Marco Fossati

The action has to be done manually via the UI, no API implementation/documentation seems available. Current UI procedure: job > settings > show advanced settings > contributors tab > geography/language...

To help workers in the crowdsourcing job.

To simplify the crowdsourcing job.

UI: job > settings > show advanced settings > quality control tab > quality control settings > minimum time per page field which is sets to `10` VS API: `"job[options][calibrated_unit_time]":...

Add sources we processed to https://tools.wmflabs.org/mix-n-match/

- [ ] filter rotten URLs - [ ] percent-decode IDs - [ ] double-check - [x] Discogs musician - [x] Discogs band - ~[ ] generate Discogs musical work~...

- [x] blocked by #19 - [ ] run `sync ids` - [ ] upload test edits - [ ] submit a request for permission at https://www.wikidata.org/wiki/Wikidata:Requests_for_permissions/Bot **Note:** we reverted...

- [ ] Make a class for the cache so it's also easier to switch to a new backend if needed. See https://github.com/Wikidata/soweego/pull/419#discussion_r691277529 - [ ] don't call public validation...

[This](https://w.wiki/46iF) SPARQL query returns all the regexes to validate a given ID. - [ ] Make a template out of the query, the first pattern should be a variable -...

Following a conversation with the MusicBrainz team, a _"relationship"_ set as _ended_ in the MusicBrainz database dump stands for a known broken URL: implement this check at import time.