<!---  
SPDX-FileCopyrightText: Hermine team <hermine@inno3.fr> 
SPDX-License-Identifier: CC-BY-4.0
-->

# Installing Hermine for production


## Getting the code

You can download latest releases from [Hermine releases page](https://gitlab.com/hermine-project/hermine/-/releases)
or clone latest development version from [GitLab](https://gitlab.com/hermine-project/hermine/-/tree/main):

```bash
git clone https://gitlab.com/hermine-project/hermine.git
```

You can also use git to clone a specific version :
    
```bash
git clone --branch v0.0.1 https://gitlab.com/hermine-project/hermine.git
```

Major versions changes mean breaking changes, either in the installation configuration or in the API.
You can find more information in [CHANGELOG.md](https://gitlab.com/hermine-project/hermine/-/blob/main/CHANGELOG.md).

For minor or patch versions, it should be safe to update your instance by pulling the latest tag from the repository or
downloading from the release page.


## Docker Compose

Hermine provides a [Docker Compose](https://docs.docker.com/compose/) configuration with the
following services:
* a PostgreSQL database
* a [gunicorn](https://gunicorn.org/) server for the Python backend
* a [Caddy](https://caddyserver.com/) server to
[serve static files](https://docs.djangoproject.com/en/4.1/howto/static-files/deployment/)
and proxy other requests to gunicorn

Beware this config is for `docker compose`
([Compose v2](https://docs.docker.com/compose/migrate/))
not for `docker-compose` (Compose v1).

Two profiles are available :
* an `https` profile where Caddy is configured with automatic HTTPS. It can easily be
deployed on a VPS.
* a `localhost` profile to use Hermine on a local machine or behind a reverse proxy (not suited for development) 

Configuration is made through a [`.env` file](https://github.com/bkeepers/dotenv) which should be
placed at the root of the project. Relevant [configuration variables
for Hermine image](#image-configuration) are exposed through Compose with
the prefix `HERMINE_`. Check [docker-compose.yml](https://gitlab.com/hermine-project/hermine/-/blob/main/docker-compose.yml)
comments for more information.

At first launch, a superadmin user is created with `admin / admin` credential.
You can update these credentials from `http://example.com/admin/auth/user/`.

In case you need to access the `django-admin` tool from outside Docker, you should use :
`docker exec -it hermine_django_1 /opt/hermine/manage.py`

To automatically init the database with [reference data](reference_data.md), just put the `shared.json` file in the `docker/` directory
before starting the containers.

### HTTPS profile (for deployment on a VPS)

You just need to set `HERMINE_SECRET` and `HERMINE_HOST` environment variables before
you can start the containers. The easiest way to do so is to
write it in a .env file.

```bash
# configure secret key
echo "HERMINE_SECRET=RANDOMSTRINGFORSECURITY" > .env
# optional : configure HOST if you use something else than localhost:80
echo "HERMINE_HOST=example.com" >> .env
# start the services in background
docker-compose --profile https up -d
```

Hermine should be accessible at `https://example.com`. Caddy automatically sets up
and renew HTTPS certificates.

To update your instance : 

```bash
cd hermine/
git switch main && git pull
docker-compose --profile https up -d --build
```

### Localhost profile (to use behind a reverse proxy)

You must set `PORT` insted of `HERMINE_HOST` variable.

```bash
# configure secret key
echo "HERMINE_SECRET=RANDOMSTRINGFORSECURITY" > .env
# configure port
echo "PORT=9000" >> .env
# start the services in background
docker-compose --profile localhost up -d
```

## Hermine Docker image 

You can also use the Hermine Docker image to run Hermine in your own Docker environment,
without using Docker Compose. You will need to set up your own PostgreSQL database, as well as a reverse proxy to
serve static files and proxy other requests to the Docker image.

You first need to build the image :

```bash
docker build -t hermine .c
```

### Image configuration

Runtime configuration is made through environment variables.
You can use a `.env` file to set them up.

The minimal configuration is :
* **SECRET**: the secret key used by Django to sign session cookies
* **DJANGO_PORT**: the port on which the container will listen for requests 
* **POSTGRES_HOST**: the hostname of the PostgreSQL database
* **POSTGRES_PORT**: the port of the PostgreSQL database
* **POSTGRES_NAME**: the name of the PostgreSQL database
* **POSTGRES_USER**: the username to connect to the PostgreSQL database
* **POSTGRES_PASSWORD**: the password to connect to the PostgreSQL database
* **SUPERUSER**: the username of the superuser created at first launch
* **PASSWORD**: the password of the superuser created at first launch
* **HOST**: the external url of your Hermine instance (ex: `example.com`), should be the same as the one used in the reverse proxy configuration
* **TRUST_PROXY_HEADERS**: should be set to "True" as you are using a reverse proxy
* **STATIC_ROOT**: the directory of the volume where static files will be stored

Optional configuration :
* **CSRF_TRUSTED_ORIGINS**: a list of trusted origins for CSRF protection, defaults to `[HOST]`
* **THREADS**: the number of threads used by gunicorn workers, defaults to number of CPU cores 
* **MAX_UPLOAD_SIZE**: the maximum size that the SBOM to import can be (in bytes), defaults to 10MB

You will need to be able to access the STATIC_ROOT directory from your reverse proxy,
and to serve its content on `example.com/static/`, so you should probably mount it as a volume.
Read more about [serving static files](https://docs.djangoproject.com/en/5.0/howto/static-files/deployment/)
in Django documentation.

Example of command to run the container :
```bash
docker run -d \
  --name hermine \
  --volume /host/path/to/static_directory:$STATIC_ROOT hermine
  --env-file .env \
  hermine
```

## Manual install

### Install python dependencies

You should run Hermine in a Python virtual environnement.
Using [poetry](https://python-poetry.org/docs/), you can create the
virtual environment and install the dependencies with:

```bash
cd hermine/
poetry install
```

### Install and build front modules

```bash
npm install
npm run build
```

### Configure your instance

Before the first run, you have to create a `config.py` file in the `hermine/hermine/` directory.

```
cp hermine/hermine/config.default.py hermine/hermine/config.py
```
Update `hermine/hermine/config.py` according to your configuration.
Ex: For development purposes, you can set `HOST` to `127.0.0.1` and `DEBUG` to `True`

Here are the parameters you may want to change in your `hermine/hermine/config.py` file

| Parameter name | Description | Default value |
| -------------- | ----- | ------------- |
| HOST |  The url of your Hermine instance | "<span>example.com</span>" |
| MAX_UPLOAD_SIZE | The maximum size that the SBOM to import can be (in bytes) |10\*1024\*1024|
| SECRET_KEY | The key that Django will use to encrypt data | "your-django-secret-key" |
| DEBUG | Enables debug functionality, should be disabled in production | False |


Update the database structure and create a superuser :
```
# activate poetry shell
poetry shell

# create database structure
python hermine/manage.py migrate

# create a superuser
python hermine/manage.py createsuperuser
```


### Run the server

How you want to serve Hermine is up to you. You should be familiar with WSGI servers and reverse proxies. You can find more information in
[Django documentation](https://docs.djangoproject.com/en/4.1/howto/deployment/wsgi/gunicorn/).

A typical installation is :

* Gunicorn or another WSGI server (running Hermine's `wsgi.py`) 
* NGinx to serve static files and proxy other requests to Gunicorn

Hermine is not different from any other Django application. You can find more information in [Django documentation](https://docs.djangoproject.com/en/4.1/howto/deployment/).

For development purpose, you can simply run :

```bash
# run the server 
python hermine/manage.py runserver
```

#### Static files

Static files should be served by your web server. After
install, you must run `collectstatic` to copy static files to the `static`
directory (or any other configured in `config.py`).

```bash
python hermine/manage.py collectstatic
```

You can find more information in [Django documentation](https://docs.djangoproject.com/en/4.1/howto/static-files/deployment/).

## OAuth

You can use an OAuth2 server as authentication backend by setting OAuth parameters.

For Docker installs, the following environment variables can be set :
* **OAUTH_CLIENT_ID**
* **OAUTH_DOMAIN**
* **OAUTH_CLIENT_SECRET**
* **OAUTH_SCOPE**: list of scope separated by ','
* **OAUTH_ID_KEY**
* **OAUTH_TOKEN_URL**
* **OAUTH_AUTHORIZE_URL**
* **OAUTH_USER_URL**
* **OAUTH_USERNAME_PROPERTY** (defaults to `username`)

On manual install, you should set these parameters in the `config.py` file.

Further information can be found in [Python Social Auth’s documentation](https://python-social-auth.readthedocs.io/en/latest/configuration/django.html) 
on which Hermine relies.

Users will be created on the fly at authentication by the OAuth server.


## Health check

Hermine provides two test endpoints which you can use in your monitoring system.

`/ping/` always send a 200 response, and can be used to check server availability.

`/ready/` does the same but also tries to connect to Hermine database. It sends a 200 response if it succeeds.