Udata

- Handle None values in dataset and resource extras endpoints [2805](https://github.com/opendatateam/udata/pull/2805)
- Fix default license being selected in form in optional select group [2809](https://github.com/opendatateam/udata/pull/2809)
- Fix only SHA1 checksum is accepted when uploading resources [2808](https://github.com/opendatateam/udata/pull/2808)
- Fix organization metrics count [2811](https://github.com/opendatateam/udata/pull/2811)
- Fix setuptools version used in CI [2813](https://github.com/opendatateam/udata/pull/2813)
- Add `udata harvest clean` command [2812](https://github.com/opendatateam/udata/pull/2812)

- Add python version requirement <3.10 [2798](https://github.com/opendatateam/udata/pull/2798)
- Fix date timezone and format for harvest previz [2799](https://github.com/opendatateam/udata/pull/2799)
- Add support for DCAT startDate and endDate in temporal coverage [2801](https://github.com/opendatateam/udata/pull/2801)
- New feature: Users can now change their email by themselves [2792](https://github.com/opendatateam/udata/pull/2792)

- :warning: Kafka removal [2783](https://github.com/opendatateam/udata/pull/2783)[#2794](https://github.com/opendatateam/udata/pull/2794):
- Resources events publication now uses webhooks
- Search indexation (at runtime and with index command) are now made through HTTP requests.
- If you use [udata-search-service](https://pypi.org/project/udata-search-service/), you need to upgrade to the >=2.0.0 version
- Add dedicated extras endpoints on resources and datasets [2779](https://github.com/opendatateam/udata/pull/2779)
- Enrich catalog with harvest infos [2789](https://github.com/opendatateam/udata/pull/2789)
- Add optionnal license select group custom setting for admin [2786](https://github.com/opendatateam/udata/pull/2786)
- Make index setup optional on init based on config [2797](https://github.com/opendatateam/udata/pull/2797)

- :warning: Upgrade to `Flask-Security-Too` version 4.0.0 [2772](https://github.com/opendatateam/udata/pull/2772):
- New User model attribute `fs_uniquifier`, migration needed.
- The `fs_uniquifier` is used to invalidate existing session in case of password reset.
- The user's `fs_uniquifier` is used instead of the `id` for auth mecanism including permissions.
- Exhaustive list of changes [here](https://flask-security-too.readthedocs.io/en/stable/changelog.html#version-4-0-0).
- Fix apiv2 swagger with harvest metadata and add apiv2 swagger tests [2782](https://github.com/opendatateam/udata/pull/2782)
- Improve frequency criterion in quality score [2771](https://github.com/opendatateam/udata/pull/2771)
- Add quality score to csv catalogs [2785](https://github.com/opendatateam/udata/pull/2785)
- Optimize DCAT harvesting on large multiple-paged catalogs, introduce `HARVEST_MAX_ITEMS` development setting [2781](https://github.com/opendatateam/udata/pull/2781)
- Add condition in security mail utils class to avoid mail sending according to config var [2788](https://github.com/opendatateam/udata/pull/2788)

- Fix resource harvest uri validation error [2780](https://github.com/opendatateam/udata/pull/2780)

- :warning: **Breaking change** Use dedicated dynamic harvest metadata for dataset and resources. A migration copies identifying fields from extras to this dedicated metadata field. Extras won't be used anymore for harvest-related information. udata-ckan, udata-ods and udata-front packages are impacted and should be upgraded accordingly [2762](https://github.com/opendatateam/udata/pull/2762)