Prepare for AI Classification

[tyler-spangler6]

Background

The VA CAIO is working to build a machine learning classifier to expand the contention classification coverage. After the model is built and ready for inference, the CAIO team is going to establish an endpoint to provide the classifications for VA.gov claims that the Conditions Team processes. This PR establishes an initial framework to process and send requests to this future service.

The endpoint is not yet established, so this PR focuses on building the initial structure to include establishing a client that will perform the API Request and structuring the request and response data to send back to vets-api.

---

Related Tickets

• #677 Create Structure for ML Classification

---

Summary of File Changes

• src/python_src/api.py
  • Create dummy endpoint for use in testing the API Client
  • Create hybrid classification endpoint to first classify with expanded classification and only send unclassified contentions to ml endpoint
• src/python_src/util/classifier_utilities
  • Helper functions to subset unclassified contentions, call the API client, and reformat the API Response to send back to vets-api
• src/python_src/pydantic_models
  • Add AiRequest and AiResponse models to send to endpoints and make it easier to process the response
  • NOTE: This might be temporary and need to be adjusted based on the endpoint that the AI team establishes
• src/python_src/util/api_client
  • Establishes the client and post method to call the dummy endpoint. After we have the url and endpoint for the AI Classifier, this should be able to be updated to accept and return the expected API Models.
• src/python_src/util/app_config
  • New fields for the dummy endpoint url and endpoint name. This will be updated after the AI team establishes the endpoint.