bridgeAI-regression-model-training

Model training

1. The data used is available here. Download the csv file and update the path to the csv file in the config.yaml file in data.raw_data or in the environment variable DATA_PATH
2. Update the python environment in .env file
3. Install poetry if not already installed
4. Install the dependencies using poetry poetry install
5. update the config and model parameters in the config.yaml file
6. Add ./src to the PYTHONPATH - export PYTHONPATH="${PYTHONPATH}:./src"
7. Run python src/main.py or poetry run python src/main.py

Model training - using docker

1. Build the docker image - docker build -t regression .
2. Bring up the dependencies by using docker compose up -d
3. Run the container with the correct DATA_PATH and MLFLOW_TRACKING_URI as environment variables. (Refer to the following Environment Variables table for complete list)
   docker run -e DATA_PATH=/app/artefacts/HousingData.csv -e MLFLOW_TRACKING_URI=http://host.docker.internal:5000 -v ./artefacts:/app/artefacts --rm regression

Environment Variables

The following environment variables can be set to configure the training:

Variable	Default Value	Description
DATA_PATH	./artefacts/HousingData.csv	File path to the raw data CSV data used for training
CONFIG_PATH	./config.yaml	File path to the model training and other configuration file
LOG_LEVEL	INFO	The logging level for the application. Valid values are DEBUG, INFO, WARNING, ERROR, and CRITICAL.
MLFLOW_TRACKING_URI	http://localhost:5000	MLFlow tracking URI. Use http://host.docker.internal:5000 if the MLFlow is running within docker container.
GITHUB_USERNAME	None	Githuib username. This is needed to pull the data form the dvc repo.
GITHUB_PASSWORD	None	Githuib token. This is needed to pull the data form the dvc repo.
DVC_REMOTE	s3://artifacts	Dvc remote
DVC_REMOTE_NAME	regression-model-remote	Dvc remote name.
DVC_ENDPOINT_URL	http://minio	The URL endpoint for the DVC storage backend. This is typically the URL of an S3-compatible service, such as MinIO, used to store and manage datasets and model files.
AWS_DEFAULT_REGION	eu-west-2	The dvc remote s3 bucket region
DVC_ACCESS_KEY_ID	None	Access key id for dvc remote. Optional. Not needed if using IAM based access for dvc remote.
DVC_SECRET_ACCESS_KEY	None	secret access key for dvc remote. Optional. Not needed if using IAM based access for dvc remote.
DEPLOY_AS_CODE	False	Should manual intervention or evaluation is needed to register the trained model or not.
DEPLOY_MODEL_NAME	house_price_prediction	The name with which the model will be registered. This name and alias will be used for deployment.
DEPLOY_MODEL_ALIAS	champion	The alias to be added to the model. This alias along with the name will be used for deployment.

Running the tests

Ensure that you have the project requirements already set up by following the Model training instructions

• Ensure pytest is installed. poetry install will install it as a dev dependency.
• • For integration tests, set up the dependencies (MLFlow) by running, docker-compose up -d
• Run the tests with poetry run pytest ./tests