search

gbv / cmo

Corpus Musicae Ottomanicae
GNU General Public License v3.0
6 stars 0 forks source link

Certain Makam names are not represented under standardized name #250

Open sgroneme opened 4 months ago

sgroneme commented 4 months ago

When I browse for makam "Sûz-ı dilârâ", I get 68 pieces displayed.

When I copy & paste "Sûz-ı dilârâ" into source catalogue search field, there are 73 pieces.

When I type "suz-i dil-ara", I get 511 pieces, which are classified under makam "Sûz-ı dilârâ".

Is there a problem with the indexing?

zhelvaci commented 4 months ago

I think the discrepancy is caused not by the indexing but by the divergent namings in the sources vs. the correct makam names (given/categorized by us). Compared with the filters on the right hand side, 68 of the pieces in the search with 73 results are listed under the standard name "Sûz-ı dilârâ", 4 under "Rûy-ı dilârâ" and 1 under "Nihâvend". It is similar with the serach with 511 results. But it is a good point to include in the user guide.

nihantahtaisleyen commented 4 months ago

In fact, through this example, we want the system to detect the simplest spelling of the maqam and list both the standard and all the different spelling versions of the same maqam. For example, when we type suzidilara without symbols, it would be good to show the available data. When I type suzidilara, I get 11 results. But we have more... No user writes names with symbols when searching for maqam and usul. They search in the simplest form. This applies to all maqams and usul... If we can find a solution, it would be very efficient for users and search options. One more example: devri kebir (simplest) 140 results, but devr-i kebir seen under 956 results...