Made the readme richer

Author: L1m3D4rkn3ss

README.md

@@ -1,6 +1,8 @@
 # Isquare client for Python
 
-This repository contains the client for [isquare](isquare.ai). It is available under the form of python classes, as well as a command-line-interface.
+This repository contains the official python client for [ISquare](isquare.ai). It is available under the form of python classes which are ready to use in your code, as well as a command-line-interface. We currently support inference with image, text & json files, as well as any numpy array or python dictionnary or string, both for input and output. 
+
+The complete documentation for ISquare can be found [here](docs.isquare.ai)
 
 ## Installation
 
@@ -16,27 +18,30 @@ pip install --editable .
 
 ### Additional requirements
 
-To be able to test your models, you need the following packages:
+To be able to test your model builds, you need the following packages:
 Docker >= 19.03.13
 
 _Note_: If you only need the client for inference, this is not required.
 
 ## Usage
+The client can be used to verify your model build (e.g. checking if they will properly run on [ISquare](isquare.ai)) and to perform inference calls to your deployed models. To use this client for inference, you need to have a model up and running on [ISquare](isquare.ai).
 
 Commands and their usage are described [here](docs/commands.md).
 
-Guidelines on the code adaptation required to deploy a model on isquare.ai can be found [here](docs/isquare_tutorial.md)
+End-to-end guidelines on the code adaptation required to deploy a model on isquare.ai can be found [here](docs/isquare_tutorial.md)
 
 ## Examples
 
-- Build your i2 compatible docker image:
+### Command line interface
 
+#### Test if your model repository is Isquare-compatible
+To verify if your code will run smoothly on [ISquare](isquare.ai), you can perform a local build & unit test. This will build a container image with all your specific dependencies and perform an inference test. We've included an example of a simple computer vision model which returns the mirrored image it is given, and it can be tested by running:
 
 ```bash
  i2 build examples/tasks/mirror.py
 ```
+When you deploy a model with [ISquare](isquare.ai), you will be provided a url for the model, and requested to create access keys. Using a valid url & access keys (the one displayed are an example), you can perform an inference with an Image model (e.g. the Mirror) and a `.png` image by running:
 
-Simple inference:
 
 ```bash
 i2 infer \
@@ -46,4 +51,22 @@ i2 infer \
 ```
 Other examples can be found [here](docs/getting_started.md).
 
+### Using a model inside your python code
+As you probably want to automate your model calls by integrating them directly into your code, we've provided you with several python classes you can directly use in your code. The main class to use for that is the `I2Client` class. A simple inference can be performed as follows:
+
+```python
+from i2_client import I2Client
+import cv2
+
+# You need your url, access key and an image
+url = "wss://archipel-beta1.isquare.ai/43465956-8d6f-492f-ad45-91da69da44d0"
+access_key = "472f9457-072c-4a1a-800b-75ecdd6041e1"
+img = cv2.imread("test.jpg")
+
+# Initialize the client & perform inference
+inference_client = I2Client(url,access_key)
+success, output = i2_client.inference(img)[0]
+```
+
+A more complex example, showing how to stream a camera to your model, can be found [here](examples/webcam_stream.py)