Models holder class

class models.modelsHolder.ModelsHolderClass(models_directory)

Class that loads and holds in memory all models for further predictions.

Attributes:

__configParser : ConfigParser()

Instance of config parser for loading configs that come with models.

__config : dict

Dictionary that contains entries of configs of every model. {‘model_no_1’ : {‘arg’:1, ‘arg2’:2}}

__models : list

List of strings with model names to load.

__model_instance_holder : dict

A dictionary containing instances of per-model specific classes (inheriting from the base ModelController class)

Methods

sendRequest(model_name, data)	Invokes the ModelController’s feed method that returns models prediction.
getAvailableModels()	Returns loaded and available models.
getModelsConfigs()	Get all models configs.

getAvailableModels()

Returns loaded and available models.

getModelsConfigs()

Get all models configs.

Returns:

dict

sendRequest(model_name, data)

Invokes the ModelController’s feed method that returns models prediction.

Parameters:

model_name : str

Name of the model to use.

data : MultiDict or str

Data sent by the client to pass to the ModelController.

Returns:

str

Printable string ready to view / send to client. Prediction results or error message.