init

[AnushaGanapathi]

What

This PR introduces an improvement to the KoboToolbox connector to support labelled outputs and labelled questions instead of only returning raw variable names.

Previously, the connector returned responses with internal field names (e.g., ${q1}, ${select_one_abc}), which made downstream transformations harder and reduced readability—especially for non-technical users.

With this enhancement, users can now retrieve:

• Human-readable question labels
• Labelled response options (in English)

This makes the data significantly easier to understand, transform, and use in downstream pipelines.

Additionally, minor changes were made to the existing KoboToolbox connector to resolve local execution errors. These fixes are small, non-functional improvements and can be excluded from review if deemed unnecessary.

How

• Extended the KoboToolbox connector to fetch and map:
  • Question labels
  • Choice labels (for select-type questions)
• Implemented a transformation layer that:
  • Replaces raw field names with their corresponding labels
  • Maps coded responses to their English-labelled values
• Ensured backward compatibility so that:
  • Existing workflows continue to function without breaking changes
• Applied minor fixes to resolve local connector execution issues (non-breaking changes).

Review guide

1. source.py – Label extraction and response mapping logic

User Impact

Positive Impact

• Outputs are now human-readable and English-labelled.
• Simplifies downstream transformations and modelling.
• Reduces the need for manual mapping during data preparation.
• Improves accessibility for non-technical users.

Negative Side Effects

• Output column names may differ if users were previously relying strictly on raw variable names (though backward compatibility is maintained).

Can this PR be safely reverted and rolled back?

• YES 💚
• NO ❌