Zimscraperlib

* Using new python-libzim based on libzim v7
* New Creator API
* Removed all namespace references
* Renamed `url` mentions to `path`
* Removed all links rewriting
* Removed Article/CSS/Binary seggreation
* Kept zimwriterfs mode (except it doesn't rewrite for namespaces)
* New `html` module for HTML document manipulations
* New callback system on `add_item_for()` and `add_item()`
* New Archive API with easier search/suggestions and content access
* Changed download log level to DEBUG (was INFO)
* `filesystem.get_file_mimetype` now passes bytes to libmagic instead of filename due to release issue in libmagic
* safer `inputs.handle_user_provided_file` regarding input as str instead of Path
* `image.presets` and `video.presets` now all includes `ext` and `mimetype` properties
* Video convert log now DEBUG instead of INFO
* Fixed `image.save_image()` saving to disk even when using a bytes stream
* Fixed `image.transformation.resize_image()` when resizing a byte stream without a dst

Intermediate release using unreleased libzim to support development of libzim7.
Don't use it.

* requesting newer libzim version (not released ATM)
* New ZIM API for non-namespace libzim (v7)
* updated all requirements
* Fixed download test inconsistency
* fix_ogvjs mostly useless: only allows webm types
* exposing retry_adapter for refactoring
* Changed download log level to DEBUG (was INFO)
* guess more-defined mime from filename if magic says it's text
* get_file_mimetype now passes bytes to libmagic
* safer regarding input as str instead of Path
* fixed static item for empty content
* ext and mimetype properties for all presets
* Video convert log now DEBUG instead of INFO
* Added delete_fpath to add_item_for() and fixed StaticItem's auto remove
* Updated badges for new repo name

* add `stream_file()` to stream content from a URL into a file or a `BytesIO` object
* deprecated `save_file()`
* fixed `add_binary` when used without an fpath (69)
* deprecated `make_grayscale` option in image optimization
* Added support for in-memory optimization for PNG, JPEG, and WebP images
* allows enabling debug logs via ZIMSCRAPERLIB_DEBUG environ

* added `wait` option in `YoutubeDownloader` to allow parallelism while using context manager
* do not use extension for finding format in `ensure_matches()` in `image.optimization` module
* added `VideoWebmHigh` and `VideoMp4High` presets for high quality WebM and Mp4 convertion respectively
* updated presets `WebpHigh`, `JpegMedium`, `JpegLow` and `PngMedium` in `image.presets`
* `save_image` moved from `image` to `image.utils`
* added `convert_image` `optimize_image` `resize_image` functions to `image` module

* added `YoutubeDownloader` to `download` to download YT videos using a capped nb of threads

* fixed rewriting of links with empty target
* added support for image optimization using `zimscraperlib.image.optimization` for webp, gif, jpeg and png formats
* added `format_for()` in `zimscraperlib.image.probing` to get PIL image format from the suffix