fix: cartesian product with gtfsdatasets in query

[cka-y]

Summary
Closes #1221.

Details

Previously, the query used to filter feeds by bounding box did not properly join GtfsDataset and GtfsFeed, resulting in a cartesian product. This caused the ST_Covers(...) spatial filter to be applied incorrectly or not at all.

The query has been updated to explicitly join GtfsDataset on feed_id when applying bounding box filters, preventing duplicate table references and ensuring the spatial filtering works as intended.

Validation

The following request now returns the correct GTFS feed for the Lisbon area (as expected):

GET /v1/gtfs_feeds?limit=10&offset=0&bounding_filter_method=completely_enclosed&is_official=true&dataset_latitudes=38.4,39.0&dataset_longitudes=-8.8,-9.5

Example response from dev:

Sample output

[
  {
    "id": "mdb-1272",
    "data_type": "gtfs",
    ...
    "provider": "Cascais Próxima, E.M., S.A.",
    ...
    "locations": [
      { "country_code": "PT", "municipality": "Cascais e Estoril", ... },
      ...
    ],
    "latest_dataset": {
      "id": "mdb-1272-202502251645",
      "bounding_box": {
        "minimum_latitude": 38.679561,
        "maximum_latitude": 38.763594,
        "minimum_longitude": -9.484887,
        "maximum_longitude": -9.313059
      },
      ...
    }
  }
]

✅ Verified locally pointing to dev.

Please make sure these boxes are checked before submitting your pull request - thanks!

• Run the unit tests with ./scripts/api-tests.sh to make sure you didn't break anything
• Add or update any needed documentation to the repo
• Format the title like "feat: [new feature short description]". Title must follow the Conventional Commit Specification(https://www.conventionalcommits.org/en/v1.0.0/).
• Linked all relevant issues
• Include screenshot(s) showing how this pull request works and fixes the issue(s)

[davidgamez]

LGTM!