Growth Nirvana Transform Authoring

Growth Nirvana Transform Authoring

Purpose

Produce reliable transformation SQL drafts with evidence-based model alignment and validation loops.

Workflow

1. Resolve dataset by displayName and inspect relevant model folders:
   • models/<slug>/query.sql
   • models/<slug>/model.json
   • models/<slug>/data_transformation.json
2. Identify target variant context:
   • combined_report
   • combined_report_ff
   • combined_report_catalog
   • other custom marts
3. Draft SQL in style consistent with neighboring models.
4. Validate with create_dry_run(context=TransformationModel).
5. Poll with get_dry_run to terminal.
6. If errors, revise and rerun dry run until success.
7. Return upload-ready handoff details.

Guardrails

• Never skip dry run.
• Never execute query unless user explicitly asks.
• Always note assumptions about source models and column semantics.
• If variant-specific remapping exists, preserve it unless user asks to change it.

Response Contract

• workflow: transform-authoring@v1
• dataset + target model slug
• model variant context
• draft summary
• dry run id/status/errors
• final SQL path/content summary
• upload handoff notes

Resources

• Examples: examples.md
• Troubleshooting: troubleshooting.md