1-click batch prediction with fusion

1-click-batchprediction.json

{
"name": "1-click batch prediction from fusion",
"description": "This script creates a batch prediction from a user-given source and fusion ID",
"inputs": [
{
"default": "",
"description": "Source for inputs",
"name": "test-source",
"type": "source-id"
},
{
"default": "Fusion ID",
"description": "Input a valid value",
"name": "fusion",
"type": "fusion-id"
}
],
"outputs": [
{
"description": "Output description",
"name": "batch-prediction",
"type": "batchprediction-id"
}
]
}

1-click-batchprediction.py

##############################################################################
# Copyright (c) 2015-2018 BigML, Inc
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
# EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
# NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
# LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
# OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
# WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
##############################################################################


#!/usr/bin/env python
# -*- coding: utf-8 -*

import argparse
import sys
import json


from bigml.api import BigML

SUMMARY = ("Creates a batch prediction from a 1-click dataset for a fusion.")


OPTIONS = {
    # fusion ID to be used
    '--fusion': {
        'required': 'True',
        'action': 'store',
        'dest': 'fusion',
        'help': "fusion/id"},

    # delimiter in the file
    '--test-source': {
        'action': 'store',
        'dest': 'test_source',
        'default': ',',
        'help': "Path to the test file."}
}



def parser_add_options(parser, options):
    """Adds the options to the parser
    """
    for option, properties in sorted(options.items(), key=lambda x: x[0]):
        parser.add_argument(option, **properties)


def create_parser(user_options):
    """Parses the user-given parameters.
    """
    parser = argparse.ArgumentParser(
        description=SUMMARY,
        epilog="BigML, Inc")

    parser_add_options(parser, user_options)
    return parser


def main(args=sys.argv[1:]):
    """Parses command-line parameters and calls the actual main function.
    """

    # If credentials are properly set in environment variables, there's no need
    # to explicitly create the api object. Otherwise, use next code to set them

    try:
        api = BigML()
    except:
        sys.exit("The connection to BigML needs credentials. Please set them"
                 " in the environment variables.")

    command_args = create_parser(OPTIONS).parse_args(args)
    # transforming args object to dictionary
    context = vars(command_args)

    source = api.create_source(context["test_source"])
    api.ok(source)
    try:
        with open("./script") as file_handler:
            script_id = file_handler.read()
    except IOError:
        with open("./1-click-batchprediction.whizzml") as file_handler:
            script_code = file_handler.read()
        args = json.load(open("./1-click-batchprediction.json"))
        script_id = api.create_script(script_code, args)
        with open("./script", "w") as file_handler:
            file_handler.write(script_id["resource"])

    execution = api.create_execution( \
        script_id, {"inputs": [["test-source", source["resource"]],
                               ["fusion", context["fusion"]]]})
    api.ok(execution)
    batch_prediction = execution["object"]["execution"]["results"][0]
    api.download_batch_prediction(batch_prediction, "./predictions.csv")


if __name__ == "__main__":
    main()

1-click-batchprediction.whizzml

(define test-dataset (create-and-wait-dataset {"source" test-source}))
(define batch-prediction (create-and-wait-batchprediction {"fusion" fusion
                                                           "dataset" test-dataset
                                                           "output_dataset" true
                                                           "all_fields" true}))

README.md

Python example: 1-click batch prediction from a fusion

Using the python bindings and WhizzML scripts, the example creates a batch prediction from an existing fusion object for the user-given test local file.

The command options available are:

-h, --help show the help message and exit --test-source FILE_PATH path to the CSV file that contains the test inputs --fusion FUSION_ID fusion/id

Usage:

python 1-click-batchprediction.py --fusion fusion/572b130a3bbd213099002274 --test-source iris.csv

Notes:

• The iris.csv in the command line needs to be replaced by the local file that contains the input data
• BigML credentials are expected to be available through the environment variables.

Requirements

The bigml module is needed. To install it you can use pip:

pip install bigml

The code has been tested in python 2.7 with bigml 4.19.4

iris.csv