ERP/README.md

# Procurement Stack v1.3 — Dockerized (PostgreSQL + FastAPI + React + MinIO)

Self-contained ship package: everything needed to run the stack is in this
folder — no external files or sibling folders required.

| Service | Container | Tech | Exposed port |
|---|---|---|---|
| **postgres** | `procurement-postgres` | PostgreSQL 18 (one instance: `procurement` + `casdoor` DBs) | `5433` |
| **minio** | `procurement-minio` | MinIO object storage (S3-compatible) | `9000` (API), `9001` (console) |
| **casdoor** | `procurement-casdoor` | Casdoor (OIDC identity provider) | `8080` |
| **api** | `procurement-api` | FastAPI + SQLAlchemy (asyncpg) | `8000` |
| **frontend** | `procurement-frontend` | React 19 + Vite, served by nginx (internal) | via Traefik |
| **traefik** | `procurement-traefik` | Traefik v3 — public reverse proxy (edge) | `80` (web), `443` (websecure) |


```

Browser ──► Traefik :80 (edge)
│
├── Host(localhost) ──► frontend (nginx SPA :80)
│                            │
│                            └── /api/* proxy ──► api :8000 (/api/v1)
│
├── Host(api.localhost) ──► api :8000
├── Host(auth.localhost) ──► casdoor :8000 (OIDC login)
└── Host(s3/storage.localhost) ──► minio (API / console)
│
├── postgres :5432 (procurement + casdoor DBs)
└── minio :9000 (invoice/proforma files)

```

---

## 📁 Layout

```text
v1.3/
├── docker-compose.yml        ← all paths are relative to this folder
├── docker-compose.prod.yml   ← production overlay (real domains + HTTPS via Let's Encrypt)
├── start.ps1                 ← PowerShell helper (build + start + status)
├── mc.cmd                    ← MinIO Client helper (inspect the bucket)
├── README.md
├── sql/
│   ├── migration.sql         ← DB schema v2.6 (auto-runs on first boot)
│   └── 02-casdoor-db.sql     ← creates the `casdoor` role+DB on the same instance
├── infra/
│   └── traefik-dynamic.yml   ← Traefik shared middlewares (HTTP→HTTPS redirect)
├── docs/                     ← ERDs + migration plan + auth guide
│   ├── erd.txt               (v2.4 — old procurement schema)
│   ├── new erd.txt           (v2.5 — material → supplier → category)
│   ├── NewERD.txt            (v2.6 — integrated procurement + inventory)
│   ├── casdoor-auth.md       (v2.6 — Casdoor login, roles, failure scenarios)
│   └── material-supplier-category-migration-plan.md
├── backend/                  ← build context of the `api` service
│   ├── Dockerfile
│   ├── requirements.txt
│   ├── .dockerignore
│   ├── app/                  (source code — procurement + inventory + auth)
│   └── tools/                (smoke-test + seed scripts, copied into the image)
└── frontend/                 ← build context of the `frontend` service
    ├── Dockerfile
    ├── nginx.conf
    ├── .dockerignore
    ├── .env.example          (copy to .env if you need a custom API base URL)
    ├── index.html
    ├── package.json
    ├── package-lock.json
    ├── tsconfig*.json
    ├── vite.config.ts
    ├── public/
    └── src/
        └── pages/
            ├── Dashboard/    (procurement office)
            ├── Wizard/       (add new offer)
            ├── Suppliers/    (suppliers & topics)
            ├── Projects/     (projects)
            ├── Units/        (unit conversions)
            └── Inventory/    (warehouse management panel — React port)

```

---

## 🔐 Login

The whole system is behind a login page (`/login`). Login is **in-page**:
the form (username / password) is sent to
`POST /api/v1/auth/login`; the backend exchanges it with Casdoor
(password grant), verifies the returned JWT, and starts the session —
no browser redirect to Casdoor.

* **مدیر پروژه (PM)** → lands in the inventory panel (their project) — can record
stock changes, manage the project's warehouses and confirm arrivals («کالاهای در راه»).
* **مدیر تدارکات / ادمین** → lands in the procurement office — can manage
quotes, suppliers, projects, and in the inventory panel: dashboard, daily control,
«کالاهای در راه» (receiving included) and «سازمان‌دهی» (users, PMs, warehouses).
The «انبارها» and «تغییر موجودی» sections are PM-only.

**Developer Mode** (top-left button on the login page) is a 3-step flow for
the admin (super user): ① superuser auth gate (admin credentials only) →
② role switcher (ادمین / مدیر تدارکات / مدیر پروژه + project) → ③ session
activation by role (admin → **پنل مدیریت ادمین** at `/admin`, مدیر تدارکات →
procurement office, مدیر پروژه → their project's inventory panel). It is
controlled by the `VITE_ENABLE_DEV_LOGIN` env var (in Docker: the
`DEV_LOGIN_ENABLED` compose build arg, default `true` — set it to `false`
for production builds).

| What | Where | Credentials |
| --- | --- | --- |
| App | http://localhost (via Traefik) | — |
| Casdoor console | http://localhost:8080 | admin credentials are set when you run the seeder (`CASDOOR_ADMIN_PW`) — never commit them |
| Seeded app users (org `inventory`) | `razaei` (admin), `mousavi` (مدیر تدارکات), `mohaghegh` (مدیر پروژه), `tester` (مدیر پروژه، پروژه «تست») | password from `CASDOOR_SEED_PASSWORD` (bcrypt-hashed at rest — see `backend/tools/seed_casdoor.py`) |

Run `backend/tools/seed_casdoor.py` once after the first boot to create the
Casdoor organization, application and users (idempotent). Full details —
addresses, credentials, role model, failure scenarios and the checklist —
are in [`docs/casdoor-auth.md`](https://www.google.com/search?q=docs/casdoor-auth.md).

> The **currently committed seed** (`backups/علی محقق/`) instead ships a
> single admin — `ali.mohaghegh` («علی محقق») — see
> [Restoring the committed seed data](https://www.google.com/search?q=%23restoring-the-committed-seed-data).

---

## 🌐 Working with the deployed system (server)

The production deployment of this stack runs on the server and is reachable
at these public addresses (HTTPS):

| What | Address |
| --- | --- |
| **App (ERP)** | https://aramesh.bennu-api-bahrestan.ir/ |
| **Casdoor admin panel** | https://casdoor.bennu-api-bahrestan.ir/ |

Log in to the app with your Casdoor account (org `inventory`); the role is
assigned inside the app (`users.role` — see «Login» above and
[`docs/casdoor-auth.md`](https://www.google.com/search?q=docs/casdoor-auth.md)).

### MinIO console on the server (SSH tunnel)

The MinIO console is deliberately **loopback-only** on the server
(`127.0.0.1:9001`) — verified unreachable from the internet — so the way in
is an SSH tunnel from your machine:

1. Open PowerShell and start the tunnel (**keep the window open** while you
use the panel):
```powershell
ssh -L 9001:127.0.0.1:9001 <user>@<server-host>

```


You will also land in a normal SSH session — that is fine; the tunnel
lives as long as the window stays open.
2. In your browser, open: **http://localhost:9001**
3. Log in with the production MinIO root credentials from the server's
`/srv/erp/.env` (`MINIO_ROOT_USER` / `MINIO_ROOT_PASSWORD`) — they are
**never** written in this repo or its history. See
[`docs/SECRET-MANAGEMENT.md`](https://www.google.com/search?q=docs/SECRET-MANAGEMENT.md).

You will see the **`procurement-files`** bucket — the invoices/proformas the
ERP stores, browsable and downloadable from there.

⚠️ Those credentials are **full admin access to the file storage**. They live
only in the server-side `.env`; if one ever leaks, rotate it immediately
(checklist in `docs/SECRET-MANAGEMENT.md`).

Close the PowerShell window (or `Ctrl+C`) when you are done to drop the
tunnel.

Notes from the server setup:

* The panel is published only on the server's loopback, so the internet
cannot reach it at all — the tunnel is what carries it to your browser.
* Host port `9000` stays reserved for the deploy-webhook receiver, which is
why the console uses `9001`.
* If you ever need the raw S3 API from your machine, or a clickable URL for
the console (e.g. `storage.aramesh.bennu-api-bahrestan.ir` behind Traefik
with a password prompt on top), that can be arranged on request — the
tunnel remains the safer default.

---

## Prerequisites (target machine)

* **Docker Desktop** (Windows) or Docker Engine + Compose plugin (Linux/macOS)
* ~3–4 GB free disk for images (`postgres:18`, `python:3.13-slim`, `node:22-alpine`, `nginx:alpine`, `minio/minio`, `minio/mc`)
* Internet access on first build (image + `npm ci` downloads)

---

## Start the whole stack

```powershell
cd <this-folder>
docker compose up -d --build

```

First run downloads base images and builds the frontend (takes a few minutes).
Subsequent runs: `docker compose up -d`.

Wait until all services are up:

```powershell
docker compose ps

```

All should show `Up` (postgres also `(healthy)`).

---

## Verify

| What | Where |
| --- | --- |
| Frontend UI | http://localhost |
| API docs (Swagger) | http://localhost:8000/docs |
| API health | http://localhost:8000/health → `{"status":"ok","service":"procurement-api"}` |
| API readiness (DB + MinIO + Casdoor) | http://localhost:8000/health/ready |
| Metrics | http://localhost:8000/metrics |
| MinIO console | http://localhost:9001 (user/pass `minioadmin`/`minioadmin`) |
| Database | `localhost:5433`, db `procurement`, user/pass `postgres`/`postgres` (Casdoor uses the same instance, db `casdoor`) |

Automated end-to-end checks (authenticates via `SMOKE_USER` /
`SMOKE_PASSWORD` — the seeder users):

```powershell
backend\tools\smoke_test.py   # read paths
backend\tools\smoke_test_write.py / smoke_test_misc.py

```

> Every API route now requires a Casdoor JWT — a bare
> `curl http://localhost/api/v1/quotes` returns **401** by design.
> Day-2 operations (deploys, backups, monitoring): [`docs/OPERATIONS.md`](https://www.google.com/search?q=docs/OPERATIONS.md).

---

## MinIO (object storage for invoice / proforma files)

| Item | Value |
| --- | --- |
| S3 API endpoint | `localhost:9000` (from the API container: `minio:9000`) |
| Web console | http://localhost:9001 |
| Root user / password | `minioadmin` / `minioadmin` (dev — change for production) |
| Bucket | `procurement-files` (created automatically by the `minio-init` container) |
| Object key layout | `{quote_id}/{kind}/{attachment_id}` where kind = `invoice` | `proforma` |

File **bytes** live in MinIO; the `quote_attachments` table holds the metadata.

Ad-hoc bucket inspection from the host (see `mc.cmd`):

```powershell
.\mc.cmd ls --recursive

```

---

## Database credentials / connection

```text
Host:     localhost
Port:     5433
Database: procurement
Username: postgres
Password: postgres

```

```powershell
docker exec -it procurement-postgres psql -U postgres -d procurement

```

---

## Recovery / seeding

The databases are **not** in git — Docker volumes hold the data, and
database dumps must never be committed (they contain user rows and signing
material). Two supported ways to bring a fresh stack up:

1. **Empty boot** — schema v2.6 auto-applies on the first boot of an empty
`postgres_data` volume; then run the seeders:
```powershell
# Casdoor org + app + users (prints/generates credentials — do not commit them)
docker compose run --rm -e CASDOOR_BASE=http://casdoor:8000 api python tools/seed_casdoor.py
# Demo dataset (optional):
docker exec procurement-api sh -c "cd /srv/backend && python -m tools.seed"

```


2. **Restore your own dump** — create dumps on the server (kept server-side,
e.g. `/srv/erp/backups/`), never in git:
```powershell
docker exec procurement-postgres pg_dump -U postgres -d procurement -Fc > /srv/erp/backups/procurement-$(date +%F).dump
docker exec procurement-postgres pg_dump -U postgres -d casdoor    -Fc > /srv/erp/backups/casdoor-$(date +%F).dump

```


Restore with `pg_restore --clean --if-exists` (stop `api` + `casdoor`
first — this **replaces all data**). The scheduled backup sidecar
(see `docs/OPERATIONS.md`) automates this.

## Common operations

```powershell
docker compose logs -f api          # see logs
docker compose down                 # stop (keep data)
docker compose up -d --build        # rebuild after code changes
docker compose down -v              # ⚠️ full reset — wipes DB + MinIO volumes

```

---

## Known behavior

* Authentication is Casdoor-based (JWT verified via JWKS) and **required on
every API route** except `/health`, `/health/ready` and `/auth/login`
(401 without a token, 403 without a role). The former dev-header fallback
(`X-User-Role` / `X-User-Name-B64`) is disabled unless
`PROCUREMENT_DEV_LOGIN_ENABLED=true` — never enable it in production.
* Roles live in the app's `users` table (Casdoor is identity-only): `pm`,
`procurement_manager`, `admin`. Role changes are **admin-only** (RBAC);
the procurement manager can only assign projects to project managers.
The admin has a dedicated panel at `/admin` (user sync from Casdoor +
grant/revoke «مدیر تدارکات»).
* **Partial goods receipt** («رسید جزیی») is supported in «کالاهای در راه»:
at least one item with a valid quantity is enough to submit; other items
stay in transit. The procurement ledger shows «رسیده X از Y» / «در راه Z»
badges and a «رسید جزیی» badge on the quote.
* **One PostgreSQL instance** serves both the app (`procurement`) and Casdoor
(`casdoor`) — the separate casdoor-postgres container is gone (data was
dumped/restored; see `docs/casdoor-auth.md`).
* **Traefik v3** is the public edge (ports 80/443); the frontend nginx only
serves the SPA and proxies `/api` internally. Production (real domains,
HTTPS via Let's Encrypt, HTTP→HTTPS redirect):
`docker compose -f docker-compose.yml -f docker-compose.prod.yml up -d --build`
(see the prod file header for the required env vars).
* Schema runs automatically ONLY on the first boot of an empty `postgres_data` volume
(see `docker-entrypoint-initdb.d`); the API also runs `CREATE TABLE IF NOT EXISTS`
for the units-dashboard tables and `quote_attachments` at startup.
* Invoice upload requires an `accepted`/`arrived` quote; proforma has no restriction.
* The supplier detail page (`/suppliers/:id`) shows the purchase history with
invoice/proforma download links from MinIO.
* **Material-Supplier Lookup**: Clicking a material in the catalog (e.g., on the `/units` page) dynamically fetches its assigned suppliers (`GET /api/materials/{id}/suppliers`). The backend uses an atomic JOIN to guarantee accurate supplier data (avoiding SQLAlchemy identity-map mismatches), and the UI displays a clean empty state («تامین‌کننده‌ای برای این متریال ثبت نشده است») if none exist.
* If builds fail with disk-full errors: `docker system prune -a`, then rebuild.
* If the frontend shows «اتصال به سرور برقرار نشد» → the api container is not
reachable; check `docker compose ps` and `docker compose logs api`.

```