Streaming GTFS and GTFS-RT parser for node.

npm install gtfs-stream --save

import { plain } from 'gtfs-stream'

request.get('https://developers.google.com/transit/gtfs/examples/sample-feed.zip') // or any other way of getting the data stream
  .pipe(plain())
  .on('data', (entity) => {
    console.log(entity)
  })

Data events emitted from the GTFS parse stream have the following shape:

• type (String, feed_info, agency, stop, route, trip, stop_time, calendar, calendar_date, fare_attribute, fare_rule, shape, frequency, transfer)
• data (Object)
  • See https://developers.google.com/transit/gtfs/examples/gtfs-feed for the available attributes of each type

If you want the raw rows, you can pass { raw: true } to this function to skip type inference. This library also provides TypeScript type definitions for all GTFS entities.

The base GTFS format is cumbersome to work with, so the enhanced parser will do a little extra work piecing things together to make it more usable. This parser will use more memory than the base parser since it needs to collect rows that need formatting while it waits for other to stream in.

import { enhanced } from 'gtfs-stream'

request.get('https://developers.google.com/transit/gtfs/examples/sample-feed.zip') // or any other way of getting the data stream
  .pipe(enhanced())
  .on('data', (entity) => {
    console.log(entity)
  })

• type (String, feed_info, agency, stop, route, trip, calendar, calendar_date, fare_attribute, fare_rule, shape, frequency, transfer)
• data (Object)
  • See https://developers.google.com/transit/gtfs/examples/gtfs-feed for the available attributes of each type

Differences from the base parser:

• trip types have a new path attribute that is a GeoJSON LineString, and shape_id is removed
• stop types have a new schedule attribute that is a collection of stop times
• route types have a human-readable route_type
• stop types have a human-readable location_type and wheelchair_boarding
• No shape or stop_time types, since they are collected into their relevant entries

import { rt } from 'gtfs-stream'

request.get('http://datamine.mta.info/mta_esi.php') // or any other way of getting the data stream
  .pipe(rt())
  .on('data', (entity) => {
    console.log(entity)
  })

Data events emitted from the GTFS Realtime parse stream have the following shape:

• id
• is_deleted
• trip_update
  • timestamp
  • delay
  • vehicle
  • stop_time_update
  • trip
    • trip_id
    • route_id
    • direction_id
    • start_time
    • start_date
    • schedule_relationship
• vehicle
  • position
  • current_stop_sequence
  • stop_id
  • current_status
  • timestamp
  • congestion_level
  • occupancy_status
  • trip
    • trip_id
    • route_id
    • direction_id
    • start_time
    • start_date
    • schedule_relationship
• alert
  • active_period
  • informed_entity
  • cause
  • effect
  • url
  • header_text
  • description_text

Only one of either trip_update, vehicle, or alert will be present in any given event.