Add the top-level runScrape() function that coordinates the full scraper
pipeline: fetch HTML (with retry), parse units, convert data types, upsert
units, insert prices, mark stale units, and update daily summary.
Features:
- dryRun mode skips all database writes while still parsing/validating
- htmlContent parameter allows injecting HTML directly (bypasses fetch)
- Calculates newUnitsCount and rentedUnitsCount by diffing against prior state
- Records every run to scraper_runs history (success or failure)
- Structured logging with jobId correlation throughout the pipeline
- Graceful error handling at each pipeline stage
Includes 15 tests covering full workflow, dry run, error handling,
trigger types, new/rented unit calculation, and empty HTML edge case.
