IlyaAndr/WarsawGTFS

Creates GTFS feed from ZTM Warsaw data

WarsawGTFS

Description

Creates GTFS data feed for Warsaw. Data comes from ZTM Warszawa FTP server and optionally mkuran.pl website.

Fetures

1. Line colors
2. Calendar exceptions
3. Trip headsigns and On-Request stops
4. Added town names to stop_name
5. Merging railway stops into one
6. Geting railway platforms from external gist
7. Geting missing stop positions from external gist
8. Inserting metro schedules from mkuran.pl
9. Realtime data
10. Shapes generator: Buses based on OSM Data, Rail/Tram based on my own graph.

Static GTFS script

First Launch

First of all you need Python3 and several modules, included in requirements.txt, so run pip3 intall -r requirements.txt.

Then start the script with python3 warsawgtfs.py

Configuration

Run python3 warsawgts.py -h to see all possible options with their descriptions. All of those are optional.

Creating GTFS

Run python3 warsawgtfs.py with desired command line options. After some time (up to 1 min, or 15 mins with the --shapes option turned on) the gtfs.zip file should be created.

Produced GTFS feed has additional columns not included in standard GTFS specification:

• original_stop_id in stop_times.txt - WarsawGTFS changes some stop_ids (especially for railway stops and xxxx8x virtual stops), so this column contains original stop_id as referenced in the ZTM file,
• is_data_source in attributions.txt - to indicate that this attribution represents entity that provides data,
• platform_code in stops.txt - A platform identifier for (most) railway stops (from Google Transit extensions),
• stop_IBNR and stop_PKPPLK in stops.txt - railway station ids shared by rail operators in Poland.
• exceptional in trips.txt - Value 1 indicates an unusual trip which does not follow common line's route (e.g. trips to depot) (from Google Transit extensions).

Realtime GTFS script

The warsawgtfs_realtime.py contains three realtime functions.

All reailtime data is created in gtfs-rt/ directory. By default only the binary protobuf file is created. Outputing human-readable representation can be done by adding the --readable flag to the script. An additional JSON file is added when the script sees flag --json.

• Alerts (option -a / --alerts):
  Arguments:

  • (optional) --gtfs-file PATH_OR_URL - Location of GTFS feed to fetch a list of valid routes. Defaults to https://mkuran.pl/gtfs/warsaw.zip

• Brigades (option -b / --brigades):
  Arguments:

  • --key APIKEY / -k APIKEY - The apikey to https://api.um.warszawa.pl,
  • (optional) --gtfs-file PATH_OR_URL - Location of GTFS feed to base brigades on. Defaults to https://mkuran.pl/gtfs/warsaw.zip

  Notes:

  • Always outputs only a json file,
  • Data is valid only on the date of creation - this process has to be run every day.

• Positions (option -p / --position):
  Arguments:

  • --key APIKEY / -k APIKEY - The apikey to https://api.um.warszawa.pl,
  • (optional) --brigades-file PATH_OR_URL - Location of the brigades.json file. Defaults to https://mkuran.pl/gtfs/warsaw/brigades.json

  Notes:

  • This script assumes that all trips are running on time.
  • If you wish to update positions in a loop, please from src import Realtime and call Realtime.positions() on your own. It returns matched vehicles, which then can be provided again to Realtime.positions() for a slightly better accuracy of matching trip_ids (instead of assuming everything runs on time).

  Arguments of Realtime.positions():

  • apikey: Apikey to https://api.um.warszawa.pl
  • brigades: Path/URL to brigades.json
  • previous: Dictionary of known vehicles, as returned by this function. If calling for the first time provide an empty dict, {}.

License

WarsawGTFS is provided under the MIT license. Please take a look at the license.md file.