Session Consolidation

Produce a single, clean .py file from a Databricks REPL session by reading session.json and the .cmd.py files.

Workflow

1. Read session.json — the steps array contains the ordered list of steps with status and command file paths.
2. Read each .cmd.py file — in step order, skipping failed steps (only successful steps survive).
3. Strip REPL boilerplate — remove or convert REPL-specific calls (see Boilerplate Rules).
4. Deduplicate — if a step was retried after an error, only keep the final successful version.
5. Resolve imports — collect all imports from across cells and deduplicate them at the top of the file.
6. Write the output — a single .py file with a clear structure.

Output Structure