I'm trying to upload the songs in the user's media library and get analysis information for them in a way that will scale across hundreds of users and make best use of the API calls I have available. Doing a catalog update seems like the right way to do this, but I have noticed that the song information (artist name, song name) has to be a precise match for the song to resolve. If one word is missing off the end of the song name, the song will not resolve correctly. On the other hand, the song search method will likely show the right song in this case (or may even show multiple matches).
My question is if there is a way to get the catalog update method to behave more like the song search method. This would drastically improve the number of matched tracks when you're uploading a real user's music collection, where the song metadata is bound to be imperfect. If catalog update picked the top result that song search would produce for the given artist/song name, that would be ideal.
I can't call song search on every song in the user library because that would quickly exhaust my API limit.