Robert Buels

NAME

Bio::JBrowse::Store::NCList::ArrayRepr - compact array-based serialization of hashrefs

DESCRIPTION

    The ArrayRepr class is for operating on indexed representations of objects.

    For example, if we have a lot of objects with similar attributes, e.g.:

        [
            {start: 1, end: 2, strand: -1},
            {start: 5, end: 6, strand: 1},
            ...
        ]

    we can represent them more compactly (e.g., in JSON) something like this:

        class = ["start", "end", "strand"]

        [
            [1, 2, -1],
            [5, 6, 1],
            ...
        ]

    If we want to represent a few different kinds of objects in our big list,
    we can have multiple "class" arrays, and tag each object to identify
    which "class" array describes it.

    For example, if we have a lot of instances of a few types of objects,
    like this:

        [
            {start: 1, end: 2, strand: 1, id: 1},
            {start: 5, end: 6, strand: 1, id: 2},
            ...
            {start: 10, end: 20, chunk: 1},
            {start: 30, end: 40, chunk: 2},
            ...
        ]

    We could use the first array position to indicate the "class" for the
    object, like this:

        classes = [["start", "end", "strand", "id"], ["start", "end", "chunk"]]

        [
            [0, 1, 2, 1, 1],
            [0, 5, 6, 1, 2],
            ...
            [1, 10, 20, 1],
            [1, 30, 40, 1]
        ]

    Also, if we occasionally want to add an ad-hoc attribute, we could just
    stick an optional dictionary onto the end:

        classes = [["start", "end", "strand", "id"], ["start", "end", "chunk"]]

        [
            [0, 1, 2, 1, 1],
            [0, 5, 6, 1, 2, {foo: 1}]
        ]

    Given that individual objects are being represented by arrays, generic
    code needs some way to differentiate arrays that are meant to be objects
    from arrays that are actually meant to be arrays.
    So for each class, we include a dict with <attribute name>: true mappings
    for each attribute that is meant to be an array.

    Also, in cases where some attribute values are the same for all objects
    in a particular set, it may be convenient to define a prototype ("proto")
    with default values for all objects in the set

    In the end, we get something like this:

        classes = [
            { "attributes"  : [ "start", "end", "subfeatures" ],
              "proto"       : { "Chrom"       : "chr1"   },
              "isArrayAttr" : { "Subfeatures" : true     }
            }
        ]

    That's what this class facilitates.

AUTHOR

Robert Buels <rbuels@gmail.com>

COPYRIGHT AND LICENSE

This software is copyright (c) 2013 by Robert Buels.

This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.