Hi! I'm trying a very little import with OpenRefine (fewer than 4 lines), but after starting it gets stuck at "0% complete". Also @Alexmar983: is having problem in starting edits. Are you noticing anything strange in these days? Thank you very much in advance!
OK, in my case it has just done the edits. Anyway, if you have noticed some slowness or strangeness, let us know! Thank you anyway!
yes because in my case I cannot start the import at all. It freezes when I log in my password on Mozilla Firefox and on the Microsoft browser... all becomes kinda grey-ish and I have to reboot OpenRefine totally to do anything at all (can't even come back). That's why I asked @Epìdosis to do that little test. this problem is going on since many days and it never disappeared so I am quite worried.
BTW, since few days I can import again. maybe it was the simple reboot of the PC, no idea. I will tell you if the issue appears again.
when trying to revert one of my batches I came to this page . It says:
The current lag is 1 day, 13:50:39 (136239 seconds) - if this is more than an hour then something is probably wrong - please notify the maintainers of the tool and try again later.
Edit group not found
Hi, I tried to run two batches in QuickStatements using curl, say:
-d action=import \
-d submit=1 \
-d username=Bargioni \
-d "batchname=add en labels to journals from issn.org" \
--data-raw 'token=***' \
Job #28473 and #28478 were executed, but they remain in a state of RUN, not DONE. No modifications were applied to items.
More, both jobs lack the edit group: Edit group "28473" not found. https://tools.wmflabs.org/editgroups/b/QSv2/28473/
Thx a lot.
Hi, this sounds like a problem between you and QuickStatements. EditGroups will not detect the edit group until at least one edit has been performed successfully on Wikidata.
Ehm, take a look at my figures... https://xtools.wmflabs.org/sc/www.wikidata.org/Bargioni :-)
Ok, it seems you refer to the single batch. If so, I have a problem with QS when accessed through curl. Sorry for the previous reply.
Property:P5318 - Property:P5319
J'ai remarqué que le site de l'Académie des César, avait eu la "brillante idée" de changer toutes les URL concernant les films et les acteurs/actrices.
Exemple avec le film De rouille et d'os :
Avant l'URL était la suivante : https://www.academie-cinema.org/ceremonie/palmares-par-recherche/,film,43.html
Maintenant c'est celle-ci : https://www.academie-cinema.org/films/de-rouille-et-dos/
Idem pour les personnalités du cinéma, exemple avec Marion Cotillard :
Si vous avez une idée comment résoudre le problème, je vous en serais reconnaissant, car moi, mes connaissances sur Wikidata sont très très limitées.
Je notifie également @Thierry Caro qui a plusieurs fois modifié ces deux éléments.
J'ai traité les films, pour commencer.
Et les gens ensuite. C'est bon maintenant.
Un grand merci Thierry Caro ! Et désolé du dérangement.
Bien cordialement. Jack
OpenRefine for de-duplication
Hi! I'd like to use OpenRefine to help de-duplicate the recent import from The Peerage website.
In particular, at Property_talk:P4638#English_Wikipedia_articles_without_ID there are some lists of items that one might strongly suspect should have The Peerage person ID (P4638) IDs, but currently don't.
Is there any neat workflow for loading one of these batches into OpenRefine and reconciling it, ignoring the known Qid, (or, second-best, looking only for Qids above Q75000000) ?
Intuitively this sounds like a use case for third-party reconciliation.
I don't know if you have access to a full dump of The Peerage (with names and identifiers) - if you do, then you could potentially try to set up a reconciliation service for it, for instance using reconcile-csv. If you can only access their website, then you could perhaps write up a small reconciliation service which wraps their own search engine in the reconciliation API, but that is a bit more work.
Once you have such a service, you can just use it in place of the lobid-GND service demonstrated in the tutorial.
Let me know if it works for you!
It's a bunch of webpages, raw html, lots of entries on each page, not much of a searh facility. The best most accessible version of names + IDs is now Wikidata, where there was a 100% upload (700,000 IDs). Of course those could be extracted reasonably easily by a SPARQL query, with their WD labels; or alternatively the catalog data could be downloaded from Mix'n'match, if the original labels are preferred -- so not hard to get a local copy of the data.
But what I was really wondering was whether there was a technique to match eg a few hundred names against Wikidata, but either excluding a specific Qid for each match (or filtering it from the suggestions post-match); or alternatively, excluding all Qids with Q-numbers lower than 75,000,000
Yeah I think I see the idea - I don't really have a workflow to propose on top of my head, reconciliation isn't really designed to do deduplication. It's something we could improve (it's not clear to me what it would look like but definitely worth thinking about).
Still, I think you could potentially use third-party reconciliation for that: download the 700,000 qids, ids and labels from wikidata, put that in reconcile-csv (if it can handle that load - I haven't tried), and then try to reconcile to this your lists of items which should have a The Peerage id.
thanks for creating Beta Masaheft property!
Thank you very much for creating the property on Wikidata for us!
OpenRefine reconciliation down?
Hi, I was just wondering whether reconciliation is down (for everyone) at the moment, or whether it's just me?
It was working for me on Monday (13th), but today I'm getting a "502 Bad Gateway" error. (At 3pm this afternoon, and again when trying again now at 10pm).
Just wondered whether this is likely to be a server-side issue; or whether some problem has come up my end (eg firewall issue?)
Interesting! It is working now as far as I can tell, but in general you are right that it is really not as stable as it should.
I don't know what I can be doing wrong then... I've tried turning Windows firewall off, tried versions 3.2 beta, 3.2 stable, 3.3 rc1, tried reconciling a different file -- still seeing the same thing.
Curiously when I added a column with a single Q-number, and asked it to reconcile that, that worked. But try a column of actual strings, and I'm getting the 502 error
It seems I can successfully reconcile to VIAF, using the service by RDM Page, just not Wikidata.
I tried putting in the Wikidata API address as an additional reconciliation service and got the following onscreen warning, even though the link seems to give a response when just put into a web-browser address bar:
"Error contacting recon service: error : Not Found - Error contacting recon service: timeout : timeout - https://tools.wmflabs.org/openrefine-wikidata/en/api"
Now seems to be working again.
Oops, you were right indeed! I restarted the service this morning. I really need to find a more reliable hosting provider than toollabs.
Hi, I was just wondering how you do to add a lot of properties' data in a single edit. Is it some gadget or something else?
Hi, Yes I use a script, which is available here: https://github.com/wetneb/wikidata-bots/tree/master/createprop
Does the bot automatically recheck if the page is edited after a fail to validate regex?
No, my property creations are supervised. I am not checking Wikidata often as I am on Wikibreak :)
I see. Well, apologies for disturbing you from it. I just updated the proposal to have a more correct length of description.
Editgroups lag is now 14 hours
I would like to revert a recent batch. Can you help decrease the lag somehow?