pm_service_api_develop_score.md

copyright	years 2016, 2017
lastupdated	2017-08-09

{:new_window: target="_blank"} {:shortdesc: .shortdesc} {:screen: .screen} {:codeblock: .codeblock} {:pre: .pre}

Scoring with a deployed predictive model

POST http://{PA Bluemix load balancer URL}/pm/v1/score/{contextId}?accesskey={access_key for this bound application}

Use this API call to post the input data to use by the deployed model to generate and return the predictive analytics in the score results.

Request example:

    Content-Type: application/json;charset=UTF-8
    Parameters:
        Path parameters:
            contextId: the identifier of the deployed model to use to process this score request
        Query Parameters:
            accesskey: access_key from env.VCAP_SERVICES
        Body: the input data, json string, eg.
            {
                "tablename":"DRUG1n.sav", 
                "header":["Age", "Sex", "BP", "Cholesterol", "Na", "K", "Drug"], 
                "data":[[43.0, "M", "LOW", "NORMAL", 0.526102, 0.027164, "drugY"]]
            }   

{: codeblock}

Example of a successful response to the previous request:

    Content-Type: application/json;charset=UTF-8
    Status code: 200
    body: the score result, a json array, eg.
        [
            {
                "header":["Age","Sex","BP","Cholesterol","Na""K","Drug","$N-Drug","$NC-Drug"], 
                "data":[[23.0,"M","NORMAL","NORMAL",0.78452,0.055959,"drugX","drugX",0.9892564426956728]]
            }
        ]

{: codeblock}

Response when scoring request fails:

    Content-Type: application/json
    Status code: 200
    body:
        {
           "flag":false, 
           "message":"reason"
        }  

{: codeblock}