[Feature Proposal] Add FastAPI wrapper for HTTP REST API endpoints

[MichielMAnalytics]

I've built a FastAPI wrapper that exposes the semantic-similarity-rating methodology through REST API
endpoints. I'd like to propose contributing this to the project as an optional component.

Repository: https://github.com/MichielMAnalytics/pipeguru-ssr-api

Motivation

While the core semantic-similarity-rating package is excellent as a Python library, many use cases would
benefit from accessing SSR functionality via HTTP endpoints:

• Microservices architecture: Services in other languages (JS, Go, etc.) could use SSR
• Web applications: Frontend apps could call SSR without Python backend
• Cloud deployment: Easy deployment as a containerized service
• API-first workflows: Teams that prefer REST APIs over Python imports

What I've Built

A lightweight FastAPI wrapper that:

• ✅ Provides POST /v1/rate endpoint for converting responses to PMFs
• ✅ Maintains clean separation from core library (thin wrapper)
• ✅ Includes Pydantic models for request/response validation
• ✅ Supports all SSR parameters (temperature, epsilon)
• ✅ Docker deployment ready
• ✅ Comprehensive tests
• ✅ OpenAPI/Swagger documentation built-in

Example Usage

curl -X POST http://localhost:8000/v1/rate \
  -H "Content-Type: application/json" \
  -d '{
    "responses": ["I really like this product"],
    "reference_sentences": [
      "I definitely would not purchase",
      "I probably would not purchase",
      "I might or might not purchase",
      "I probably would purchase",
      "I definitely would purchase"
    ],
    "temperature": 1.0
  }'

Response:
{
  "success": true,
  "ratings": [{
    "response": "I really like this product",
    "pmf": [0.05, 0.10, 0.15, 0.30, 0.40],
    "expected_value": 3.90,
    "most_likely_rating": 5,
    "confidence": 0.40
  }]
}

Proposed Integration Options

I'm flexible on how this integrates with the project. Here are a few options:

Option A: Separate directory with optional dependencies

semantic-similarity-rating/
├── semantic_similarity_rating/  # Core library (unchanged)
├── api/                          # New: FastAPI wrapper
│   ├── __init__.py
│   ├── main.py
│   ├── endpoints/
│   └── schemas.py
├── Dockerfile                    # New: For API deployment
└── pyproject.toml               # Add optional [api] dependencies

Install with: pip install semantic-similarity-rating[api]

Option B: Keep as separate repo with link

- I maintain the wrapper separately
- You add a link to it in your README under "Related Projects" or "Web API"
- Clean separation, no scope creep

Option C: Example/tutorial

- Add as examples/fastapi-wrapper/
- Shows users how to build on top of SSR
- Less maintenance commitment

[hiddensalmon]

Nice.