Jump to content

Template talk:Infobox drug/Archive 12

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 5←Archive 10Archive 11Archive 12Archive 13Archive 14Archive 15

There are currently links to the IUPHAR/BPS Guide to Pharmacology (formerly IUPHAR-DB) database on many drug infoboxes but this database has grown massively over the past couple of years and now has over 7500 ligands. Many relevant drug infoboxes do not have links to IUPHAR. As the curators of this resource we would like to help add in these new links. Please could you tell us how we can do a bulk upload of IUPHAR links to Wikipedia drug infoboxes?

We provide downloadable lists of ligands on our website which might contain information helpful to do a bulk update, e.g. PubChem CIDs, InChIkeys, etc. See http://www.guidetopharmacology.org/download.jsp for the current full ligand list.

IUPHARcurators (talk) 10:29, 19 May 2015 (UTC)

Probably through WP:BOTREQUEST. The bot will have to solve the drug identification (a page may not be named by one of the identifiers; or the identifiers present in a page may not align).
An other route could be to add them to wikidata right away, which has other bot operators. IIRC Snipre works on wikidata chamicals.
At the moment this en wiki has 6,000 Drugbox articles and 10,000 Chembox articles (=chemicals not specified as drugs by editing, but they could be. Of these 10k, today 115 have a ligand entered). -DePiep (talk) 19:10, 19 May 2015 (UTC)

@IUPHARcurators: If you use PubChem CID we can easily integrate your ID in Wikidata. Snipre (talk) 11:31, 22 May 2015 (UTC)

We do index our chemicals by PubChem CID. What exactly do we need to do to make this happen? Thanks, IUPHARcurators (talk) 14:23, 27 May 2015 (UTC)

Perseverance. -DePiep (talk) 22:24, 27 May 2015 (UTC)
;-) Snipre (talk) 12:23, 28 May 2015 (UTC)
@IUPHARcurators: We need to match Wikidata items with your IDs. For example Wikidata identifier (called Q number) for ethanol is Q153 and the item is [1] and your ID is QNXXX1 (just an example). We have to define the link between Q153 in Wikidata and QNXXX1 in your database. You can do that manually by looking for each entry in your database which is the corresponding item in WD or you can use the power of programming to do that using PubChem CID as matching criteria (if PubChem CID for Q153 is 702 and PubChem CID for QNXXX1 is 702 so Q153 is QNXXX1). So step by step this gives:
  • Provide a list of your IDs with the corresponding PubChem CID
  • Get a list of Wikidata Q numbers with the corresponding PubChem CID
  • Match your IDs with Wikidata Q numbers
  • Use a bot and add your ID in the Wikidata items using a specific property. Snipre (talk) 12:23, 28 May 2015 (UTC)
By the way you can use CAS number or other common identifiers for matching purpose. The best will be to do the matching using several IDs in order to be sure that you link correctly both entries for 2 different databases. Snipre (talk) 12:53, 28 May 2015 (UTC)
Whattttttt? We need, We have to, You can do that manually you write? Snipre
Sure. Listen, wikidata people. This IUPHARcurators editor is offering to add a 7500 associated id's to wikidata (if you do not get this then leave wikidata right now). Sure you can ask: "is that PubChem -- IUPHAR id 1:1" (as anyone with knowledge about data would ask), but hey, that's the only question you need to ask[DP1]. And to be clear, IUPHAR is the Authority control (rings a bell? If not, leave wikidata).
The boring thing is that wikidata people still do not have a bot that can connect good data. See you in three years.
Were wikidata usable only in this one thing, I'd use it right away in {{Infobox drug}} and {{Chembox}} (16k transc's). -DePiep (talk) 22:23, 10 June 2015 (UTC).
[DP1]: duh, even wikidata people did not ask that question. Whatever: the opening remark was: "We need to match Wikidata items with your IDs". "We need ...". Says it all. -DePiep (talk) 23:03, 10 June 2015 (UTC)
@DePiep: Please keep your comment for yourself or come to help to improve the thing. I try to organize the things for chemistry field but I feel lonely enough without having to support such comments. If you want to help you can curate quite a lot of discrepancies for the different identifiers like this one. Snipre (talk) 11:27, 11 June 2015 (UTC)
I'm helping you, maybe in disguise. I'm actually pointing to the reason why you are alone in this. Wikidata is a wiki no one can edit. -DePiep (talk) 11:21, 12 June 2015 (UTC)

Thank you for all the info. We can easily provide PubChem CID to IUPHAR Id 1:1 mappings (we are confident these are correct because we manually curate them), but I'm still unclear how exactly we obtain a list of Wikidata Q numbers and which bot we use to do the upload. With >7500 ligands we'd obviously want to do this programmatically. Sorry to keep asking this question, but this is all new to us because in the past we have just provided our ligand list with PubChem mappings (obtainable at the link given above) and someone else did the matching on our behalf. IUPHARcurators (talk) 09:33, 11 June 2015 (UTC)

Serious point 1: Why would you add that Q-info while the IUPHAR -- CID 1:1 data relation you presented is there already? (I repeat: wikidata should do that)
Serious point 2: Is wikidata secretly trying to become another Authority control? Why didn't wikidata people say so? -DePiep (talk) 22:46, 12 June 2015 (UTC)
I will see if I can finish this job myself. The solutions above will work, but IMHO using Wikidata is an unnecessary complication. It is more straight forward to produce a list of article that contain {{drug box}} template. Using Pywikibot, this can be done with:
  • python pagegenerators.py -transcludes:"Infobox drug" -ns:0 > drugbox.txt (5128 aritcles as of 12 April 2015 5852 as of 11 June 2015)
and then extract the stored identifiers and maps these identifiers to the IUPHAR ID. Many but not all of the articles already contain an IUPHAR ID. After the mapping is done, a second bot can add missing IUPHAR ID parameters. I have a Pywikibot script to do both tasks (see for example User:BogBot/Source code/Task 06), but for reasons not entirely clear, the first bot kept crashing and I had to keep manually restarting it after every few dozen articles. I will try again this weekend. Hopefully fresh eyes will help me fix the problem. I already have all the data I need to complete this task. Cheers. Boghog (talk) 11:34, 11 June 2015 (UTC)
@IUPHARcurators: You extract the data yourself by using a query service developed for WD (still under development but able to perform simple request for the moment). Look at this UI and use this code to generate a list of items with their Qnumber and the value for the PubChem CID:
PREFIX wd: <http://www.wikidata.org/entity/> PREFIX wdt: <http://www.wikidata.org/prop/direct/> PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#> PREFIX p: <http://www.wikidata.org/prop/> PREFIX v: <http://www.wikidata.org/prop/statement/> SELECT ?h ?CID WHERE { ?h wdt:P662 ?CID . }
Or this URL (press execute to generate the list). Then you can copy-paste your data and perform your matching with the tool you want to use. When you have a list you can contact a bot using that page by defining your request. Sorry to not offering you to do the job but I am updating my bot and I don't have plenty of time currently to do it. Snipre (talk) 12:11, 11 June 2015 (UTC)
@Snipre: Your solution is way too complicated and does not directly address what the IUPHAR curators have asked for, namely (1) a mapping of IUPHAR ID with English Wikipedia drug article names and (2) populate the IUPHAR_ID parameters in English Wikipedia drug articles. Finally I have offered to complete the job myself. Boghog (talk) 12:24, 11 June 2015 (UTC)
@Boghog: I agree with you but 1) WD has 15221 items with a PubChem CID so more than your 5128, this means more chance to have a better overlapping with WD database and its database, 2) this will store data even if the English article doesn't exist meaning the data will be available for the future articles, 3) using WD means English infobox drug articles but all other infoboxes of all WPs even Chinese, Thai or Russian. So more complex but larger use. Snipre (talk) 13:00, 11 June 2015 (UTC)
@Snipre:The purpose of this request is to make reciprocal links between IUPHAR database and the English language Wikipedia drug articles. Therefore foreign language Wikipedia articles are not relevant. Furthermore many of the CIDs are neither drugs nor ligands hence would not be found in the IUPHAR database and therefore are also not relevant. Some drugs/ligands may be found in {{chembox}}, but at the moment, the chembox does not support an IUPHAR parameter. It probably should. Boghog (talk) 18:16, 11 June 2015 (UTC)
{{Chembox}} has |IUPHAR_ligand= in the Identifiers. See Glycine. -DePiep (talk) 09:27, 12 June 2015 (UTC)
OK, so for completeness, I will also extract identifier data from the chemboxes. IUPHAR_ligand is not included in the main documentation but I now see it is documented in the collapsed {{Navbox Chembox}}. Boghog (talk) 09:57, 12 June 2015 (UTC)
There are ~105/10000 {{Chemboxes}} having this input (by {{Chembox IUPHAR ligand}}). You might consider skipping this for effort efficiency. -DePiep (talk) 10:02, 13 June 2015 (UTC)
Many thanks Boghog for offering to complete this task for us. Thank you Snipre for the tips on using Wikidata, I didn't easily find any documentation on this. Maybe it's something we could consider in future if it seems useful, but at the moment it does seem to over complicate things since Boghog has helpfully provided another way. IUPHARcurators (talk) 13:48, 16 June 2015 (UTC)
Update: I have extracted the necessary data from {{infobox drug}} and {{chembox}}. In articles with drug boxes, there are currently 443 articles where the IUPHAR_ligand parameter has been specified. Using CID and InChiKey as cross references, I have identified 1417 - 443 = 974 drugbox articles with missing IUPHAR_ligand parameter values. For articles with chemboxes, 111 articles currently have IUPHAR_ligand values and 585 - 111 = 474 have missing values. I will now add these missing values and provide a mapping of IUPHAR_ligand IDs to Wikipedia article names. Boghog (talk) 14:15, 16 June 2015 (UTC)
@IUPHARcurators: Update2: BogBot has completed the addition of missing IUPHAR_ligand parameters to articles containing {{drugbox}} and {{chembox}} templates. I will e-mail the mappings of IUPHAR IDs to Wikipedia article names. Cheers. Boghog (talk) 06:57, 17 June 2015 (UTC)
Thank you very much Boghog! Great to see so many new links. IUPHARcurators (talk) 13:33, 17 June 2015 (UTC)
@Boghog: Just checking some of the new links and noticed that a small number of ligands from the spreadsheet you sent earlier are missing links in their {{Infobox drug}}. Two examples: Orlistat [IUPHAR id 5277] and Canagliflozin [IUPHAR id 4582]. Many others seem to be linking OK. What could be the reason for this? Thanks - IUPHARcurators (talk) 16:23, 17 June 2015 (UTC)
Thanks for pointing out the missing links. I believe that a large majority are linking OK. The few that are not linking are due to unforeseen special cases with the drugbox templates. As I see these exceptions, I modify the script to account for these exceptions. Hopefully I will catch these in the next run with a debugged script. Boghog (talk) 19:45, 17 June 2015 (UTC)
Do I understand correctly that pre-existing IUPHAR data entries, in both infoboxes, are not checked or touched in any sense? IOW, only additions were made? (no consequences for me, more out of curiousity). -DePiep (talk) 09:03, 18 June 2015 (UTC)
That's a shame - is it something specific with those special cases? It does look like a small number but if I spot any more I'll let you know. IUPHARcurators (talk) 14:19, 18 June 2015 (UTC)

Boghog and IUPHARcurators, have you seen wikidata:Wikidata:Data donation yet? It lists a couple of tools that might be helpful, and it might be another place to find help if you still need it. WhatamIdoing (talk) 02:06, 12 June 2015 (UTC)

To paraphrase Jamie Zawinski, if you use Wikidata to solve a problem, now you have two problems. The KISS principle applies here. Other bots already exist to extract information from infoboxes and upload this data to Wikidata. No need to duplicate the effort of these existing bots. I have tracked down the problem in my script to mwparserfromhell crashing when encountering links and in article section headings (see for example diff). I am making progress in compiling the drugbox data by fixing section heading errors as I encounter them. I hope to have the entire task completed in a few days. Boghog (talk) 03:52, 12 June 2015 (UTC)
Sigh. My point is: why does it need a manual approach to connect id data for wikidata. -DePiep (talk) 20:55, 13 June 2015 (UTC)
re WhatamIdoing: --It says: "Wikidata is editable by everyone." Duh, even I can not. --It says: "values high quality content from excellent partners." ask IUPHARcurators. "Wikidata offers several ways to get data into Wikidata" then prove it, right here right now. DePiep (talk) 21:59, 13 June 2015 (UTC)
@Boghog No, no new imports of data from WPs to WD will occur because this was only a first step in order to get data. Currently WD is curating the data coming from several WPS with a high number of discrepancies. So continuing the import of data from WPs without corrections from their sides is a headache for WD. Snipre (talk) 08:29, 17 June 2015 (UTC)
@Snipre: If it is a headache for WD, it is even a bigger headache for WP. From what you have written above, it is very clear that phase II of WD is not ready to be used by WP. It is essential that WD and WP are synched on a regular basis. It is also essential that discrepancies are resolved, but sounds like a lot of manual work requiring input from editors that understand the data. As participation in Wikipedia as a whole it declining, it will be difficult to find editors that have the time to do this. Finally the tools for up- and down-loading data are very primitive. WD has a lot of potential, but a lot more work needs to be done before we can use WD productively. Boghog (talk) 09:05, 17 June 2015 (UTC)
It is a headache mainly because people are not working in team: they add their own data without any discussion with other persons, bots import data and often delete previous data by replacing with their data or no matching process is done to control the links between WP articles and WD items. I hope until the end of the year to be control the data of the chemical compounds by doing cross-referencing between databases. Snipre (talk) 09:51, 17 June 2015 (UTC)

Side issues

This section is about IUPHAR topics that are not directly related to the mass update.
  • About target link IUPHAR ligand. This is the lefthand wikilinked text in both infoboxes. At the moment they link like this:
[[International Union of Basic and Clinical Pharmacology|IUPHAR ligand]]
IMO, this is not good enough. Our Reader expects a link to the definition of the ligand, not to the institute. I propose to have it linked to simply [[IUPHAR ligand]] (that's IUPHAR ligand, I just created). Whenever and whereever that definition is in our wiki, that page can be edited redirect to it (for example: #REDIRECT [[International Union of Basic and Clinical Pharmacology#Ligand]]. The infoboxes then don't need an edit, that's Good too. -DePiep (talk) 09:37, 18 June 2015 (UTC)
Done for {{Chembox}}. -DePiep (talk) 09:57, 18 June 2015 (UTC)
Done for {{Infobox drug}}. Anyone is invited to provide some IUPHAR ligand definition to enwiki. -DePiep (talk) 11:05, 18 June 2015 (UTC)
I agree that the current link to IUPHAR ligand is not ideal. I suggest that the link be changed to [[Guide to Pharmacology|IUPHAR/BPS ligand]]. Guide to Pharmacology is the name of the database where the ligands are stored and is equivalent to the other Identifier descriptions currently linked in the drugbox such as PubChem. DrugBank, ChEMBL, etc. Boghog (talk) 10:27, 18 June 2015 (UTC)
For starters, you can change the Redirect page to that one. (That's content change I am not familiar with, so I'll leave it to others). Now if the LH labeltext (in both infoboxes) should change to "IUPHAR/BPS ligand" instead of current "IUPHAR ligand", I'll wait the outcome of this proposal. My first thought is: what could "BPS" possibly mean, and why is such a code in the template at all? Mustn't I, lay reader, first research the word ligand to learn? Shouldn't it be "IUPHAR/BPS ligand" (I propose this one formally). -DePiep (talk) 11:15, 18 June 2015 (UTC)
To keep things simple and short, I propose that the word "ligand" be removed. The identifier section of the infobox links to external databases and the identifier refers to the registration number of the compound in that external database. BPS stands for British Pharmacological Society. BPS is part of the official name of the database and therefore it is important to include. In summary, the link should be [[Guide to Pharmacology|IUPHAR/BPS]]. Boghog (talk) 11:28, 18 June 2015 (UTC)
I have added the proposed change to the IUPHAR link in the sandbox and a side-by-side comparison of the current vs. sandbox can be found here. Boghog (talk) 11:49, 18 June 2015 (UTC)
Looks good to me (btw, I was not asking BPS to be explained here, I'd like to have it clear at face value in the infobox, or removed. However, this latest proposal makes a nice solution: as a name not an adjective). -DePiep (talk) 13:49, 18 June 2015 (UTC)
Maybe that target page better be moved to IUPHAR/BPS Guide to Pharmacology, as they title the main page. Would prevent confusion between "a guide" and "the guide". -DePiep (talk) 13:57, 18 June 2015 (UTC)
Thanks for making these changes. We would ask that the link displays 'Guide to PHARMACOLOGY' please, if that is not too long (the capitals are intentional). 'IUPHAR/BPS Guide to PHARMACOLOGY' is the full name but it is probably too long and/or would look less neat. Agree that the word ligand could be removed. We also propose to change the name of the main page to IUPHAR/BPS Guide to PHARMACOLOGY and agree the target should be IUPHAR/BPS Guide to PHARMACOLOGY. IUPHARcurators (talk) 14:13, 18 June 2015 (UTC)
To this link text I object. As with other data here, we want to mention the identifier's authorityshow What It Is concisely, not require clicks & puzzling. However inviting in a first approach, there is no need (or space) trying to explain or define the identifier. Therefor the visible text best should be: "IUPHAR/BPS" (Boghog's propsal). The actual invisible target link target may be different, longer and changing/moving all right. (btw, link directly, unlabeled, to IUPHAR/BPS must be OK too). -DePiep (talk) 14:41, 18 June 2015 (UTC)
I don't see that "IUPHAR/BPS" is less puzzling to a lay reader than "Guide to PHARMACOLOGY". In favour of the latter is that this is a name used frequently on other sites that link to us including UniProtKB, ChEMBL, DrugBank etc, so would be more familiar to many readers. IUPHARcurators (talk) 15:01, 18 June 2015 (UTC)
We want a name there, not "click here" or "we write PHARMACOLOGY here so you should understand that it has something to do with pharmacology, while the capitalization remains a mistery as does why only this should be that Guide as if there exists no other guide", or "click here to read here what should actually be written here". If a name like "IUPHAR/BPS" is not good enough to sit there, then we better remove the whole data row. -DePiep (talk) 15:23, 18 June 2015 (UTC)
  • I've implemented Boghog's proposal i.e. IUPHAR/BPS, albeit to that page (now a redirect, able to follow any change of target article). If the subsequential outcome here is to change the visible name once more, we'll do that. -DePiep (talk) 20:01, 18 June 2015 (UTC)


  • About code changes to {{Chembox}}, now nicely having 580 IUPHAR ligand aritcles. I have prepared some code cleanup in the |IUPHAR_ligand= area. My question is whether you mind me updating the subtemplates, or better wait more days while you are working with this? The changes are:
1. IUPHAR ligand now takes indexed values, just like other main identifiers do: |IUPHAR_ligand=, |IUPHAR_ligand1=, ... |IUPHAR_ligand5= each with comment option: |IUPHAR_ligand1_Comment=. This is mainly a useful for manual editing, allowing to add & link multiple identifiable chemicals (parallel with CASNo, CASNo1, ... CASNo5 etcetera). See testcases2. Existing |IUPHAR_ligand= is unaffected, and will be able to have a |IUPHAR_ligand_Comment= outside of the link.
2. Internally, code was cleaned up, including removing old wrongs links and {{Chembox IUPAC ligands/format}} was added. This have no effect on the resulting infobox (IUPHAR should work & show as it does now).
So, can I edit {{Chembox}}, or better wait? -DePiep (talk) 13:41, 17 June 2015 (UTC)
Thanks for the update on indexed values. I am essentially done adding IUPHAR_ligand parameters for now, so please go ahead an update the template as you were planning. There were a few ambiguities as to which IUPHAR ID to add. These were mainly caused by multiple forms of a ligand being registered in the IUPHAR database (e.g., salts vs. parent, unlabeled vs. labeled, etc). Where there were multiple forms, I generally retained/added only the ID corresponding to the unlabeled parent compound. Boghog (talk) 19:38, 17 June 2015 (UTC)
 Done. See this Chembox talk for more details. Or ask me. If Boghog's "ambiguities as to which IUPHAR ID to add" can be retracked or published, someone might be able to add multiple ligands to one infobox. (I can not). Changed the target wikilink as mentioned below. -DePiep (talk) 09:57, 18 June 2015 (UTC)
Alternative IUPHAR_ligand entries are easy to find by search the IUPHAR database. For example a search for Valsartan gives two hits, valsartan (Ligand Id: 3937) and [3H]valsartan (Ligand Id: 593). I recommend that |IUPHAR_ligand=3937 and |IUPHAR_ligand_Comment=parent while |IUPHAR_ligand1=593 and |IUPHAR_ligand_Comment1= tritiated. Boghog (talk) 10:27, 18 June 2015 (UTC)

Add Metabolites to Drugbox

Currently {Drugbox} and {Chembox} both have |metabolism=. However, {Chembox} has |metabolites= while {Drugbox} does not (as far as I can see). I think this would be a useful parameter for {Drugbox}. Can this be implemented? Sizeofint (talk) 19:20, 18 June 2015 (UTC)

 Done without much ado, |metabolites= added. Now it's up to you to edit 6000 articles ;-). -DePiep (talk) 20:19, 18 June 2015 (UTC)
Time to check the order of this set of data rows. There is a talk at Chembox. -DePiep (talk) 21:15, 18 June 2015 (UTC)

Presentation changes: no data no header

Currently the infobox shows headers even when there is no data supplid. This is about sections Clinical data, Identifiers, and Chemical data. I have prepared the sandbox to do not show a header when there is no data below. See Template:Infobox drug/testcases (and pages /testcases1–10). At the moment, there is a checking feedback text showingclean.

Internally , I've renumbered the headers to be a ten-fold. And the {{longitem}} template is incorporated in |labelstyle=, centrally, for the infobox. -DePiep (talk) 13:41, 22 June 2015 (UTC)

 Done -DePiep (talk) 22:33, 25 June 2015 (UTC)

Template edits 6 July 2015

I've made the next edits:

  • Add parameter |SMILES= next to existing |smiles=, accepting & handling the same input. SMILES in capitals is the proper name, as we usually have for parameter (eg PubChem). Old |smiles= is kept available unchanged.
  • {{EMA-EPAR}} formats input |licence_EU=. I have moved it to {{Infobox drug/EMA-EPAR}}: a fragile url (that was broken, see next point), better keep under tight control of {{Infobox drug}}. Today none of the ~150 usages was outside of {{Infobox drug}}.
  • Fix url link for |licence_EU= (see for example Thalidomide)
-DePiep (talk) 20:14, 6 July 2015 (UTC)
 Done -DePiep (talk) 20:45, 6 July 2015 (UTC)

Add secondary identifiers?

At the moment, a lot of {{Drugbox}}'s have a second identifier added incode. But that parameter does not show. Example: DASB has

|PubChem=656408 [2]
|PubChem2=10446567 [3]

This is about identifiers: smiles2, pubchem2, InChI2, InChIKey2, ChemSpiderID2, DrugBank2, UNII2, KEGG2 (and maybe CAS_number2, see below).

The main question is: what to do with these #2 id's? We can either remove them, or allow & show them.

Related question: there exists |CAS_supplemental= (Example: HDMP-28). Can someone clarify whether this 'secondary' is a common word in drug-world, or does this just mean a 'second CAS RN' i.e. CAS_number2? -DePiep (talk) 10:59, 28 June 2015 (UTC)

FYI: Most articles listed in Category:Chemical articles with unknown parameter in Infobox drug (now 100 P) have double id's this way. -DePiep (talk) 14:19, 28 June 2015 (UTC)

Proposal

CAS_number2
ChEBI2
ChEMBL2
ChemSpiderID2
DrugBank2
IUPAC_name2
IUPHAR_ligand2
KEGG2
PubChem2
StdInChI2
StdInChIKey2
UNII2
SMILES2

I propose to add secondary identifiers as listed. It follows the IUPHAR/BPS ligand updates (discussed here), and should show the already added second parameter in some 45 articles, listed in Category:Chemical articles with unknown parameter in Infobox drug. See /testcases5#index-2 for a demo.

It might also support articles that have |type=combo set, but only for two components not four. (If I am mistaken here, please clarify the relation between combo components and identifiers).

The pattern follows {{Chembox}} indexing setup, but only for the #2 index. When added, they will be followed by a |CAS_number2_Ref= bot-maintained check parameter (see {{cascite}} documentation).

Remarks? -DePiep (talk) 22:55, 13 July 2015 (UTC)

IUPAC name is systematic name?

See Acetate. It says: 

IUPAC name: Acetate
Systematic IUPAC name: Ethanoate

So these two labels do not mean the same (as they can lead to different names).

Now Drugbox says "Systematic (IUPAC) name" only. Does not look OK. -DePiep (talk) 21:10, 16 July 2015 (UTC)

Need for InChI next to StdInChI?

In Category:Chemical articles with unknown parameter in Infobox drug, some 40 (out of current 48) articles have |InChI= in the {{Drugbox}}. But Drugbox only knows & shows |StdInChI= and |StdInChIKey=.

Are there good reasons to add InChI, InChIKey to {{Drugbox}}? If not, I'll delete that input. Remarks? -DePiep (talk) 10:11, 8 July 2015 (UTC)

 Done rm all InChI parameters when there is a StdInChI present (few dozens). (We could keep those because of the discussion below, but these InChI's were never shown or checked/sourced ever). -DePiep (talk) 21:50, 16 July 2015 (UTC)
@DePiep: (and I would argue the same for {{Chembox}}): For me, the current content of InChI/InChIKey could be ignored throughout, and StdInChI and StdInChI should be the standard in the boxes. (and after that, a bot removing all InChI and InChIKey parameters that are currently there and renaming all StdInChI and StdInChIKey's to InChI resp. InChIKey would be great - though probably in a bit smarter way than just blind delete/move: if InChI = std -> use that, delete StdInChI, otherwise if InChI is set and StdInChI is blank, keep the old one, if InChI is not std and StdInChI is -> overwrite InChI and delete StdInChI, leave rest to maintenance cat). --Dirk Beetstra T C 11:23, 15 July 2015 (UTC)
For the original question, this is a "delete unless there is no Std entered" (all by pairs wrt the Key).
To refine: 1. you want "Std" removed from shown text (in article). All InChI says "InChI", and up to the reader to see that it is Std (by "1S"). 2. What input value is shown can be performed by the template by internal logic (detect presence of Std, detect whether an InChI is Std, then apply the logic you mention wrt what is shown). 3. Renaming params (moving changing param names in the article, maybe delete a whole parameter) is a late-action (can be done afterwards, it is just a deprecation).
One Q: the Key, which does not show whether it is Std, is always paired to a shown InChI? Is Key without InChI valid input for our infoboxes (or to categorize for maintenance)? -DePiep (talk) 11:37, 15 July 2015 (UTC)
Better start a new thread with this (copied). No hurry for that. Is very // with {{Chembox}}. I understand the algorithm for my OP. -DePiep (talk) 01:42, 16 July 2015 (UTC)
  • I think there is no real need to have any non-standard InChI's (but am willing to hear other input) - an argument could be that people might want to be able to find a Wikipedia page through a search on an external search engine using a non-standard InChI, and hence that we therefore need to render the non-standard InChIs for each page (maybe invisible).
  • The key is generated from the InChI by an algorithm. InChIs are unique, but through the algorithm there is a certain chance that two different InChIs result in the same Key. People do use either as the search key in search engines, and their rendering in the final page should both be 'perfect' (so that they get properly indexed, and people would be able to find the Wikipedia page through an external search on either InChI or InChIKey).
I agree with the point that the rename should be done post-implementation (Lua might be a help in maintenance categorisation). --Dirk Beetstra T C 05:29, 16 July 2015 (UTC)
All clear. Only one InChI will show, pref the Std. Key follows. There can be some maint categories (like 'Key missing', 'only non-Std InChI', ...). I will ask around about a second InChI (a non-Std one) being present in the article wrt wiki search philosophy. Now all I need is a lot of time :-) -DePiep (talk) 18:43, 16 July 2015 (UTC)

Reopening the topic: CAS RN link?

I have reopened the (sleaping) topic of #the CAS RN link. Comments best go there. -DePiep (talk) 20:19, 6 October 2015 (UTC)

Retrieved from "https://en.wikipedia.org/w/index.php?title=Template_talk:Infobox_drug/Archive_12&oldid=687281836"