Transfer Pattern Journey Planner
Implementation of Hannah Bast's transfer pattern journey planner. This repository does not generate transfer patterns, they need to be created in a pre-processing step.
In addition to the algorithm described in the paper this implementation:
- Checks calendars to ensure services are running on the specified day
- Origins and destinations may be a set of stops
- Interchange time at each station is applied
- Pickup / set down marker of stop times are obeyed
- Overtaken trains are removed
- Transfers (footpaths) can be used
It will work with any well formed GTFS data set.
Node +12 is required for all examples.
npm install --save transfer-pattern-planner
The algorithm expects transfer patterns to be stored in a MySQL compatible table:
CREATE TABLE `transfer_patterns` ( `journey` char(6) NOT NULL, `pattern` varchar(255) NOT NULL, PRIMARY KEY (`journey`,`pattern`) ) ENGINE=InnoDB DEFAULT CHARSET=latin1
Where the journey is the origin and destination concatenated and the pattern are comma separated stops, excluding the origin and destination.
The following environment variables can set the database credentials and gtfs file location:
DATABASE_HOST=localhost DATABASE_USER=root DATABASE_NAME=jp DATABASE_PASS= GTFS=/path/to/gtfs.zip
Depart After Query
Find the first results that depart after a specific time
const Container = ;const container = ;const query = await container;const results = await query;await container; // database connection must be closed
Issues and PRs are very welcome. To get the project set up run:
git clone firstname.lastname@example.org:planarnetwork/transfer-pattern-planner npm install --dev npm test
If you would like to send a pull request please write your contribution in TypeScript and if possible, add a test.
This software is licensed under GNU GPLv3.