[02:53:09] <wm-bb>	 <cvictorovich> Some of them aren't merely spams; they're trolling (re @Jan_ainali: We had one a couple of days ago with about ten messages)
[05:56:02] <wm-bb>	 <cvictorovich> https://integration.wikimedia.org/ci/job/quibble-vendor-mysql-php74-selenium-docker/80983/console
[05:58:35] <wm-bb>	 <cvictorovich> How can this "script timeout" be solved? I'm currently working on FlaggedRevs, and seems it's caused by confliction with VisualEditor (very likely to be related to some backend deps (node_modules?). This is the only remaining test that cannot pass.
[08:53:59] <wm-bb>	 <dpriskorn> I have a question about a database I found for Entity Linking.
[08:54:00] <wm-bb>	 <dpriskorn> There are 5 mio rows with QIDs and I need to fetch the following from Wikidata for each one:
[08:54:02] <wm-bb>	 <dpriskorn> sv label
[08:54:03] <wm-bb>	 <dpriskorn> sv description
[08:54:05] <wm-bb>	 <dpriskorn> sv aliases
[08:54:06] <wm-bb>	 <dpriskorn> How do I best do that using Python? 
[08:54:08] <wm-bb>	 <dpriskorn> Using the dumps in PAWS?
[08:54:09] <wm-bb>	 <dpriskorn> Which Wikidata dump reading library do you recommend?
[08:54:11] <wm-bb>	 <dpriskorn> https://github.com/egerber/spaCy-entity-linker/tree/master
[10:38:54] <wm-bb>	 <Nikki> the dumps are just json
[10:48:27] <wm-bb>	 <Nikki> I would probably use quarry for something like that though, the dumps are *huge*
[10:49:33] <wm-bb>	 <Nikki> (i.e. I'd generate a list of all swedish labels, etc, and then use that subset to extract the data I actually want)
[10:57:13] <wm-bb>	 <Nikki> if I were doing it from a dump anyway, I'd probably use `grep` (no point wasting time parsing an entity that doesn't have any swedish), `jq` (to extract the data) and `join` (to take the data I'd extracted and the list of ids I'm interested in and only return the intersection of them), but that's not python