GeneTrail 2

[axa9070]

I've really started liking this tool called GeneTrail 2. There is a REST API for it and I'd like to take full advantage of it in R. There is already a client written in Python that I haven't tried using. So much of my pipelines are in R though I'd hate to have to switch.

Using R6Class I've been able to get the bare bones of the client working:

##
# Gene Trail 2 Client class
#
GeneTrail2Client <- R6Class(
    'GeneTrail2Client',
    public = list(
        initialize = function(
            username, password, sessionId )
        {
            if ( !missing( username ) && !missing( password ) )
                self$login( username, password )
            if ( !missing( sessionId ) )
                private$sessionId <- sessionId
            else private$sessionId <-self$getSessionId()
            cat( private$sessionId, '\n' )
        },
        getAPIdoc = function()
        {
            return( self$doGet( '/api-docs' ) )
        },
        login = function( username, password )
        {
            return( self$doPost( '/user/login?username=%s&password=%s', username, password ) )
        },
        getSessionId = function()
        {
            if ( is.na( private$sessionId ) )
                private$sessionId <- self$doGet( '/session' )[['session']]
            return( private$sessionId )
        },
        uploadGenes = function( genesFile, displayName )
        {
            cat( paste0( 'https://genetrail2.bioinf.uni-sb.de/results.html?session=', private$sessionId, '&show_all_results=true\n' ) )
            return( fromJSON( RCurl::postForm(
                sprintf( paste0( private$GT_URL, '/upload?session=%s' ),
                    private$sessionId ),
                file = fileUpload( genesFile ) ) ) )#, displayName = displayName ) ) )
        },
        getJobAlgorithms = function()
        {
            return( self$doGet( '/job/algorithms' ) )
        },
        getJobAlgorithmParameters = function( algorithm )
        {
            return( self$doGet( '/job/parameters/%s', algorithm ) )
        },
        startJob = function( contact )
        {
            if ( missing( contact ) )
                return( self$doGet( '/job/start?session=%s', private$sessionId ) )
            else
                return( self$doGet( '/job/start?session=%s&contact=%s', private$sessionId, contact ) )
        },
        setupJob = function( algorithm, ... )
        {
            return( self$doPostForm(
                sprintf( paste0( private$GT_URL, '/job/setup/%s?session=%s' ),
                    algorithm, private$sessionId ), ... ) )
        },
        stopJob = function()
        {
            return( self$doGet( '/job/stop?session=%s', private$sessionId ) )
        },
        runGSEA = function( genes, jobName )
        {
            job <- self$uploadGenes( genes, jobName )
            return( job )
        },
        getGT_URL = function()
        {
            return( private$GT_URL )
        },
        doGet = function( endpoint, ... )
        {
            return( fromJSON( RCurl::httpGET(
                sprintf( paste0( private$GT_URL, endpoint ), ... ),
                httpheader = list(
                    Accept         = 'application/json',
                    `Content-type` = 'application/x-www-form-urlencoded' )
                ) ) )
        },
        doPost = function( endpoint, ... )
        {
            return( fromJSON( RCurl::httpPOST(
                sprintf( paste0( private$GT_URL, endpoint ), ... ),
                httpheader = list(
                    Accept         = 'application/json',
                    `Content-type` = 'application/x-www-form-urlencoded' )
                ) ) )
        },
        doPostForm = function( url, ... )
        {
            return( fromJSON( RCurl::postForm( url, style = 'POST', ... ) ) )
        },
        doPut = function( endpoint, content, ... )
        {
            return( fromJSON( RCurl::httpPUT(
                sprintf( paste0( private$GT_URL, endpoint ), ... ),
                content,
                httpheader = list(
                    Accept         = 'application/json',
                    `Content-type` = 'application/x-www-form-urlencoded' )
                ) ) )
        },
        api = list(
            API_DOC = '/api-docs',
            USER    = list(
                login = '/user/login?username=%s&password=%s'
            ),
            SESSION = list(
                get    = '/session',
                PUTseal   = '/session/%s/seal',
                DELETE = '/session/%s'
            ),
            UPLOAD = list(
                uploadGenes = '/upload?session=%s&displayName=%s'
            ),
            JOB     = list(
                getAlgorithms = '/job/algorithms',
                getParameters = '/job/parameters/%s',
                startJob      = '/job/start?session=%s',
                setupJob      = '/job/setup/%s?session=%s',
                stopJob       = '/job/stop?session=%s'
            ),
            RESOURCE = list(
            )
        )
    ),
    private = list(
        GT_URL = 'https://genetrail2.bioinf.uni-sb.de/api',
        sessionId = NA
    )
)

My hope is someone will eventually just create the R package. (I wish I had the ability to do so myself.) Until that happens though, is there anyone that could assist me in figuring out how to, through the API, retrieve recommended parameters. This is a feature you get through the web portal where the web applications reports: "Based on your uploaded data we recommend the following parameters." but this is something I'd like to simply be able to kick off with default parameters.

If it helps, I've attached a POSTMAN json file with a bunch of calls to the API which is listed here with Swagger UI. (Sadly, the Swagger UI is BROKEN... the url in the api calls
'https://genetrail2.bioinf-uni-sb.de/api/api-docs/' when it is supposed to be
'https://genetrail2.bioinf.uni-sb.de/api/api-docs/'. smh)

Thanks,
Alex