As SkyAlps I would like to convert the currently available GTFS export in NeTEx, so that this data can be integrated in the Italian National Access Point and be used by national MaaS Operators

[rcavaliere]

At national level there are GTFS converters that can convert GTFS exports in NeTEx, ensuring the compliance with the Italian specifications. In order to use one of them and produce the requested NeTEx export also for the SkyAlps dataset, we need to fix the GTFS export as follows:

• we need to add the feed_info.txt file (see https://gtfs.org/schedule/reference/). Many fields can be copied from the already existing agency.txt file
• we need to add the shapes.txt file containing mainly the distances between start / destination couples. This could be automatically calculated by considering in input the coordinates of the airports considered.
• the existing file trips.txt has to include the field shape_id, i.e. the reference to the shapes.txt file
• the existing file stop_times.txt has to include the field shape_dist_traveled, i.e. the distance between two consecutive stops. Must be consistent with the distances provided in the shapes.txt file.

Activities to do:

• [x] Specify how the additional fields requested have to be populated.
• [x] Implement a module that integrates the additional parts to the GTFS exports (or integrate all this in the module that automatically generates the GTFS export)

[rcavaliere]

@clezag some indications for the provision of the additional elements in the GTFS:

• feed_info.txt

• • feed_publisher_name = agency_name (in agency.txt)
• • feed_publisher_url = agency_url (in agency.txt)
• • feed_lang = 'IT'

• shape.txt: each shape associated to a route is characterized by two points. So, two records in the table for every shape_id.

• • shape_id: could be an automatically generated ID based on the reference route_id (e.g. route_id = BZO_SUF -> shape_id = BZO_SUF_shape)
• • shape_pt_lat: latitude of the origin / departure airport
• • shape_pt_lon: longitude of the origin / departure airport
• • shape_pt_sequence: 1 or 2 (origin = 1, departure = 2)
• • shape_dist_traveled: if shape_pt_sequence = 1 then shape_dist_traveled = 0 otherwise Haversine distance between the origin and airport

For the implementation of the Haversine distance check e.g. https://it.martech.zone/calculate-great-circle-distance/

Remember also to include the references to the shapes.txt fields as mentioned in the main description, namely in the trips.txt file and in the stop_times.txt file.

[clezag]

@rcavaliere now up in testing. https://flighdata-skyalps-gtfs-test.s3.eu-west-1.amazonaws.com/GTFS.zip

[rcavaliere]

@clezag very good, I will test it and let you know in case we need some adaptations

[rcavaliere]

@clezag seems to be OK, the only thing which I think is not correctly implemented is in my view the shape_id, which is equivalent to the route_id. And why we have multiple records for each shape_id? We should have maximum two records per shape_id in the shapes.txt file.

[clezag]

@rcavaliere good catch! Fixed now in testing

[rcavaliere]

@clezag now shapes.txt looks fine! The only thing which is probably still not optimal is that route_id and shape_id are identical, maybe we could append something in the shape_id to differentiate them (see my proposal above).