User talk:Egon Willighagen

From Wikidata
Jump to: navigation, search

User talk Archives[edit]

Bug in tutorial[edit]

I spend a while debugging why the tutorial in the example does not work. Seems like only the variable ?item is currently allowed in pywikibot. I will file a bug or maybe send a patch to the pywikibot team. I adapted the example here: Wikidata:Pywikibot_-_Python_3_Tutorial/Big_Data#Selecting_Item_by_Wikidata_Statement

Did you run into any mistakes in the tutorial? --Tobias1984 (talk) 19:53, 4 May 2016 (UTC)

Please check and correct the statements[edit]

You created items and imported data for two chemicals which seem to be only one (they share several identifiers) without any sitelink. Please can you check these 2 items (PIP2[4',5'] (Q22082911), no label (Q23927257)) and merge them if necessary or providing enough description to understand what is the difference between them ? Thank you Snipre (talk) 20:09, 25 May 2016 (UTC)

Indeed :( I'm sorry about that. I still need to check why I did not pick this up, as I do checking if something is present. I guess something went wrong with the check on the PubChem Compound ID. Thanks for the catch and letting me know! I merged to two entries. Egon Willighagen (talk) 04:10, 26 May 2016 (UTC)


exact match (P2888) is ready. --Tobias1984 (talk) 19:28, 5 June 2016 (UTC)

guanosine cyclic 3',5'-phosphate (Q422511)[edit]

Hi, I think the data your bot has added is not correct. There are now two chemical compound (Q11173) statements, one with found in taxon (P703) as a qualifier (which should not be used as a qualifier, but as a property). ∼Wostr (talk) 18:20, 27 August 2016 (UTC)

User:Wostr, point taken. How would you model the fact that something is a 'human metabolite'? Subclassing metabolite doesn't sound like a good idea, as there are too many species... BTW, the related bot permission request is Wikidata:Requests_for_permissions/Bot/UreomiczBot 1. --Egon Willighagen (talk) 18:26, 27 August 2016 (UTC)
I don't understand why the found in taxon (P703) must be a qualifier for instance of (P31) = chemical compound (Q11173), as it should be added as a property, not as a qualifier; like in Reelin (Q13561329). I have no idea how the fact that something is a 'human metabolite' should be modelled, but neither adding found in taxon (P703) as a qualifier for instance of (P31) = chemical compound (Q11173) nor duplicating the 'instance of (P31) = chemical compound (Q11173)' statement are correct. ∼Wostr (talk) 18:38, 27 August 2016 (UTC)
That may work... most metabolites (which is what 95% is of what WikiPathways chemical compounds are) are not species specific. But it means that found in taxon (P703)must support 1-n relations between chemical compound and species... (and you will likely get very many 'found in taxon' values. Would that be acceptale? --Egon Willighagen (talk) 18:43, 27 August 2016 (UTC)
Using found in taxon (P703) as a property and not as a qualifier for instance of (P31) = chemical compound (Q11173) will not affect the relation between chemical compound and the species and I really don't know how exactly it could affect any such relation. The relations compound–spiecies remain the same, only the way the relations are added changes, so the number of 'found in taxon' values will be identical in both options. ∼Wostr (talk) 19:28, 27 August 2016 (UTC)


Op 29 oktober vieren we de vierde verjaardag van dit prachtige project. Het lijkt mij heel leuk om af te spreken met diverse Nederlandstalige vrijwilligers in Utrecht, zie Wikidata:Fourth Birthday voor meer informatie. Sjoerd de Bruin (talk) 18:49, 3 October 2016 (UTC)

@Sjoerddebruin: oh, dan zit ik in de trein naar huis van een reis voor het werk :( Egon Willighagen (talk) 18:55, 3 October 2016 (UTC)
Jammer! Je kunt natuurlijk nog een week later naar Berlijn komen maar dat is ook even reizen. Bedankt voor je snelle reactie! Sjoerd de Bruin (talk) 19:09, 3 October 2016 (UTC)


I saw this and this today... and don't understand. The property found in taxon, based on its discussion page, appears to have been created to hold gene/protein data - like gene X is endogenous to species Y.

You appear to be filling it with chemicals/metabolites that can be found in humans sometimes....  ? Jytdog (talk) 19:15, 13 December 2016 (UTC)

@Jytdog: The description of the properties says "taxon in which the molecule or substance can be found"... it seems to me the domain information is incomplete, indeed. See also the discussion on that page about the use for molecules from July and October. --Egon Willighagen (talk) 20:23, 13 December 2016 (UTC)
I understand that you want a place to put metabolites but you are pushing the boundaries of what the field means, per the way it is already used (genes and proteins endogenous to X) and the "equivalent class" here. If you proceed you are going to render the field useless for making queries as it will have different kinds of data in it. Why in the world would you want to do that? Someone else raised a similar concern above. And really - should every chemical found in humans be listed there, whether it be some intermediate in a biosynthesis pathway or something like Bisphenol-A that many, many people have inside them? This seems crazy. Jytdog (talk) 20:50, 13 December 2016 (UTC)
You work at a (unnamed) company in drug discovery and would probably know very well that this is not actually crazy at all (and why it is needed). Besides, you claim something to be clear which it is not, and did not respond to that: the discussion page of found in taxon (P703) most certainly has enough to suggest this is appropriate use. I suggest we continue the discussion there to get the use clarified. --Egon Willighagen (talk) 21:18, 13 December 2016 (UTC)
Where i work has zero relevance. OK fine will discuss there. Jytdog (talk) 21:42, 13 December 2016 (UTC)