Rename Non-Compliant Target Variables to Follow ODR-007 Naming Convention

[Polichinel]

Connected to issue 17 in views-pipeline-core and ODR-007

Description

To ensure full compliance with ODR-007: Target Variable Naming Convention, all existing target variables must be renamed according to the prefix rules (lr_, ln_, lx_, by_).

Currently, some features in the dataset do not conform to this naming standard, which can cause:

• Silent errors in ensemble reconciliation.
• Incorrect evaluation and feature interpretation.
• Pipeline inconsistencies due to naming mismatches.

Steps to Complete

1. Identify all non-compliant target variable names in existing datasets, querysets, and model configurations.
2. Rename each variable according to the appropriate transformation prefix:
   • lr_ → Linear (untransformed)
   • ln_ → Natural Log (log(1 + target))
   • lx_ → Natural Log (log(exp(-100) + target))
   • by_ → Binary (0/1 only)
3. Update all references if needed to the renamed features in:
   • Model querysets (config_queryset)
   • Metadata configurations (config_meta)
   • Documentation, datasets, and stored model configurations
4. Verify correctness by running pipeline tests to ensure that the renamed variables work as expected.

Expected Outcome

• All target variables follow the correct prefix convention.
• No pipeline errors occur due to incorrect naming.
• Evaluation and reconciliation algorithms correctly interpret target variables.

Checklist

• Identify all non-compliant target variables.
• Rename variables to match ODR-007 naming rules.
• Update references in config_queryset and config_meta.
• Update any relevant documentation.
• Validate changes by running tests.

Branch

• Work on a new branch: rename_target_vars