Installation#
Package installation#
Predictable can be installed using pip:
Creating a project#
Once installed, the predictable CLI can be used to create a new project:
This will scaffold a new project in the my_project directory. The structure of this project will look similar to:
my_project/
├── __init__.py
├── data/
│ └── modelpoints.csv
├── model.py
├── output/
├── run.py
└── tables/
Project structure#
| File | Description |
|---|---|
__init__.py |
This file allows us to use python modules |
data/modelpoints.csv |
Example modelpoint file used (as input) when the model is run |
model.py |
Contains the model logic |
output/ |
Stores model results and artifacts |
run.py |
Used to orchestrate model runs |
tables/ |
Stores table inputs used by the model |