AIDP Tool

A script for getting predictions from the AIDP model

Installation

Running

pip install -e .

in the aidp directory should make the aidp command available to use anywhere.

Usage

aidp --help

for a full list of options and argument descriptions.

Train

Pass train as the second parameter to aidp to train a new model based on the data in the excel sheet passed as the third parameter. E.g,

aidp train /path/to/input_file.xlsx [--model_key='default']

The input excel sheet must adhere pretty closely to the format of the sample excel sheet in ./tests/resources/text.xlsx.

You can optionally provide a model key (using --model_key=<key>) when making a call to train the models. This will save all models generated during this session to /resources/models/<key>/ folder. If no model key is provided, a timestamp will be used.

Predict

Pass predict as the second parameter to aidp to make predictions using the model specified by the model_key parameter based on the data in the excel sheet passed as the third parameter. E.g,

aidp predict input_file.xlsx [--model_key='default']

Again,the input excel sheet must adhere pretty closely to the format of the sample excel sheet in ./tests/resources/text.xlsx. However, no GroupIds are required for predictions.

You can optionally provide a model key (using --model_key=<key>) when making a call to predict. This will use the models saved in /resources/models/<key>/ folder for predictions. If no model key is provided, default is used.