imageInf

AI-powered image inferencing service that applies domain-specific categorization tags to uploaded datasets to support research workflows and data discovery.

API Endpoints

• /api/status: Health check
• /api/inference/sync: Run inference on a single image (synchronous)
• /api/inference/async: Submit a batch for async processing (coming soon)

API Documentation

• Swagger UI: http://localhost:8080/api/docs
• ReDoc: http://localhost:8080/api/redoc

Running locally

Build

make build

Start

make start

Go to: http://localhost:8080/api/status

Examples

Run a demo script or notebook to test the image inference API.

Testing and Linting

With the imageinf service running via Docker, you can open a shell inside the container to run tests:

docker exec -it imageinf bash

Run Tests

pytest

Note Some tests may download large models. To skip those, run only non-slow tests:

pytest -m "not slow"

Run Linting checks

black --check .
flake8 .

Auto-fix formatting with black:

black .

Client example (frontend)

cd client
npm ci
npm run dev

Go to http://localhost:8080/

For more info, see Client README