Udata

- Improve search serialization perfs for datasets in big topics [2937](https://github.com/opendatateam/udata/pull/2937)
- Migrate to Flask-Babel because of Flask-BabelEx deprecation [2897](https://github.com/opendatateam/udata/pull/2897)
- Contact points feature [2914](https://github.com/opendatateam/udata/pull/2914) [#2943](https://github.com/opendatateam/udata/pull/2943):
- Users and Organizations can now define a list of contact points
- Api endpoint for creating, updating and deleting contact points
- Datasets can define one contact point, among the list of the organization or the user owning the dataset.
- Defining a contact point for a dataset is done throught a form field
- Allow wildcards in redirect_uris for Oauth2Client [2935](https://github.com/opendatateam/udata/pull/2935)
- Allow for being one day late on update fulfilled in time [2941](https://github.com/opendatateam/udata/pull/2941)
- When a topic is deleted, corresponding discussions are purged [2944](https://github.com/opendatateam/udata/pull/2944)

- Add sorting in geozone suggest API endpoint to return zones based on their admin level [2936](https://github.com/opendatateam/udata/pull/2936)

- Prevent sending post_save signals on extras update [2919](https://github.com/opendatateam/udata/pull/2919)
- Add topic filter on datasets list [2915](https://github.com/opendatateam/udata/pull/2915)
- Topics: API v2 endpoints [2913](https://github.com/opendatateam/udata/pull/2913)
- Allow for discussions on Topics [2922](https://github.com/opendatateam/udata/pull/2922)
- Raise for status on DCAT harvester calls [2927](https://github.com/opendatateam/udata/pull/2927)
- Harvest dcterms:hasPart as 'other' resource in DCAT [2928](https://github.com/opendatateam/udata/pull/2928)
- Make sure harvested resources are marked as remote [2931](https://github.com/opendatateam/udata/pull/2931)
- Use GET and POST harvest BaseBackend utility to have user-agent [2930](https://github.com/opendatateam/udata/pull/2930)
- Use LazyReferenceField on Topic datasets and reuses [2924](https://github.com/opendatateam/udata/pull/2924)
- Use harvested dates and not max with internal [2932](https://github.com/opendatateam/udata/pull/2932)
- Better rdf frequency resilience [2933](https://github.com/opendatateam/udata/pull/2933)
- :warning: **breaking change** Geozone refactor [2878](https://github.com/opendatateam/udata/pull/2878):
- Complete Geozone model refactor, keeping only fields `slug`, `name`, `code`, `level` and adding `uri`
- Removed parent and validity concept
- To deploy:
- Datasets with geozone will return a 500 due to `mongoengine.errors.FieldDoesNotExist: The fields "{'flag', 'dbpedia', ...}" do not exist on the document "GeoZone"`
- `udata spatial load -d` to load new geozones into the database
- `udata spatial migrate` to migrate datasets geozones to new ones
- Reindex datasets (`udata search index dataset`) if using [udata-search-service](https://github.com/opendatateam/udata-search-service)
- Removed forgotten fields in search [2934](https://github.com/opendatateam/udata/pull/2934)

New & Improved

- Topics have been refactored and are no more deprecated:
- Topics creation, update and deletion are now opened to all users [2898](https://github.com/opendatateam/udata/pull/2898)
- Topics are now `db.Owned` and searchable by `id` in dataset search [2901](https://github.com/opendatateam/udata/pull/2901) [#2917](https://github.com/opendatateam/udata/pull/2917)
- Remove `deleted` api field that does not exist [2903](https://github.com/opendatateam/udata/pull/2903)
- Add `created_at`field to topic's model [2904](https://github.com/opendatateam/udata/pull/2904)
- Topics can now be filtered by `tag` field [2904](https://github.com/opendatateam/udata/pull/2904)
- Topics can now be queried by test search in `name` field with `q` argument [2904](https://github.com/opendatateam/udata/pull/2904)
- Add support for a CSW harvester using DCAT format [2800](https://github.com/opendatateam/udata/pull/2800)
- Add German to udata translations [2899](https://github.com/opendatateam/udata/pull/2899)[2909](https://github.com/opendatateam/udata/pull/2909)
- Add harvesters count to site metrics [2890](https://github.com/opendatateam/udata/pull/2890)
- Use a single session for reindex [2891](https://github.com/opendatateam/udata/pull/2891)

- Fix site title and keywords never get updated [2900](https://github.com/opendatateam/udata/pull/2900)
- Reuse's extras are now exposed by API [2905](https://github.com/opendatateam/udata/pull/2905)
- Add configuration settings to enhance cookies security [2910](https://github.com/opendatateam/udata/pull/2910)
- Add items.dataset to HarvestJob indexes [2907](https://github.com/opendatateam/udata/pull/2907)
- Consider acronym when suggesting organization [2918](https://github.com/opendatateam/udata/pull/2918)

Fixes

- Return 400 instead of 500 in case of not ObjectID arg in API [2889](https://github.com/opendatateam/udata/pull/2889)
- Fix default community resource sort parser [2908](https://github.com/opendatateam/udata/pull/2908)

Deprecation

- Python 3.7 is now deprecated and will be removed in upcoming release [2859](https://github.com/opendatateam/udata/pull/2859)
- GeoZone model will be heavily refactored for a simplified version [2878](https://github.com/opendatateam/udata/pull/2878)

- Fix slug overflow with index suffix when reaching max_length [2874](https://github.com/opendatateam/udata/pull/2874)
- Add extra field to topic model and add it to the dataset search adapter [2876](https://github.com/opendatateam/udata/pull/2876)
- Upgrade pyyaml in develop and doc deps [2880](https://github.com/opendatateam/udata/pull/2880)
- Expose dataset's `*_internal` dates in a nested `internal` nested field in api marshalling [2862](https://github.com/opendatateam/udata/pull/2862)
- Add `business_number_id` metadata for organizations [2871](https://github.com/opendatateam/udata/pull/2871) [#2887](https://github.com/opendatateam/udata/pull/2887)
- Return 403 when posting comment on discussion closed [2881](https://github.com/opendatateam/udata/pull/2881)
- Ensure rdf parsed frequency is lowercase [2883](https://github.com/opendatateam/udata/pull/2883)
- Add a dict of URIs to replace in a RDF graph at harvest time [2884](https://github.com/opendatateam/udata/pull/2884)
- Fix duplicate recipients in new comments mail [2886](https://github.com/opendatateam/udata/pull/2886)
- Add type to resource csv adapter [2888](https://github.com/opendatateam/udata/pull/2888)

- Improve DCAT harvest of mime type [2857](https://github.com/opendatateam/udata/pull/2857)
- Don't crash on files not found when purging resources [2858](https://github.com/opendatateam/udata/pull/2858)
- Improve DCAT catalog exposed [2860](https://github.com/opendatateam/udata/pull/2860)
- Use the resource's extra `analysis:last-modified-at` in the `last_modified` property [2863](https://github.com/opendatateam/udata/pull/2863)
- Add optionnal harvest validation form [2864](https://github.com/opendatateam/udata/pull/2864)
- Fix dataset list default sorting [2867](https://github.com/opendatateam/udata/pull/2867)
- Update API doc link [2866](https://github.com/opendatateam/udata/pull/2866)
- Update admin quality progress bar [2872](https://github.com/opendatateam/udata/pull/2872)