# Network
## Information
The motivation for using this wrapper is to make easier the call APIs and gather responses and errors from the server. The idea is to make it as verbose as possible using a builder pattern so the request can be read at first glance.

Before making the network call, the wrapper will check a number of conditions within the "check_request_data" method which will throw a number of errors or warnings if the request is not well built. The wrapper is permissive with the definition of REST APIs, so you can have a GET call with body or a POST with GET parameters in the url (TBD).

Note: To achieve this effect, it is necessary to call new() and pass a node to the "launch_request" method, which will self manage all that's necessary for it to work.

## Usage
To use the wrapper for network requests simply build a Network_Call object and start configuring it with the to, with, etc... methods.

* to (String) -> url destination
* with (String) -> request body
* add_get_param (String,String) y add_all_get_param(Dictionary) -> add get params to request
* add_header(String,String) y add_all_header(Dictionary) -> add headers to request
* verb(Http_Method) -> set the method for REST request
* in_time(int) -> set the timeout time to the request
* set_on_call_success (Callable) -> call this function if the request was ok (200)
* set_on_call_fail (Callable) -> call this function if the request was fail (>400)
* no_cache -> set no cache for request ( only used in GET requests )

## Example
```GDScript
	func _onReady():
		Network_Call.new().to("https://catfact.ninja/fact").set_on_call_success(on_cat_fact).set_on_call_fail(on_error).launch_request(self)
	
	func on_cat_fact(result):
		$Label.text = str(result)
	
	func on_error(error):
		$Label.text = "Error requesting fact about cats :("
```
## Cache
The network module has a cache for GET requests to save time and bandwidth for similar requests in 'short' time spans.

Since nodes are ephemeral (network requests are nodes that disappear) the cache is stored on disk and not in memory (this resposability was left for the upper layers).

The cache works as follows: 
* 1. The request is hashed ( using the url ) 
* 2. A file with the specific name ( the hash ) is searched for on disk. 
* 3.a If it exists and it has not passed 300 seconds ( arbitrary and improvable time using an etag? ) the request is resolved and returns the file content. 
* 3.b If the file does not exist or it has been more than 300 seconds, the request is made and cached using the same hash.

The time validity of this cache is on CACHE_TIME_IN_SECONDS constant in network_call.gd

The cached content is an array of bytes due to the heterogeneous nature of the possible requests (text, images, sound...).