Pipeline output compatibility

[jonathanstelman]

Epic 6: Pipeline Integration

Ensure the pipeline integrates cleanly with the new monorepo structure, generates stable resort UUIDs, and exposes metadata to the backend.

Tasks

• Generate `resort_id` UUIDs for all resorts on first pipeline run; persist mapping in `data/resort_id_map.csv`
• On subsequent runs, look up existing UUIDs from the mapping and only assign new UUIDs to new resorts — never reassign
• Ensure `data/resorts.csv` (with `resort_id` column) is committed to the repo and readable by the backend at startup
• Write `data/pipeline_metadata.json` on each run (already implemented in `pipeline.py`); ensure backend exposes `last_pipeline_run` from this file in `GET /meta`
• Document how to run the pipeline from the new `/pipeline` directory in the monorepo

Data delivery approach

`data/resorts.csv` is committed to the repo. The backend reads it at startup. The scheduled pipeline (Issue #77) commits the updated CSV back to the repo, and a new deploy picks it up. This keeps the backend stateless with no external storage dependency.

Acceptance Criteria

• Pipeline generates UUIDs on first run and writes `data/resort_id_map.csv`
• Re-running the pipeline preserves existing UUIDs for existing resorts
• `data/resorts.csv` includes `resort_id` column after pipeline runs
• `GET /meta` response includes `last_pipeline_run` timestamp from `pipeline_metadata.json`
• README (or `/pipeline/README.md`) documents how to run the pipeline in the new structure