# ResumeLens

ResumeLens is a web app that grades a resume against a job description.

- Backend: Go HTTP API (`/api/analyze`)
- Frontend: React + Vite
- AI provider: OpenAI (server-side only)

## What the project does

1. Accepts a PDF resume upload and job description text.
2. Extracts text from the PDF on the backend.
3. Sends resume text + job description to the LLM.
4. Returns structured JSON with score and feedback.

## Tech stack

- Go 1.25.x
- Node 20+
- React 19 + Vite 7
- Docker + Docker Compose (optional)

## Prerequisites

- `go` installed
- `node` and `npm` installed
- `OPENAI_API_KEY` set for real analysis calls

Example:

```bash
export OPENAI_API_KEY="your_key_here"
```

## Run locally

### 1) Start backend API

From repo root:

```bash
go run ./cmd/server
```

Backend listens on `http://localhost:3000`.

### 2) Start frontend

In another terminal:

```bash
cd web
npm ci
npm run dev
```

Frontend runs on Vite default port (usually `http://localhost:5173`).

By default in development, frontend calls backend at `http://localhost:3000`.

## Run with Docker Compose

From repo root:

```bash
docker compose up --build
```

- Frontend is exposed at `http://localhost:3005`
- Frontend proxies `/api` traffic to the backend container
- Backend uses values from root `.env`

## API contract

`POST /api/analyze` with `multipart/form-data`:

- `resume`: PDF file
- `job_description`: plain text

Returns JSON analysis result.

Notes:

- Request size is capped at 10 MB.
- Rate limit is 10 analyze requests per hour per IP.

## Tests

Run all default Go tests:

```bash
go test ./...
```

Run tests with race detector:

```bash
go test -race ./...
```

There is an opt-in live OpenAI consistency test. It is skipped by default.

```bash
RUN_LIVE_OPENAI_TESTS=1 OPENAI_API_KEY="your_key_here" go test ./internal/services -run TestCallLLM_Live_ScoreConsistencyPlusMinus10 -v
```

Frontend checks:

```bash
cd web
npm run lint
npm run build
```

## Handoff notes

- Verification strategy doc: `doc/verification-strategy.md`
- Main backend entrypoint: `cmd/server/main.go`
- API route mount: `internal/api/routes.go`
- Analyze handler: `internal/handlers/analyze.go`
- LLM service: `internal/services/analyzer.go`

If analysis fails in a fresh environment, first check that `OPENAI_API_KEY` is set in the backend process environment.