Skip to content

Latest commit

 

History

History
39 lines (27 loc) · 966 Bytes

File metadata and controls

39 lines (27 loc) · 966 Bytes

Prophet Service

A companion service that provides time-series forecasting for Tractor-Beam using Facebook's Prophet library.

Features

  • Demand forecasting with external regressors (MTI, inflation)
  • Supplier performance prediction (quality ratings and lead time reliability)
  • Visualization of forecasts with confidence intervals

Installation

  1. Clone the repository
  2. Set up a Python virtual environment:
    python -m venv venv
    source venv/bin/activate  # On Windows: venv\Scripts\activate
  3. Install dependencies:
    pip install -r requirements.txt

Usage

Start the service:

python app.py

The service will try to run on port 5001, with fallbacks to 5002 and 5000 if needed.

API Endpoints

  • POST /predict/demand - Generate demand forecasts
  • POST /predict/supplier-performance - Predict supplier performance metrics

See the API documentation for request/response formats and examples.