1
0
mirror of https://github.com/mikf/gallery-dl.git synced 2024-11-22 18:53:21 +01:00

[directlink] improve URL pattern

This commit is contained in:
Mike Fährmann 2017-08-02 21:06:49 +02:00
parent d443822fdb
commit c4713404c8
No known key found for this signature in database
GPG Key ID: 5680CA389D365A88

View File

@ -16,22 +16,28 @@ class DirectlinkExtractor(Extractor):
"""Extractor for direct links to images and other media files"""
category = "directlink"
filename_fmt = "{domain}/{path}"
pattern = [r"https?://([^/]+)/([^?&#]+\."
r"(?:jpe?g|jpe|png|gif|web[mp]|mp4|mkv|og[gmv]|opus))"]
test = [(("https://photos.smugmug.com/The-World/Hawaii/"
"i-SWz2K6n/2/X3/IMG_0311-X3.jpg"), {
"url": "32ee1045881e17ef3f13a9958595afa42421ec6c",
"keyword": "1abd2f2c115cdf2cf2671d2611349b4213c3ab3e",
})]
pattern = [r"https?://(?P<domain>[^/]+)/(?P<path>[^?&#]+\."
r"(?:jpe?g|jpe|png|gif|web[mp]|mp4|mkv|og[gmv]|opus))"
r"(?:\?(?P<query>[^/?#]*))?(?:#(?P<fragment>.*))?$"]
test = [
(("https://photos.smugmug.com/The-World/Hawaii/"
"i-SWz2K6n/2/X3/IMG_0311-X3.jpg"), {
"url": "32ee1045881e17ef3f13a9958595afa42421ec6c",
"keyword": "2427b68c14006489df1776bb1bcd3bc24be25e10",
}),
("https://example.org/path/file.webm?que=1&ry=2#fragment", {
"url": "fd4aec8a32842343394e6078a06c3e6b647bf671",
"keyword": "ed008f35fc18dddb2f448a18d160c949bb3b054c",
}),
]
def __init__(self, match):
Extractor.__init__(self)
self.domain, self.path = match.groups()
self.data = match.groupdict()
self.url = match.string
def items(self):
data = {"domain": self.domain, "path": self.path}
text.nameext_from_url(self.url, data)
text.nameext_from_url(self.url, self.data)
yield Message.Version, 1
yield Message.Directory, data
yield Message.Url, self.url, data
yield Message.Directory, self.data
yield Message.Url, self.url, self.data