introduce two new metadata fields:
- filename_media: original filename of file uploaded to 4chan
- timestamp_ms : timestamp with millisecond precision (tim)
- either {_lit[foo]} or {'foo'}
- useful as alternative for empty metadata fields: {title|'no title'}
- due to using '_string.formatter_field_name_split()' to parse format
strings, using certain characters will result in an error: [].:!
- rename
- load_cookiestxt -> cookiestxt_load
- save_cookiestxt -< cookiestxt_store
- in cookiestxt_load, add cookies directly to a cookie jar
instead of storing them in a list first
- other unnoticeable performance increases
Only call os.makedirs() when an open() call
fails with a FileNotFoundError
Might cause errors with some post processors etc,
but for now I was unable to find any.
- add 'date' metadata to avatar/background files when available
and use that in default filenames / archive ids
- remove deprecation warnings as their option names clash with
subcategory names
- use media timeline + search for default user URLs like
https://twitter.com/SCREEN_NAME
- fetches all/most media for the type of twitter URL that most users
use with gallery-dl
- can be disabled by setting 'strategy' to any truthy value,
like "timeline"
- split 'user' extractor and its 'avatar' and 'background' options into
separate extractors ('artworks', 'avatar', 'background')
- avatars can now be downloaded with
https://www.pixiv.net/en/users/ID/avatar
as URL and will use a proper archive key; similar for backgrounds
- options for the 'user' subcategory must be moved to 'artworks' to have
the same effect as before
- start all endpoints with '/'
- use extractor.wait() for rate limit
- retry with while loop instead of recursion
- in case of error, write entire response to debug log