kedro-templar 0.0.1

No project description provided

Kedro Templar

About

A plugin build for Kedro to support templating configuration files. Using this plugin you can specify a templates for your configuration files such as catalog.yaml, paramters.yaml and fill them automatically using one file with specified variables. Using this approach you can easily parametrize you pipeline from single point.

Usage examples:

• you can store multiple simultaneous configurations and switch between them on demand.
• change prefixes in catalog.yaml to reuse the same pipeline for different data,

Install

Plugin can be easily installed using pypi repository.

pip install kedro-templar

Templates

kedro-templar is using jinja2 templating engine for rendering files.

Example

catalog.yaml

Template created for catalog.yaml file.

sample_data:
  type: pandas.CSVDataSet
  save_args:
    index: True
  filepath: s3://your_bucket/{{run_name}}/{{another_subname}}/data/sample_data.csv

definition.yaml

Your file that contains all variables used for templating.

run_name: run_1
another_subname: subfolder

Commands

Currently, this plugin supports 3 commands:

apply

A core logic of this plugin. This command is used to create config files based on provided templates and a file with variables.

You can setup a default values using environment variables:

• TEMPLAR_TEMPLATES_DIR - a default directory for your templates
• TEMPLAR_OUTPUT_DIR - a default directory where templates will be rendered

If no value is specified explicitly and env var is empty, the commands will use default value specified in the code.

kedro templar apply -c definition.yaml

download

A helper function used to download a config file from a given S3 path

kedro templar download -i <S3_PATH>/definition.yaml -o definition.yaml

upload

A helper function used to upload a config file to a given S3 path

kedro templar upload -i definition.yaml -o <S3_PATH>/definition.yaml

Contact

Plugin was created by the Data Science team from WebInterpret.