Zimscraperlib

Added

- `zim.creator.Creator(language=)` can be specified as `List[str]`. `["eng", "fra"]`, `["eng"]`, `"eng,fra"`, "eng" are all valid values.

Changed

- Fixed `zim.providers.URLProvider` returning incomplete streams under certain circumstances (from https://github.com/openzim/kolibri/issues/40)
- Fixed `zim.creator.Creator` not supporting multiple values in for Language metadata, as required by the spec

- Using pylibzim v2.1.0 (using libzim 8.1.0)

Added

- [libzim] `Entry.get_redirect_entry()`
- [libzim] `Item.get_indexdata()` to implement custom IndexData per entry (writer)
- [libzim] `Archive.media_count`

Changed
- [libzim] `Archive.article_count` updated to match scraperlib's version
- `Archive.article_counter` now deprecated. Now returns `Archive.article_count`
- `Archive.media_counter` now deprecated. Now returns `Archive.media_count`

Removed

- [libzim] `lzma` compression algorithm

Added

- `download.get_session()` to build a new requests Session

Changed

- `download.stream_file()` accepts a `session` param to use instead of creating one

Added

- `zim.Creator` now supports `ignore_duplicates: bool` parameter to
prevent duplicates from raising exceptions
- `zim.Creator.add_item`, `zim.Creator.add_redirect` and `zim.Creator.add_item_for`
now supports a `duplicate_ok: bool` parameter to prevent an exception
should this item/redirect be a duplicate

Added

- `download.stream_file()` supports passing `headers` (scrapers were already using it)

Changed

- Fixed `filesystem.get_content_mimetype()` crashing on non-guessable byte stream