search

capitalone / rubicon-ml

Capture all information throughout your model's development in a reproducible way and tie results directly to the model code!
https://capitalone.github.io/rubicon-ml/
Apache License 2.0
129 stars 34 forks source link

Dataframe plot #465

Closed jeh362 closed 3 months ago

jeh362 commented 3 months ago

closes: #60

Description Issue: Currently we are trying to build an extension for serving visualizations through a Juptyer lab extension. To do this we want to include visualization information in the catalogs. Initially, we created a base visualization data source similar to the existing one for projects and experiments. Now we need to create a concrete implementation for Dataframe Plots.

Solution: For this issue, we added a new DataframePlotDataSource that is going to extend VizDataSourceMixin. We will read in the catalog data and metadata from the passed arguments . Then _get_schema creates a Dataframe Plot visualization and sets it as the visualization object attribute.

Changes

  1. Updated viz.py within the intake_rubicon folder. This file holds the code for the new DataframePlotDataSource.
  2. Updated the __init__ function within the intake_rubicon folder to now be able to include DataframePlotDataSource for module importing purposes.
  3. Updated rubicon_ml/viz/publish.py to take in a DataframePlot as a visualization object
  4. Updated the setup.cfg file to include the DataframePlotDataSource as an Intake driver.
  5. Updated test_viz.py to include a DataframePlotDataSource test method as well.
CLAassistant commented 3 months ago

CLA assistant check
All committers have signed the CLA.

ryanSoley commented 3 months ago

one last thing - can we add a publish test like how tests/unit/intake_rubicon/test_publish.py was updated in #464