Update BNF authority file links

ahankinson commented 2 months ago

The BNF authority links are a bit of a mess.

Some cite the full ARK, e.g., https://muscat.rism.info/admin/institutions/40006238. This doesn't work because the link formatter expects just a number. So it turns it in to https://data.bnf.fr/ark:/12148/cbark:/12148/cb14783478b.

Some records cite just the number, e.g., https://muscat.rism.info/admin/institutions/51005989 gives 15386579.

However, this also does not work, because it's missing the check digit (4) in this case, which is part of the ARK. In other words:

ark:/12148/cb153865794 -> cb is the prefix, 15386579 is the identifier, 4 is the check digit.

But wait! It gets better.

According to this page: https://www.bnf.fr/fr/lidentifiant-ark-archival-resource-key, the BNF have several schemes for their ARK prefixes: dp for digitized documents, cb for catalogue records, and mm for educational resources. I happen to also know that there are others, e.g., ark:/12148/btv1b105513309. The best way to untangle all of these is to go through their ARK resolver, https://ark.bnf.fr.

So, this is just a long-winded way of saying that:

I think the BNF identifier in Muscat should be configured to point to https://ark.bnf.fr/. This requires the use of the full ARK in our identifiers, but corresponds to the "Identifier" that the BNF themselves cite on their catalogue pages: https://catalogue.bnf.fr/ark:/12148/cb14783478b -> "Identifiant de la notice", and, if we ever require it, on their Gallica site: https://gallica.bnf.fr/ark:/12148/btv1b8422976g -> "Identifier".
the BNF identifiers in Muscat should have the full ARK in them, and not just the "identifier" bit (since it's really hard to figure out the check digit otherwise, and if we don't have the prefix we might be mixing up the different types of records.)

Tagging @Docudoctor, since I've already roped him in to this on Muscat.

xhero commented 2 months ago

I can surely change the hardcoded link from https://data.bnf.fr/ark:/12148/cb to to https://ark.bnf.fr/, but how should we go to fix all the identifiers?

xhero commented 2 months ago

I updated the link so it should work in the next release. @alexandermarxen and @Docudoctor could you have a look to see what is the path to fix the data here? We can probably isolate the identifiers that do not start with "ark:" but they will need to be checked by hand.

alexandermarxen commented 2 months ago

Thank you very much! I'm afraid that there are a lot of records, especially when it comes to persons. Could I have a list of the affected records?

ahankinson commented 2 months ago

I think these are the people records that have a bnf identifier, but that identifier doesn't start with "ark". ~2,100.

ark_people.csv

alexandermarxen commented 2 months ago

Thank you very much! It's about the order of magnitude I was expecting.

BaMikusi commented 2 months ago

(FYI: Guido will be able to look into this only next week.)

Docudoctor commented 2 months ago

Hi! There are round about 30 institution records with a BNF identifier. Some with ark, some without. I can easyly find them in "Any field". My question is, how can I cite the BNF identifiers?

ahankinson commented 2 months ago

You should include the full ARK, including the ark:/ prefix. So for example:

https://muscat.rism.info/admin/institutions/40006238/edit

This is cited correctly, even though the link does not work. The link will start working after the next Muscat release.

This record, however:

https://muscat.rism.info/admin/institutions/40001144

Is not correct. The BnF identifier should be: ark:/12148/cb12229245w

You can see this in the BnF catalogue: https://catalogue.bnf.fr/ark:/12148/cb12229245w Look under the "Identifiant de la notice" section.

Docudoctor commented 2 months ago

I know, but I was a little bit confused, because of the change of the hardcoded link. So good luck! :)

rism-digital / muscat

Update BNF authority file links #1564