rt106SpecificAdaptorCode.py

# Copyright (c) General Electric Company, 2017.  All rights reserved.

# Rt 106
# Template for integrating a new algorithm.
# This is one of two required source files that need to be adapted for integrating an algorithm:
#     This Python script is for marshalling data and running the algorithm.
#     A JSON file that contains definitions with declarations.

# This template file has places where you need to add your own code for marshalling data and running your algorithm.
# If left unchanged, the template will result in a skeleton "algorithm" the registers itself and trivially responds
# to request messages.

# Please follow the numbered steps 1-5 below.

import os, glob, uuid, time, json, string, logging

# function: run_algorithm() -- Python function for marshalling your data and running your algorithm.
# parameters:
#   datastore: object to be used when interacting with the Data Store
#   context:  A JSON structure that should contain all the inputs and parameters your algorithm needs.
def run_algorithm(datastore, context):

    # 1.    Add code here for marshalling your inputs.  You may receive URIs in 'context' that reference image or data
    #       files.  You need to fetch that data using the Data Store API.

    #       The 'datastore' object provides a method named 'retrieve_series', which takes two inputs. The 1st input identifies
    #       which series data should be retrieved from the datastore. The individual components of the path 'patient', 'study' and
    #       'series', identify the series data. The 2nd input specifies where the data should be placed within the container.
    #
    #        Following ia an example of using this method:
    for f in glob.glob('/rt106/input/*') + glob.glob('/rt106/output/*'):
        os.remove(f)

    # Make sure we have received the needed input.  We can't do anything without the inputSeries.
    if (context['inputSeries'] == "" or not context['inputSeries']):
        status = "ERROR_NO_INPUT_SERIES"
        return { 'result' : {}, 'status' : status }

    input_path = context['inputSeries']
    #logging.info('input_path: %r' % input_path)
    response_retrieve = datastore.retrieve_series(input_path,'/rt106/input')
    #logging.info('response_retrieve: %r' % response_retrieve)
    
    # 2.    Add code here for calling your algorithm.  This may be an "exec" of an external command-line driven process.
    #       Alternately, it may be a call into a Python library that you provide.
    #       This call should wait for your algorithm to complete to receive status and results.
    #       (Asynchronism and parallelism are handled elsewhere.)

    # Your algorithm should be called here.

    # 3.    Based on the success or failure of your algorithm execution, set the 'status' variable to an appropriate string.
    status = "EXECUTION_FINISHED_SUCCESS"
    time.sleep(2)  # Temporary sleep to test sequencing of events in the execution scenario.  Remove this when you add your algorithm.

    # 4.    If you have result files, you need to store them in the Data Store, and obtain the URIs that refer to those files.

    #       Tha 'datastore' object provides a method named 'upload_series', which takes two inputs. The 1st input specifies
    #       the identity of the series to be created on the datastore. The 2nd input identifies where the data should be retrieved
    #       from the container, and placed into the series.
    #
    #       Following is an example of using this method:
    #
    #           datastore.upload_series('patientID/studyID/seriesID','/tmp/output')
    #
    
    output_path = datastore.get_upload_series_path(input_path)
    #logging.info('output_path: %r' % output_path)
    response_upload = datastore.upload_series(output_path,'/rt106/input')
    #logging.info('response_upload: %r' % response_upload)
    
    # 5.    You need to create the result_context JSON structure containing your results, including URIs obtained above in step 4.
    #       This example returns input series from about as an output series.
    #       Replace this code with whatever is appropriate for your algorithms outputs.

    output_info = ""
    if response_retrieve == 404:
        status = "EXECUTION_ABORTED_FILE_NOT_FOUND"
    elif response_upload == 409:
        status = "EXECUTION_ABORTED_UPLOAD_PATH_CONFLICT"
    else:
        output_info = response_upload.get('path')

    result_context = {
        "outputSeries" : output_info,
        "calculatedValue"  : 999
    }

    return { 'result' : result_context, 'status' : status }