[![pipeline status](https://gitlab.com/patatoid/boruta_auth/badges/master/pipeline.svg)](https://gitlab.com/patatoid/boruta_auth/-/commits/master)
[![coverage report](https://gitlab.com/patatoid/boruta_auth/badges/master/coverage.svg)](https://gitlab.com/patatoid/boruta_auth/-/commits/master)
# Boruta OAuth provider core
Boruta is the core of an OAuth provider giving business logic of authentication and authorization.
It is intended to follow RFCs:
- [RFC 6749 - The OAuth 2.0 Authorization Framework](https://tools.ietf.org/html/rfc6749)
- [RFC 7662 - OAuth 2.0 Token Introspection](https://tools.ietf.org/html/rfc7662)
- [RFC 7009 - OAuth 2.0 Token Revocation](https://tools.ietf.org/html/rfc7009)
- [RFC 7636 - Proof Key for Code Exchange by OAuth Public Clients](https://tools.ietf.org/html/rfc7636)
As it, it helps implement a provider for authorization code, implicit, client credentials and resource owner password credentials grants. Then it follows Introspection to check tokens.
## Documentation
Documentation can be found [here](https://patatoid.gitlab.io/boruta_auth/Boruta.html)
## Live example
A live example can be found [here](http://oauth.boruta.patatoid.fr/)
## Installation
1. Schemas migration
If you plan to use Boruta builtin clients and tokens contexts, you'll need a migration for its `Ecto` schemas. This can be done by running:
```sh
mix boruta.gen.migration
```
2. Implement ResourceOwners context
In order to have user flows working, You need to implement `Boruta.Oauth.ResourceOwners`.
Here is an example implementation:
```elixir
defmodule MyApp.ResourceOwners do
@behaviour Boruta.Oauth.ResourceOwners
alias Boruta.Oauth.ResourceOwner
alias MyApp.Accounts.User
alias MyApp.Repo
@impl Boruta.Oauth.ResourceOwners
def get_by(username: username) do
with %User{id: id, email: email} <- Repo.get_by(User, email: username) do
{:ok, %ResourceOwner{sub: id, username: email}}
else
_ -> {:error, "User not found."}
end
end
def get_by(sub: sub) do
with %User{id: id, email: email} = user <- Repo.get_by(User, id: sub) do
{:ok, %ResourceOwner{sub: id, username: email}}
else
_ -> {:error, "User not found."}
end
end
@impl Boruta.Oauth.ResourceOwners
def check_password(resource_owner, password) do
User.check_password(user, password)
end
@impl Boruta.Oauth.ResourceOwners
def authorized_scopes(%ResourceOwner{}), do: []
end
```
3. Configuration
Boruta provides several configuration options, to customize them you can add configurations in `config.exs` as following
```elixir
config :boruta, Boruta.Oauth,
repo: MyApp.Repo,
cache_backend: Boruta.Cache,
contexts: [
access_tokens: Boruta.Ecto.AccessTokens,
clients: Boruta.Ecto.Clients,
codes: Boruta.Ecto.Codes,
resource_owners: MyApp.ResourceOwners,
scopes: Boruta.Ecto.Scopes
],
expires_in: [
authorization_code: 60,
access_token: 3600
],
token_generator: Boruta.TokenGenerator
```
## Integration
This implementation follows a pseudo hexagonal architecture to invert dependencies to Application layer.
In order to expose endpoints of an OAuth server with Boruta, you need implement the behaviour `Boruta.Oauth.Application` with all needed callbacks for `token/2`, `authorize/2` and `introspect/2` calls from `Boruta.Oauth`.
This library has specific interfaces to interact with `Plug.Conn` requests.
Here is an example of a token endpoint controller:
```elixir
defmodule MyApp.OauthController do
@behaviour Boruta.Oauth.Application
...
def token(%Plug.Conn{} = conn, _params) do
conn |> Oauth.token(__MODULE__)
end
@impl Boruta.Oauth.Application
def token_success(conn, %TokenResponse{} = response) do
conn
|> put_view(OauthView)
|> render("token.json", response: response)
end
@impl Boruta.Oauth.Application
def token_error(conn, %Error{status: status, error: error, error_description: error_description}) do
conn
|> put_status(status)
|> put_view(OauthView)
|> render("error.json", error: error, error_description: error_description)
end
...
end
```
## Feedback
It is a work in progress, all feedbacks / feature requests / improvements are welcome