Records flat file storage

[ArnaudBuchholz]

• Handle array of records (objects)
• Want to know if some changed (use of prototype & hasOwnProperty)
• IO in a flat file
  • CSV / JSON?
  • Need to abstract host & streams (IFileStorage, IStream)
  • Append mode for performances, save only differences (i.e. read must adapt consequently)
  • Propose a compact feature (regenerate the file)
• Propose filtering helpers (1 syntax but several parsers: LDAP, ODATA, ...)
• Propose sorting helpers (might include merging several arrays)
• Propose paging helpers? (not sure if useful in step 1 as I might need to get a long array of results anyway, slice should be enough)

For future:

• By default, every properties will be saved (considering the record is flatten'd). Using attributes, this can be restricted or controlled more finely
• Records could be validated through JSON schema

[ArnaudBuchholz]

gpf:createObjectStorage({
    id: "property" || [properties]
    output: IWritableStream
}) : IObjectStorage

IObjectStorage
    readAll(IReadableStream).then(records => undefined)

    add(object)
    update(object)
    remove(object)

    writeAll(IWritableStream).then(count => undefined)

[ArnaudBuchholz]

Interesting behavior to implement:

• When loading the storage, computes how much data is split over the file (0 means 1 line = 1 record, 100% means all records have more than one line)
• Over a treshold, consolidate the file automatically