mapbox / mapbox/nepomuk

Serialise / Deserialise Output of GTFS Parser

Open
#20 0 comments 0 reactions 0 assignees View on GitHub
Backend Data
Dominant language
C++
Stars
24
Forks
7
PR merge metrics
No merged PRs in 30d

Description

Some items remain from the initial todo list of https://github.com/mapbox/directions-transit/pull/6.
To finalise the GTFS portion of the code, we need to implement full support for the protocol buffer scheme.

The Input part of the code should probably perform the following actions:

- [ ] make IDs consecutive integers (with the possibility of supporting multiple types)
- [ ] decide whether to keep access to original IDs (e.g. agency ID?)
- [ ] de-duplicate all string formats (dictionary + id look-up)
- [ ] de-duplication includes an additional .proto message to serialise out the dictionary
- [ ] create PBF encoders / decoders for the different gtfs types

This task will most likely change the internal message file format. While this is not intended to happen in a forward compatible PBF format, before the first release we will adjust the file format freely and as necessary.

Contributor guide

Open the contributing guide

Research direction

Start with the GTFS parser's Input path and the existing protocol buffer scheme referenced in the issue. Decide the representation for consecutive and original IDs, string dictionaries, and GTFS-type encoders and decoders. Done means the GTFS data can be serialized and deserialized with the revised internal message format.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp
Domain
backend, data
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.