pyflowapi 0.2.0

Async service toolchain

PyFlowApi

PyFlowApi is a framework for creating REST API services, powered by PyFreeFlow and FastAPI. It allows you to configure and expose endpoints through a simple YAML configuration file, without writing additional code.

Definition

The YAML configuration requires two main keys:

• pyfreeflow: defines which non-preloaded pyfreeflow extensions should be loaded.
• api: defines the list of API endpoints to be exposed.

API configuration

Each API entry under the api key supports the following parameters:

• min_size: minimum request size.
• max_size: maximum request size.
• methods: list of allowed HTTP methods (GET, POST).
• route: endpoint route.
• version (optional): API version, defaults to 0.0.
• pipeline: the pyfreeflow pipeline definition to be executed when the endpoint is called. (See pyfreeflow documentation.)

Endpoint composition

Each endpoint is automatically exposed following the rule:

v<X>/<Y>/<route>

Where:

• X = major version number
• Y = minor version number
• route = the configured route string

Example:

v1/0/sleep

Example configuration

pyfreeflow:
  ext:
  - pyfreeflow.ext.sleep_operator
api:
- max_size: 128
  methods:
  - GET
  - POST
  min_size: 4
  pipeline:
    digraph:
    - I -> A0
    - A0 -> A1
    - A1 -> A2
    - A2 -> O
    name: sleep
    node:
    - config:
        transformer: |2
          data.I = true
          state.I = true
      name: I
      type: DataTransformer
      version: '1.0'
    - config:
        transformer: |2
          data.A = true
          state.A = 0
      name: A0
      type: DataTransformer
      version: '1.0'
    - config:
        sleep_min: 0
        sleep_max: 2
      name: A1
      type: RandomSleepOperator
      version: '1.0'
    - config:
        transformer: |2
          data.A = true
          state.A = state.A + 1
      name: A2
      type: DataTransformer
      version: '1.0'
    - config:
        transformer: |2
          local result
          for k, v in pairs(state) do
            result[k] = v
          end
          data = {body = result}
      name: O
      type: DataTransformer
      version: '1.0'
  version: '1.0'
  route: /sleep

This configuration exposes a single endpoint:

• Route: v1/0/sleep
• Method: GET/POST
• Pipeline: sleep, defined using pyfreeflow operators

Quickstart

Run the API server by providing a configuration file:

pyflowapi-server.py -c config.yaml

Reference

PyFreeFlow -> core library for pipeline definition and execution. FastAPI -> web framework powering the API layer.

License

This software is available under dual licensing:

Non-Commercial Use - AGPL v3

For non-commercial use, this software is released under the GNU Affero General Public License v3.0. Any modifications must be shared under the same license.

Commercial Use

For commercial use, please contact me.