Feat/realtime reproducible fixtures regeneration mbta tests subissue 17

[jloria2703]

Sub-issue #17 — Reproducible Sample Data

This sub-issue focused on providing reproducible GTFS-Realtime sample data, and the work completed fully satisfies the requirements.
Two complementary testing workflows were implemented: a deterministic fixture generator and a live ingestion pipeline connected to the MBTA realtime server.

The deterministic module produces small, self-contained fixtures following GTFS-Realtime v2.0. These files are stored in the repository, regenerate consistently using fixed seeds, and serve as stable inputs for both automated tests and project documentation.
In parallel, the MBTA streaming workflow validates the system with real external data, ensuring correct decoding, storage, and structural consistency under real operating conditions.

Together, these components deliver what the sub-issue requires:

• Small fixtures included in the repository
• Scripts to regenerate them whenever needed
• Outputs used directly in tests and technical docs

Execution Commands

# Subissue #17 — Deterministic fixtures
python -m gtfs.scripts.regenerate_fixtures

# Subissue #17 — MBTA streaming and validation
python -m gtfs.scripts.stream_mbta_feeds
sqlite3 db.sqlite3
.tables
SELECT COUNT(*) FROM gtfs_tripupdate;
SELECT * FROM gtfs_vehicleposition LIMIT 3;
.exit