Skip to content

async-mode.md

Latest commit

 

History

History
61 lines (43 loc) · 1.62 KB

async-mode.md

File metadata and controls

61 lines (43 loc) · 1.62 KB

Asynchronous Mode Configuration

This guide provides a step-by-step setup for using the asynchronous endpoints of the model wrapper.

1. Enable Asynchronous Mode

To enable asynchronous mode, set the ASYNC_ALLOW environment variable to True.

export ASYNC_ALLOW=True

2. Configure Redis

The model wrapper uses Redis for managing asynchronous jobs. Make sure you have a Redis server running and configure the following environment variables to connect to it:

  • REDIS_HOST: The hostname of the Redis server (default: localhost).
  • REDIS_PORT: The port of the Redis server (default: 6379).
  • REDIS_DB: The Redis database to use (default: 0).
  • REDIS_PASSWORD: The password for the Redis server (default: None).

3. Submit an Asynchronous Job

To submit an asynchronous job, add the async field to your request and set it to True.

Example:

{
  "service_type": "get_molecule_property",
  "service_name": "MySimplePredictor",
  "parameters": {
    "property_type": ["LogP", "TPSA"],
    "subjects": ["CCO", "CCC"]
  },
  "async": true
}

The server will respond with a job ID that you can use to retrieve the results later.

{
  "job_id": "f1b2c3d4-e5f6-7890-1234-567890abcdef"
}

4. Retrieve the Results

To retrieve the results of an asynchronous job, send a request with the service_type set to get_result and the url field set to the job ID.

Example:

{
  "service_type": "get_result",
  "url": "f1b2c3d4-e5f6-7890-1234-567890abcdef"
}

The server will respond with the results of the job, or a status indicating that the job is still pending.