online_predict.sh

#!/bin/bash
# Copyright 2020 Google Inc. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

# Makes API calls for AutoML Tables online prediction.
#
# Usage: bash deploy_model.sh config_file.yaml operation payload.json
#
# config_file.yaml: The configuration file used to run the training pipeline.
# operation: The specific operation to run.
#   "deploy" to deploy the AutoML Tables model.
#   "predict" to make a prediction with a payload .json file.
#   "undeploy" to undeploy the AutoML Tablem dole.
# payload.json: When doing "predict", the .json file of features.
# 
# Uses application default credentials. You may be prompted to authenticate.
# Model deployment takes as long as 15 minutes before you can make predictions.
#
# The deployment is a long-running operation (LRO), and you'll need to check
# the operation to determine when it is complete. The "name" field of the deploy
# response is used to check the operation with a command like so, replacing
# operation-name with the name from the deploy response:
# curl -X GET \
#   -H "Authorization: Bearer $(gcloud auth application-default print-access-token)" \
#   -H "Content-Type: application/json" \
#   https://automl.googleapis.com/v1beta1/operation-name

if [ "$1" == "" ]; then
  echo "Provide a configuration YAML file as the first argument."
  exit 0
fi

CONFIG_FILE_PATH=$1

do_shyaml() {
# Get a value from the config file for a given key.
  local key=$1
  < "${CONFIG_FILE_PATH}" shyaml get-value "${key}"
}

echo Reading config.
PROJECT_ID="$(do_shyaml global.destination_project_id)"
REGION="$(do_shyaml global.automl_compute_region)"
MODEL_NAME="$(do_shyaml global.model_display_name)"
CREDENTIALS="$(gcloud auth application-default print-access-token)"
MODEL_ID="$(curl -s -X GET -H "Authorization: Bearer ""${CREDENTIALS}" \
https://automl.googleapis.com/v1beta1/projects/"${PROJECT_ID}"/locations/"${REGION}"/models \
| jq -r --arg MODEL_NAME "${MODEL_NAME}" \
'.model[] | select(.displayName==$MODEL_NAME) | .name' \
| cut -d'/' -f6)"

if [ "$2" == "deploy" ]; then
  echo "Deploying model."
  OPERATION="deploy"
  PAYLOAD="\"\""
elif [ "$2" == "undeploy" ]; then
  echo "Undeploying model."
  OPERATION="undeploy"
  PAYLOAD="\"\""
elif [ "$2" == "predict" ]; then
  echo "Predicting."
  OPERATION="predict"
  if [ "$3" == "" ]; then
    echo "Specify a .json features file as the third argument when predicting."
    exit 0
  fi
  PAYLOAD=@$3
else
  echo "Unknown operation in second argument."
  exit 0
fi

CURL1="curl -X POST -H \"Authorization: Bearer \"${CREDENTIALS}"
CURL2="-H \"Content-Type: application/json; charset=utf-8\" -d ${PAYLOAD}"
CURL3="https://automl.googleapis.com/v1beta1/projects/${PROJECT_ID}/locations/${REGION}/models/${MODEL_ID}:${OPERATION}"
CURL="$CURL1 $CURL2 $CURL3"
echo "$CURL"
eval "$CURL"