remove_page_from_job
Permanently remove a page and all its data from a cloning job
Overview
Hard-delete a single page from a cloning job when it no longer exists on the source site. Removes all child rows (issues, comparison results, lens data, junction table entries) and performs best-effort cleanup of any files stored in B2 (HTML snapshots, screenshots, videos, diff images). The operation is permanent and irreversible. Use discover_all_pages with its built-in discover_then_delete mode for bulk removal.
How It Works
- Looks up the page by source URL across all of the current user's cloning jobs.
- Returns not_found if the URL is not tracked — the call is safe to make speculatively.
- Collects all B2 file keys for the page (HTML, screenshots, videos, diff images) and deletes them on a best-effort basis before opening the DB transaction.
- Opens a single transaction and cascades: issue sub-tables → issues → lens sub-tables → page_compare_lens → page_compare → junction tables (page_stylesheet_page, route_page_a, content_type_field_page, content_type_page) → page.
- Returns removed: true plus page_id, source_url, and job_id on success.
Input Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
source_url |
string |
required | URL of the source (original) page to remove |
What You Get Back
- removed — true on success
- page_id — the database ID of the deleted page
- source_url — the stored source URL of the deleted page
- job_id — the cloning job the page was removed from
- result_status — 'not_found' when the URL is not tracked (no mutation occurs)
Example Use Case
A source page returned 404 on re-compare and is confirmed gone. Call remove_page_from_job with the source URL to hard-delete it and stop it surfacing in get_next_actionable_page.
Tips
The call is idempotent for the not_found path — calling it twice on the same URL is safe.
Job-level tables (shared_page_element, design_system, route, content_type, page_stylesheet) are not affected — only per-page rows and junction table entries are removed.
B2 cleanup is best-effort. If cleanup fails the DB transaction still commits and the page is deleted.
For removing many stale pages at once, use discover_all_pages — it has a built-in discover_then_delete mode that re-crawls and removes pages no longer on the source site.
After removal, run_full_comparison and get_next_actionable_page will naturally exclude the deleted page.