debiai-data-provider 1.1.2

Start your own Data-provider from simple Python

DebiAI Data Provider Python module

This DebiAI Data Provider Python module allows you to easily deploy your own data-provider through the data-provider API.

A data-provider allows you to provide data to DebiAI so that no duplication of data is needed.

DebiAI Data-providers documentation

Getting started

Installation

Install debiai_data_provider with pip:

pip install debiai_data_provider

Usage

Create a new Python file, for example my_data_provider.py and copy the content of the simple_project.py Data-provider.

Run the Python file and your project is now available through the DebiAI Data Provider API!

This way of creating a data-provider is the most generic one, you can customize it to support any data-sources.

We also provide a higher level of abstraction to create data-providers:

• ParquetDataProvider is a data-provider that provides data from a parquet file. It is a simple way to create a data-provider with a parquet file as input.

Plug-in your data-provider with DebiAI

To link your data-provider with DebiAI, you can follow our Creation of a data provider guide

Examples

Find out how to use the DebiAI Data Provider Python module in the examples folder:

Generic Data-providers

• simple_project.py shows how to create a simple data-provider with a project
• project_with_results.py shows how to create a data-provider with a project that also provides model results
• parquet_data_provider.py shows how to create a data-provider that provides data from a parquet file

Data-providers with a higher level of abstraction

• easy_parquet_data_provider.py does the same as the previous example but with a higher implementation level, using the ParquetDataProvider class.

Testing

To run the tests, install the dependencies and run the tests with pytest:

pip install -r requirements.txt
pip install pytest pytest-cov
python -m pytest --cov=debiai_data_provider

Tested versions:

• Python 3.9
• Python 3.10
• Python 3.11
• Python 3.12

Roadmap

• Publish to Pypi
• Provide project data
  • Provide project metadata
  • Provide project samples
  • Provide project models & model results
  • Provide project selections
• Make available project interactions
  • Project deletion
  • Model deletion
  • Selection creation
  • Selection deletion
• High level data-providers
  • CSV data-provider
  • Json data-provider
• Start DebiAI along with the data-provider
• Create a welcome page that shows the data-provider status and projects
• LLM improved data-provider for auto configuration

---

DebiAI is developed by And is integrated in

---