test-power-generation

This repo includes files to build and deploy models as described in the Domino Get Started with Python documentation.

I have made modifications to the tutorial so that it runs on a development version of Domino 5.3 and its Domino Standard Environment Py3.9 R4.2.

It includes the following files:

1. Forecast_Power_Generation.ipynb -- This is a Jupyter notebook to create a power generation forecast model and export it to a file from Step 5: Develop your Model.
2. requirements.txt -- This is a modified version of the requirements.txt file outlined in Step 7: Package Setup.
3. Forecast_Power_Generation_Dynamic_30days.ipynb -- This is an adapted notebook that downloads the last 30 days worth of power data and uses it to generate and export the model from Step 7: Scheduled Reports.
4. Forecast_Power_Generation_Launcher.ipynb -- This is an adapted notebook that exposes the start date and fuel types as parameters, downloads the data, and it to generate and export the model from Step 7: Launchers.
5. forecast_launcher.sh -- This is a wrapper script for a launcher to start the Forecast_Power_Generation_Launcher.ipynb notebook and substitute the provided parameters using papermill from Step 7: Launchers. It is modified to set up a python venv, install the required papermill and prophet packages, and run notebook using the venv papermill. I had to do this to get the papermill command to run.
6. Dockerfile -- This Dockerfile contains the line that needs to be added into Dockerfile Instructions when creating a new compute environment to run the Model API in Step 7: Model APIs.
7. forecast_predictor.py -- This is the predictor python script that the Model API from Step 7: Model APIs uses to load the serialized model and runs a prediction for the provided year, month, and day inputs.