Recover select.group_by value counts from #1114

[ebhills]

Summary

Recovers the select.group_by counts aggregation from #1114 onto current main. The original PR was accidentally merged into the legacy dev branch; this PR carries only that logical two-file feature and does not merge dev.

What changed

• Adds a counts aggregation to select.group_by.
• Returns a dictionary mapping each distinct value to its frequency.
• Converts dictionary keys to JSON-safe strings.
• Represents missing values as "null" and booleans as lowercase "true" / "false".
• Preserves existing per-column output renaming and auto_rename_columns behavior.

Example:

wrangles:
  - select.group_by:
      by: Group
      counts:
        - Selection: Selection Counts
      auto_rename_columns: false

Compatibility and risk

This is an additive aggregation option. Existing select.group_by operations and naming behavior are unchanged.

Validation

• Complete TestGroupBy class: 30 passed.
• wrangles/recipe_wrangles/select.py compiles successfully.
• git diff --check origin/main...HEAD passes.
• Branch is exactly one commit ahead and zero commits behind current origin/main.

The full repository suite is left to GitHub Actions.