Added documentation

README.md

@@ -0,0 +1,44 @@
+# Sentiment Analysis Tool
+
+Small web app that predicts facial emotions from images using an ONNX model and a Next.js frontend.
+
+Key pieces
+- Frontend: `frontend/` — Next.js app (TypeScript) serving the UI and demo at `/demo`.
+- Model API: `model-lab/` — FastAPI app that loads `models/onxx_models/emotion-ferplus-8.onnx` and exposes a `/predict` endpoint.
+
+Quick start (frontend)
+
+```bash
+cd frontend
+npm install
+npm run dev
+# open http://localhost:3000
+```
+
+Quick start (model API, local Python)
+
+```bash
+cd model-lab
+python -m venv .venv
+source .venv/bin/activate
+pip install -r requirements.txt
+# from project root (recommended):
+uvicorn app.main:app --reload --host 0.0.0.0 --port 8080
+# or if you're inside model-lab/app:
+uvicorn main:app --reload --host 0.0.0.0 --port 8080
+# API will be available at http://localhost:8080
+```
+
+Quick start (model API, with Docker)
+
+```bash
+docker build -t sentiment-model -f model-lab/Dockerfile model-lab
+docker run -p 8080:8080 sentiment-model
+# API will be available at http://localhost:8080
+```
+
+Usage
+- Demo UI: http://localhost:3000/demo
+- API: `POST /predict` with multipart `file` (image) and optional `model_option` form field.
+
+More detailed developer instructions are in the `docs/` folder.

docs/architecture.md

@@ -0,0 +1,23 @@
+# Project Architecture
+
+Overview
+
+This repository contains two primary parts:
+
+- Frontend (`frontend/`): Next.js application that provides the UI and a small demo at `/demo`.
+- Model service (`model-lab/`): FastAPI-based Python service that loads an ONNX emotion detection model and exposes a `POST /predict` endpoint.
+
+Key files and responsibilities
+
+- `frontend/` — Next.js app and UI components in `frontend/components` and `frontend/app`.
+- `model-lab/app/main.py` — FastAPI app; accepts multipart image uploads and returns predicted emotion and probabilities.
+- `model-lab/app/model.py` — Model wrapper that loads the ONNX model and exposes a `predict()` method.
+- `model-lab/models/onxx_models/emotion-ferplus-8.onnx` — The ONNX emotion recognition model file.
+- `model-lab/Dockerfile` — Container image for the model service (runs Uvicorn on port 8080).
+
+Runtime flow
+
+1. User uploads an image via the frontend (or `curl`).
+2. Frontend sends a multipart `POST` to `http://<model-host>:8080/predict`.
+3. FastAPI endpoint validates the image, instantiates `Model(model_path, model_option)` and calls `predict()`.
+4. The API returns JSON `{ "emotion": <label>, "probabilities": <map> }`.

docs/setup.md

@@ -0,0 +1,48 @@
+# Setup and Development
+
+This document lists commands to run the frontend and model API for development.
+
+Prerequisites
+- Node.js (18+ recommended) and `npm` (or `pnpm`/`yarn`) for the frontend.
+- Python 3.10 for the model API (the Docker image uses 3.10-slim).
+- Docker (optional) if you prefer containerized model service.
+
+Frontend (development)
+
+```bash
+cd frontend
+npm install
+npm run dev
+# opens on http://localhost:3000
+```
+
+Frontend (build)
+
+```bash
+cd frontend
+npm install
+npm run build
+npm start
+```
+
+Model API (local Python)
+
+```bash
+cd model-lab
+python -m venv .venv
+source .venv/bin/activate
+pip install -r requirements.txt
+# run from repo root or inside model-lab as described in README
+uvicorn app.main:app --reload --host 0.0.0.0 --port 8080
+```
+
+Model API (Docker)
+
+```bash
+docker build -t sentiment-model -f model-lab/Dockerfile model-lab
+docker run -p 8080:8080 sentiment-model
+```
+
+Notes
+- The FastAPI app configures CORS; by default the frontend origin is `http://localhost:3000`.
+- The ONNX model used is at `model-lab/models/onxx_models/emotion-ferplus-8.onnx`.

docs/usage.md

@@ -0,0 +1,22 @@
+# Usage Examples
+
+Using the demo UI
+
+1. Start the frontend: `cd frontend && npm run dev` (port 3000).
+2. Start the model API (see `docs/setup.md`) on port 8080.
+3. Open `http://localhost:3000/demo` and upload an image.
+
+Direct API example (curl)
+
+```bash
+curl -X POST "http://localhost:8080/predict" \
+  -F "file=@/path/to/face.jpg" \
+  -F "model_option=1"
+
+# Example response:
+# { "emotion": "happy", "probabilities": {"happy": 0.88, "neutral": 0.07, ...} }
+```
+
+Notes
+- The API expects an image file (content-type starting with `image/`).
+- `model_option` is an optional form field that is forwarded to the model wrapper.