Feat/realtime validation json helpers and serialization tests s14 s15

[jloria2703]

Completion of Sub-issue #14 and Implementation Work on Sub-issue #15

Work continued on the unit tests for Sub-issue #14, which were originally embedded inside realtime.py.
Following Django conventions, the testing logic was moved into a dedicated tests/ directory, and a new file test_realtime.py was created to centralize all Realtime test cases.

A new manage.py file was introduced, enabling direct execution of tests through Django commands.
To resolve import-path conflicts during this migration, an additional tests/ directory containing only __init__.py was temporarily added to maintain compatibility with existing modules.

The project configuration was updated by modifying settings.py to include the project root (gtfs-django/) in sys.path, ensuring proper resolution of internal imports.
After these changes, test_realtime.py was successfully integrated into the final test structure, completing the environment setup required for GTFS-Realtime unit testing.

The functions get_vehicle_positions(), get_trip_updates(), and get_service_alerts() were rewritten to correctly download, decode, and store GTFS-Realtime feeds retrieved from providers configured in the database.

---

Sub-issue #15 — Timestamp/ID Validation and JSON Export Utilities

Several components were implemented to satisfy the validation and JSON export requirements:

• RealTimeValidator was added to check timestamp monotonicity across consecutive messages, fulfilling the first acceptance criterion.
• JSONExporter was implemented with export_to_json() and export_batch_to_json() to convert FeedMessage objects into GTFS-Realtime-compliant JSON structures, covering the “JSON conversion utilities documented” criterion.
• In the FeedMessage model, the clean() method was added to enforce strictly increasing timestamps and to verify identifier consistency against the Schedule layer.
• A new to_json() method was incorporated into FeedMessage to generate reusable JSON objects compatible with the exporter.
• The Alert model was updated with a JSONField (informed_entity) to better align with the GTFS-Realtime specification.

---

Results

Unit tests were executed successfully using Python’s unittest runner.
A dedicated directory was created to store the generated protobuf binary samples.
Because these files are not directly human-readable, an additional decoding step was implemented: each .pb file is parsed with gtfs_realtime_pb2.FeedMessage() and converted to JSON using json_format.MessageToJson() to verify correctness.

Tests and binary generation were executed with:

python manage.py test tests