annotation overview data

[nconrad]

I'm reporting this because I didn't understand the data organization.

From my code:

// The annotation overview structure seems to consist of hashes with
// values of "1", instead of flat arrays.  This should probably be fixed.
// Note: the 'role' structure is correct
function parseOverview(data) {
    for (var i=0; i<data.length; i++) {
        data[i].pathways = Object.keys(data[i].pathways);
        data[i].classes = Object.keys(data[i].classes);
        data[i].features = Object.keys(data[i].features);
        data[i].subsystems = Object.keys(data[i].subsystems);
        data[i].reactions = Object.keys(data[i].reactions);
    }

    return data;
}

[samseaver]

Why is this an issue again? Does it cost more than a few milliseconds of browser time to retrieve the keys?

[nconrad]

Because the data is overly complicated and larger than needed. No, it doesn't cost more than a few milliseconds. The user downloads an extra 5kb (139kb uncompressed) when visiting the annotation page, though, and I ship around 10 extra lines of code.

[nconrad]

nc$ ls -lh annotation_overview*
-rw-r----- 1 nc staff 343K Aug 26 12:28 annotation_overview
-rw-r--r-- 1 nc staff 204K Aug 26 12:30 annotation_overview_parsed
-rw-r--r-- 1 nc staff 217K Aug 26 13:04 annotation_overview_spaces_removed
nc$ ls -lh gzipped/*
-rw-r----- 1 nc staff 34K Aug 26 13:07 gzipped/annotation_overview.gz
-rw-r--r-- 1 nc staff 29K Aug 26 13:07 gzipped/annotation_overview_parsed.gz
-rw-r--r-- 1 nc staff 29K Aug 26 13:07 gzipped/annotation_overview_spaces_removed.gz