How to reconcile Cantus Index using Catalogue Numbers?

[linneakirby]

The data we have for Cantus Index includes feasts and genres, which we can reconcile, but the rest of the data is not as straightforward. One field is labeled "CAO" and references these, which we need more guidance on. @annamorphism and @kyrieb-ekat , perhaps you can help?

[annamorphism]

So, the CAO field is basically a tiny concordance table, telling you which of these 12 sources has this particular chant text. (The Cantus Index concordances should also tell you whether it's in any of the other several hundred indexed sources....)
I suppose one option would be to treat the CAO data as its own little property, which could potentially have 4096 values (C, CG, CGB, C B,...). Not sure that's really helpful, since it's not really paralleled in any other database, but it's something, and it would let you find things that are in e.g. C and B but not G, or things that are just in R.
Maybe better would be to (in addition?) transform it to look like the rest of the concordances, so that e.g. https://cantusindex.org/id/001001 is found in R (RISM: CH-Zz Rh. 28) as well as e.g. CDN-Mrb MS Medieval 0234 . (It wouldn't tell you WHERE in R it is, but it would be connected...)

[dchiller]

I've never even thought to ask this before: are all CAO sources also indexed in a Cantus Network DB?

[annamorphism]

No, though I wish they were. Cantus DB has "H"; MMMO has "C" "D" and "F". Not sure about the others.

[annamorphism]

ah, and SEMM has S: https://musicahispanica.eu/source/20199

[kyrieb-ekat]

There are a few ways to tackle the catalogue numbers question; manuscripts have shelfmarks and catalog numbers, both (or either) of which might end up in Wikidata. A good example here is this SPARQL search for items in wikidata with either a BNF or Gallica ID; all of our BNF items in Cantus Database are in this list, for example, and can be pulled up in wikidata by their shelfmark, gallica, or BNF ID. (https://w.wiki/EXbN)

BNF, just as a starting point, lets you search their catalogues and then export your results, which you can then narrow down to what you want (Latin, music, etc), and have sent to you as a CSV (I recommend, if you do these, to then power query and delimit by semi colons, not commas). I've attached my initial test query csv, which from the thirty odd random samples I took from CDB's french sources page link up (there are 49 total BNF explicit manuscripts in CDB right now)(this csv can be used to confirm that what you're reconciling to is in fact the object you're looking for and not something identically/similarly named).

Several of the other municipal libraries do also have a presence on Wikidata; I'll drop more links and match procedures as I come across them here. I'll knock out french stuff first, because they have a lot of linked library projects.

wikidata_bnf&gallica.csv
https://cantusdatabase.org/sources/?segment=&general=&indexing=&country=France&provenance=&century=&prodMethod=&sourceCompleteness=1&sourceCompleteness=4&sourceCompleteness=2&sourceCompleteness=3

[annamorphism]

....except it's a little bit more complicated, because the index for F on MMMO seems to just be the mass chants, and it looks like C's is just not complete, so the CAO values still provide information about those that isn't found elsewhere. The only ones for which it's also information indexed in the DBs seem to be H, F, and S.