2, 3, 4])
testPL(5, 2, (2,), [2, 3, 4])
testPL(5, 2, (4,), [4])
testPL(5, 2, (0, 3), [0, 1, 2])
testPL(5, 2, (1, 4), [1, 2, 3])
testPL(5, 2, (2, 99), [2, 3, 4])
testPL(5, 2, (20, 99), [])
if __name__ == '__main__':
unittest.main()