Methodology

Reference pages covering the parts of pyGTFSHandler that involve a non-obvious algorithm or design decision:

• Calendar handling — how calendar.txt/calendar_dates.txt are resolved into "active service_ids on date X", and how multi-day/ overnight trips are handled without duplicating rows.
• Frequency-based trip expansion — the headway math that turns frequencies.txt rows into concrete stop_times departures.
• Direction & headway methodology — how geometric direction_id is derived from shape bearings, and the RMS-style headway formula used to summarize departure regularity.
• Edge & filtering logic — the distance/geometry math behind edge (stop-to-stop segment) construction and the feed-wide date/time/route/AOI filter pipeline.