com.mbreissi.edgecommons.ImageProcessor is the EdgeCommons Python inference component for image
models. It loads signed, content-addressed ONNX model bundles, runs inference on finalized image
files from a spool it owns or on subscription-triggered images, and publishes one durably confirmed
inference result per image.
camera spool ─┐ ┌─► app/inference/result (confirmed, cleanup-gating)
├─► readiness ─► ledger ─┼─► data/<signal> (decision mirror, best effort)
trigger topic ┘ │ │ ├─► <image>.inference.json (evidence sidecar)
│ │ └─► processed/ | failed/ (the input, after confirmation)
│ └─ executor cell (ONNX Runtime, CUDA or CPU)
└─ model bundle cache (verified, warmed, atomically activated)
| Doc | Start here when you want to… |
|---|---|
| Tutorial | learn by doing — run one image through the component and read every output it produced |
| How-to guides | accomplish a task — build and sign a bundle, add a route, deploy, repair a stuck job |
| Reference | look up an exact config option, topic, command, or metric |
| Explanation | understand how an image becomes a job, and why the walk is authoritative |
- "I'm new here." → Tutorial.
- "What config option does X?" → Reference — Configuration.
- "What message on which topic, and what does a command reply look like?" → Reference — Messaging interface.
- "What does this metric mean?" → Reference — Metrics.
- "What does a task family produce, and how do decision rules read it?" → Reference — Normalized output and decision rules.
- "What do I put in a config for a camera route, a trigger route, or a GPU device?" → Sample configurations.
- "Why does a lost camera announcement not lose an image?" → Explanation.
These pages are for the integrator wiring a camera to a model and the operator running the result.
They describe what the component does; DESIGN.md and LLD.md in the repository root record why
it does it that way and how the modules fit together.