diff options
Diffstat (limited to 'gallery_dl/extractor/directlink.py')
| -rw-r--r-- | gallery_dl/extractor/directlink.py | 32 |
1 files changed, 20 insertions, 12 deletions
diff --git a/gallery_dl/extractor/directlink.py b/gallery_dl/extractor/directlink.py index 77a19f6..80db096 100644 --- a/gallery_dl/extractor/directlink.py +++ b/gallery_dl/extractor/directlink.py @@ -15,25 +15,30 @@ from .. import text class DirectlinkExtractor(Extractor): """Extractor for direct links to images and other media files""" category = "directlink" - filename_fmt = "{domain}/{path}" - archive_fmt = "{domain}/{path}" + filename_fmt = "{domain}/{path}/{filename}.{extension}" + archive_fmt = filename_fmt pattern = (r"(?i)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://en.wikipedia.org/static/images/project-logos/enwiki.png"), { "url": "18c5d00077332e98e53be9fed2ee4be66154b88d", - "keyword": "e81b9fe3022e971365dd859f38e4ef717a6c69ed", + "keyword": "105770a3f4393618ab7b811b731b22663b5d3794", + }), + # empty path + (("https://example.org/file.webm"), { + "url": "2d807ed7059d1b532f1bb71dc24b510b80ff943f", + "keyword": "29dad729c40fb09349f83edafa498dba1297464a", }), # more complex example - ("https://example.org/path/file.webm?que=1&ry=2#fragment", { - "url": "fd4aec8a32842343394e6078a06c3e6b647bf671", - "keyword": "ff75764b1ae66615b723a6357b8193fa2de84678", + ("https://example.org/path/to/file.webm?que=1&ry=2#fragment", { + "url": "114b8f1415cc224b0f26488ccd4c2e7ce9136622", + "keyword": "06014abd503e3b2b58aa286f9bdcefdd2ae336c0", }), # percent-encoded characters ("https://example.org/%27%3C%23/%23%3E%27.jpg?key=%3C%26%3E", { "url": "2627e8140727fdf743f86fe18f69f99a052c9718", - "keyword": "4d19dc12e41ffcb4cbec2013e335cf482377c35e", + "keyword": "831790fddda081bdddd14f96985ab02dc5b5341f", }), # upper case file extension (#296) ("https://post-phinf.pstatic.net/MjAxOTA1MjlfMTQ4/MDAxNTU5MTI2NjcyNTkw" @@ -46,11 +51,14 @@ class DirectlinkExtractor(Extractor): self.data = match.groupdict() def items(self): - text.nameext_from_url(self.url, self.data) - for key, value in self.data.items(): + data = self.data + for key, value in data.items(): if value: - self.data[key] = text.unquote(value) + data[key] = text.unquote(value) + data["path"], _, name = data["path"].rpartition("/") + data["filename"], _, ext = name.rpartition(".") + data["extension"] = ext.lower() yield Message.Version, 1 - yield Message.Directory, self.data - yield Message.Url, self.url, self.data + yield Message.Directory, data + yield Message.Url, self.url, data |