Execution Workflow

Run the notebook in order.

Core Workflow

1. Environment Setup

   • Imports dependencies.
   • Verifies the OpenAI Agents SDK version.
   • Reads OPENAI_API_KEY.
   • Configures tracing and model selection.

2. Input

   • Defines CHANGE_TEXT.
   • This is the only required business input for the core workflow.

3. Optional PDF RAG Context

   • Leave PDF_PATH = None to run without retrieval.
   • Set PDF_PATH to a local PDF to enable File Search context for impact analysis.

4. Stages 1-2

   • Parse the change request.
   • Analyze impact.
   • Optionally use File Search during impact analysis.

5. Stages 1-2 Guardrails

   • Confirm parse output is well-formed.
   • Confirm impact output includes the target.
   • Confirm impacted objects contain required fields.

6. Stages 3-4

   • Generate an execution plan.
   • Generate SQL across landing, staging, core, and mart layers.

7. Stages 3-4 Guardrails

   • Confirm plan sections are populated.
   • Confirm data type propagation.
   • Confirm nullability behavior matches the request.

8. Stage 5 SQL Sanity Checks

   • Check for empty SQL.
   • Check expected target table and columns.
   • Check required SQL actions implied by the request.

9. Final Bundle and Artifact

   • Assemble the full output bundle.
   • Save it as JSON.
   • Verify the artifact round-trips successfully.

Optional Eval Workflow

10. Pre-Promptfoo Checks

    • Confirm the notebook state is ready for evals.

11. Promptfoo Runtime Generation

    • Create a reusable SchemaFlow core module.
    • Write a Promptfoo provider.
    • Write a Promptfoo assertion file.
    • Generate Promptfoo test cases and config.

12. Promptfoo Eval Execution

    • Run parse-only and full-flow evals.
    • Save timestamped JSON and HTML reports.
    • Refresh schemaflow_cookbook_eval_latest.* aliases.