# Setup

## Setup: As new project

{% hint style="info" %}
Clear offers a CLI (*Command Line Interface*) which is still in Alpha. \
Documentation for building a new project with Clear + [Kemal](https://github.com/kemalcr/kemal) will be written once the feature is done. As of now, you can just follow the paragraph below.
{% endhint %}

## Setup: In existing project

```
$ crystal init app <yourappname>
$ cd <yourappname>
```

### In \`shard.yml\`

Add your dependency in the dependencies list of your `shard.yml`

{% tabs %}
{% tab title="/shard.yml" %}

```yaml
dependencies:
  clear:
    github: anykeyh/clear
    branch: master
```

{% endtab %}
{% endtabs %}

Then download the library:

{% tabs %}
{% tab title="terminal" %}

```
$ shards install
```

{% endtab %}
{% endtabs %}

### In your source code

Assuming your main entry point of your application is `src/main.cr` , you can require and initialize Clear:

{% tabs %}
{% tab title="src/main.cr" %}

```ruby
# append to your require list on top:
require "clear"

# initialize a pool of database connection:
Clear::SQL.init("postgres://postgres@localhost/my_database", 
    connection_pool_size: 5)
```

{% endtab %}
{% endtabs %}

#### Step by Step

* `require "clear"` load the source code of Clear and provide everything needed to use the library.
* `Clear::SQL.init` prepare a certain number of connection to your database. The URL is a convention used to connect to the database, and follow this schema:

```
postgres://USER[:PASSWORD]@HOST/DATABASE[?*OPTIONS]
```

More information about the URL notation can be found [here](https://crystal-lang.org/docs/database/)

* `connection_pool_size: 5` is optional but offers the possibility to concurrent fibers to run query at the same time. It's useful if you use an event-driven server, like Kemal.

### Installation customization

You may want to install a smaller version of Clear by calling :

```ruby
require "clear/core"
```

This will add clear without the build-in CLI and without some extensions (jsonb, bcrypt etc...).