Check for existence of a shape_id for each trip

[owades]

Describe the problem

The California Minimum General Transit Feed Specification (GTFS) Guidelines require that schedule feeds, "Include shapes.txt, with valid route shapes for each trip."

While gtfs-validator performs some validation on the shapes in shapes.txt, it does not (to my knowledge) confirm that there exists a shape for each trip in trips.txt.

Describe the new validation rule

If there exists at least one row in trips.txt with a null value in the shape_id column, trigger a "missing_shape_id" notice.

Sample GTFS datasets

No response

Severity

Warning

Additional context

Note that, once a shape_id is referenced in trips.txt, the foreign_key_violation rule will check whether that shape_id actually appears in shapes.txt

[isabelle-dr]

Hello @owades,

If I understand correctly, a valid dataset according to this Cal-ITP's Guidelines would mean

• shapes.txt is required
• every row of trips.txt would need to have shape_id associated

As mentioned in the other issues, we plan to change the validator architecture to allow users to do such changes. What it would look like on your end could be either:

• Making shapes.txt Required instead of Optional and trips.shape_id Required instead of Optional in Cal-ITP's validation set. This way, you'd have the existing missing_required_file and missing_required_field notices triggered.
• Adding two new rules that would only apply to Cal-ITP's validation set, which would trigger new notices: missing_shape_file, and missing_shape_id.

It would be great to have your comments in Google's proposal here.

[isabelle-dr]

Labeling as blocked, let's come back to this issue when we close issue #1067.