google-drive-blobs

a simple content-addressable blob store API on top of google drive

compatible with the abstract-blob-store API and passes its test suite

API

var blobs = require('google-drive-blobs')(options)

options must be an object that has the following properties:

refresh_token: the refresh token from your current google auth session (this module doesn't handle authenticating, only refreshing) client_id: your google app client id client_secret: your google app client secret

blobs.createWriteStream(options, cb)

returns a writable stream that you can pipe data to.

options should be an object that has options filename (will be the filename in your drive) and parent (google drive parent node) properties

cb will be called with (err, response) where response is the response metadata

blobs.createReadStream(opts)

opts should be {key: md5}

returns a readable stream of data for the first file in your drive whose md5 checksum matches the md5 argument

blobs.get(md5, cb)

gets google drive metadata for the first file in your drive whose md5 checksum matches the md5 argument

blobs.remove(opts, cb)

opts should be {key: md5}

deletes file by md5

blobs.addProperty(fileId, key, val, cb)

adds a key/value pair to a file by drive fileId, calls cb with (err)

google drive API gripes

you can only search by certain properties, and md5Checksum isnt one of them.

so i either have to hash the file ahead of time (not stream upload friendly) or i have to do a second request that copies the md5Checksum that i get back from the upload to the 'custom properties' field, which IS searchable

so basically to upload a file i have to do 4 requests, one to check if a file w/ that hash exists already, one to create the upload metadata, one to upload the file contents, and one to move md5Checksum to the custom properties field

to stream a file by hash i have to do two queries, one to find the file by the custom hash property and a second to stream the downloadUrl from the result of the first request