Cdspt 15587 align functionality of totals methods for all extractor subclasses

[scottbw]

📌 Summary

This change ensures that each Extractor subclass contains:

get_row_count()
get_total(measure:str)

Calling code should now call either:

• get_row_count() if you want a row count
• get_total(measure) if you want to sum by a specific measure
• calling get_total() is discouraged.

The behaviour is consistent across implementations:

• get_row_count() returns the row count
• get_total(measure) returns SUM of measure if measure is not None and that measure exists in dataset.measures
• get_total(measure) raises ValueError when measure is not None and the measure does not exist in dataset.measures
• get_total(measure) returns SUM of the first measure in dataset.measures if measure is None and there are measures in dataset.measures
• get_total(measure) returns the row count if measure is None and there are no measures in dataset.measures

Notes:

1. this is slightly different from the previous behaviour, where if a measure was specified that did not exist, the row_count was returned.
2. 'Find a default measure' logic was previously also in SubsetQueryBuilder. I've removed this - the only place default measures are substituted for None is now in the Extractor layer.

---

🧪 What Changed?

• New feature
• Bug fix
• Refactor
• Documentation only
• CI/CD or tooling
• Other (please explain):

---

✅ Checklist

Code Quality

• [x ] Code is clean, commented, and follows our conventions
• [x ] No hardcoded values or secrets
• [x ] Reusable and modular where possible

Testing

• I have tested this change locally
• Relevant tests have been added/updated
• All existing tests still pass

Documentation

• Code comments or inline explanations added where needed
• README / documentation updated (if applicable)
• Any environment variables added are documented

Security & Deployment

• No sensitive data in code or logs
• No breaking changes introduced
• Deploy scripts / jobs updated (if needed)

Note that there is a change in behaviour for where a non-existent measure is specified for a total.

⚙️ Actions Required After Pulling

• Yes.
• No.

If you checked Yes, please provide instructions below:

---

🧠 Context or Screenshots

---

🚨 Risk & Impact

• No user-facing changes
• Minimal risk — small changes isolated to specific areas
• Medium risk — changes affect multiple areas
• High risk — could impact critical paths or user flows

Notes for reviewers (optional):

[TiffanyCheng27]

• Add a unit test for the Athena get_row_count() function.
• The current unit tests cover HyperFile and DataExtractor. Add unit tests for StreamingDataExtractor to cover get_row_count() and get_total().
• (Optional) Additionally, since everything works fine for now, this can be change in the future. It might be worth considering making the get_total() function more consistent across extractors. Currently some extractors delegate internally to __get_total__() while others don't use that sub-layer at all. And there are duplication in StreamingDataExtractor and AthenaDataExtractor. We can think about standardising the pattern so all extractors follow the same structure.

[scottbw]

I've added the missing unit tests and refactored so they all call get_total() internal metho