A fast, forgiving GTFS reader built on pandas DataFrames
Clone or download
invisiblefunnel Merge pull request #46 from remix/write-feeds-concurrently
Write feeds concurrently to reduce the overall time
Latest commit 8b67dae Sep 3, 2018



Partridge is python library for working with GTFS feeds using pandas DataFrames.

The implementation of Partridge is heavily influenced by our experience at Remix ingesting, analyzing, and debugging thousands of GTFS feeds from hundreds of agencies.

At the core of Partridge is a dependency graph rooted at trips.txt. Disconnected data is pruned away according to this graph when reading the contents of a feed. The root node can optionally be filtered to create a view of the feed specific to your needs. It's most common to filter a feed down to specific dates (service_id), routes (route_id), or both.

dependency graph


The design of Partridge is guided by the following principles:

As much as possible

  • Favor speed
  • Allow for extension
  • Succeed lazily on expensive paths
  • Fail eagerly on inexpensive paths

As little as possible

  • Do anything other than efficiently read GTFS files into DataFrames
  • Take an opinion on the GTFS spec


Reading a feed

import datetime
import partridge as ptg

path = 'path/to/sfmta-2017-08-22.zip'

service_ids_by_date = ptg.read_service_ids_by_date(path)

date = datetime.date(2017, 9, 25)
service_ids = service_ids_by_date[date]

feed = ptg.feed(path, view={
    'trips.txt': {
        'service_id': service_ids,
        'route_id': '12300',

assert service_ids == set(feed.trips.service_id)

#  88

#  route_id agency_id route_short_name route_long_name route_desc  route_type  \
#     12300     SFMTA               18     46TH AVENUE        NaN           3
#  route_url route_color route_text_color
#        NaN         NaN              NaN

Extracting a new feed

import partridge as ptg

inpath = 'gtfs.zip'
outpath = 'gtfs-slim.zip'

date, service_ids = ptg.read_busiest_date(inpath)

ptg.writers.extract_feed(inpath, outpath, {'trips.txt': {'service_id': service_ids}})

assert service_ids == set(ptg.feed(outpath).trips.service_id)


  • Surprisingly fast :)
  • Load only what you need into memory
  • Built-in support for resolving service dates
  • Easily extended to support fields and files outside the official spec (TODO: document this)
  • Handle nested folders and bad data in zips
  • Predictable type conversions


pip install partridge

Thank You

I hope you find this library useful. If you have suggestions for improving Partridge, please open an issue on GitHub.