Weird message when trying to rerun map_deploy_to_devel pipeline

[pombredanne]

When I try to add a new map_deploy_to_devel on a project that already had a run of map_deploy_to_devel I get

duplicate key value violates unique constraint "scanpipe_codebaseresource_unique_path_within_project"
DETAIL:  Key (project_id, path)=(e621cfb0-1477-4daf-8c01-2948964196be, to) already exists.
....

[tdruez]

@pombredanne Re-running any duplicated pipeline following an existing run will generally be a source of issues.
You should use the "reset" feature for a clean re-run.

[reehassan]

Rerunning map_deploy_to_develop on a project that already has resources blows up on the (project, path) unique constraint for CodebaseResource. Second run tries to insert to/from again, Postgres rejects it, raw IntegrityError hits the UI.

Root cause is in scanpipe/pipes/init.py, collect_and_create_codebase_resources(). Since rerun without reset isn't really supported anyway (per the comment above), wrapped the bulk_create there in try/except and re-raise with a message pointing at Reset instead of the bare trace:

except IntegrityError as e:
    raise IntegrityError(
        "Codebase resources already exist for this project. "
        "Reset the project before re-running."
    ) from e

This second SS is the after. Traceback's still there below it for debugging, just not the first thing you see.

Added a test in scanpipe/tests/pipes/test_pipes.py, right next to the existing test_scanpipe_pipes_collect_and_create_codebase_resources, that reruns the function and checks for the new message. Ran the full suite locally, passes except one qcow2 extraction test that's already broken on main too (checked with git stash).