hch/elastic-ml-bundle

Introduction

The ElasticMlBundle is a Symfony bundle that integrates Elastic Machine Learning (ML) features from Elasticsearch into your Symfony application. This bundle allows you to create and manage Machine Learning jobs via Elasticsearch's ML API.

Features

• Create ML jobs in Elasticsearch
• Retrieve ML job results
• Flexible configuration for Elasticsearch hosts
• Example Symfony command to easily create jobs via the CLI

Installation

1. Install the bundle via Composer:

   composer require hch/elastic-ml-bundle
   

2. Enable the bundle in config/bundles.php:

   return [
   	Hch\ElasticMLBundle\ElasticMLBundle::class => ['all' => true],
   ];
   
   

3. Configure the Elasticsearch host in config/packages/elastic_ml.yaml:

   elastic_ml:
       host: 'http://localhost:9200'
   

Usage

Creating an ML Job via Symfony Command:

php bin/console elastic:ml:create-job --jobId=my-job-id --index=my-index

Example Output:

Job created successfully!

• Handling Elasticsearch Exceptions: In case of errors such as misconfigured requests or Elasticsearch being down, you will receive detailed error messages from the API.

Troubleshooting:

• 403 Forbidden: License non-compliant for ML: If you encounter this error, it means your Elasticsearch license does not support Machine Learning. You have the following options:

Activate a trial license:

curl -X POST "http://localhost:9200/_license/start_trial?acknowledge=true"

. Upgrade your Elasticsearch license to a Gold or Platinum subscription to access ML features.

. No alive nodes: All the nodes seem to be down: This error indicates that Elasticsearch is either not running or unreachable.

Ensure Elasticsearch is running:

   sudo systemctl start elasticsearch

If using Docker, ensure the container is running:

   docker start <elasticsearch-container-id>

• Invalid job request: analysis_config doesn't support START_ARRAY: This error occurs when the analysis_config in the request is incorrectly formatted. Ensure the detectors field in analysis_config is an array of objects. Here's a valid example:

  analysis_config: bucket_span: "5m" detectors: - function: "mean" field_name: "response_time"

Configuration

• Ensure correct Elasticsearch configuration in config/packages/elastic_ml.yaml:

  elastic_ml: host: 'http://localhost:9200'

Development

Code Structure:

• Service: The service ElasticMLService provides the core functionalities to create ML jobs and retrieve results from Elasticsearch.

• The constructor takes the Elasticsearch host URL as a parameter.

• Command: The Symfony command CreateMLJobCommand uses the service to create ML jobs via the CLI.

• Example Service Usage: In your Symfony controllers or services, you can use the ElasticMLService like this:

  $mlService = $this->get(ElasticMLService::class); $params = [ 'job_id' => 'my-job-id', 'analysis_config' => [ 'bucket_span' => '5m', 'detectors' => [ [ 'function' => 'mean', 'field_name' => 'response_time' ] ] ], 'data_description' => [ 'time_field' => '@timestamp' ], 'datafeed_config' => [ 'indices' => ['my-index'] ] ]; $mlService->createJob($params);

Testing

php bin/phpunit

Future Enhancements

• Allow users to pass dynamic job parameters via the Symfony command.
• Extend the bundle to support other machine learning engines, such as Google AI or AWS SageMaker.
• Automatically handle license upgrades or fallback gracefully when ML features are unavailable.

License

This bundle is licensed under the MIT License. See LICENSE for more information.