I think that there may be an opportunity to conceptually simplify the code by
a) changing map types to plain old objects
b) not extending the map type
c) just parsing the object in the constructor rather than overriding Map.add to this end
d) (optional) removing the re-stringifying code, it adds extra overhead and complexity to the parsing
My ideas above come from trying to investigate a little bit on the project with the single file hub branch, I found it difficult to use the Map. I think at the very least, the SingleFileHub will be overhauled a little bit (it makes a TrackDbFile[], which is a bit odd, it should be just one TrackDbFile? or just an array of Track objects instead of using the notion of a TrackDbFile?). Additionally it seems we have pick the objects out one by one with .get, which is a little cumbersome compared to using an object.
We are not yet at a v1.0.0 semver release, so there is an opportunity there
I think that there may be an opportunity to conceptually simplify the code by a) changing map types to plain old objects b) not extending the map type c) just parsing the object in the constructor rather than overriding Map.add to this end d) (optional) removing the re-stringifying code, it adds extra overhead and complexity to the parsing
My ideas above come from trying to investigate a little bit on the project with the single file hub branch, I found it difficult to use the Map. I think at the very least, the SingleFileHub will be overhauled a little bit (it makes a TrackDbFile[], which is a bit odd, it should be just one TrackDbFile? or just an array of Track objects instead of using the notion of a TrackDbFile?). Additionally it seems we have pick the objects out one by one with .get, which is a little cumbersome compared to using an object.
We are not yet at a v1.0.0 semver release, so there is an opportunity there