A generic API framework designed for quick model development, model retraining to classify and predict real-time data input.

Download keycloak from https://www.keycloak.org/downloads unzip the downloaded file and then run the application following the instruction here.

Create a new client called "api", then set client Authenticator to "Client Id and Secret".

Create a new user for the Client.

Then use the client use along side of the secret to get token from keycloak using postman.

then use the token in the header of the call to the API

finally set the body of the post with the data needed to be classified.

you can adjust the config file via the config.py file . The default config is as below:

DATASET = {
'days': 15,
'path': './',
'dataset_file_name': 'data/*.csv',
'label':'label',
'label_map':{
  'norm': 0,
  'mirai':1,
  'udp': 2, 
  'dns':3,
  'ack':4,
  'ack2':5
}
}

MODEL = {
'model_name': 'model.json',
'model_weight': 'model.h5',
'model_name_tfjs': 'model.js',
'validation_split': 0.3,
'epochs': 1,
'loss': 'mse',
'optimizer': 'Adam',
'activation': 'elu',
'batch_size':1
}