aboutsummaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* nix: strip ghc dependencies from outputstuebinm2022-11-071-1/+12
|
* gtfs-rt: don't remove SCHEDULED from schedule_relationshipstuebinm2022-10-161-2/+2
| | | | | (since it's the default value it's not serialised; but best practices say to have it there explicitly)
* fix nix buildstuebinm2022-10-161-2/+2
|
* fix gtfs tripupdatesstuebinm2022-10-161-83/+92
|
* simple prometheus metricsstuebinm2022-10-165-9/+55
|
* update glossarystuebinm2022-09-141-5/+8
|
* remove some extrapolation bugsstuebinm2022-09-144-61/+66
|
* on-board-unit: display estimated delay etc.stuebinm2022-09-116-7/+35
|
* update todo filestuebinm2022-09-111-4/+10
|
* correct the generated openapi descriptionstuebinm2022-09-113-5/+23
|
* gtfs realtime: add tripUpdate feedstuebinm2022-09-105-33/+104
|
* use websockets for the on-board-unitstuebinm2022-09-103-51/+79
|
* fix the close-to-a-station bugstuebinm2022-09-091-22/+31
| | | | | | (previously tracktrain could end up in a situation where the next and last station weren't actually adjacent stops, which messed up the prediction)
* init onboard-unitstuebinm2022-09-034-0/+121
|
* fix the nix buildstuebinm2022-09-022-15/+21
|
* reasonable delay forecastsstuebinm2022-09-026-66/+100
|
* guess at future delays (horrible, incorrect, and unfinished)stuebinm2022-08-315-43/+95
|
* some config thingyesodstuebinm2022-08-289-43/+108
| | | | | works kinda well, but doesn't complain about unknown config values in json, which is kinda hmpf tbh
* this does way too much tbh (also functioning delays)stuebinm2022-08-288-99/+351
| | | | most of it deals with timezones, and all the weird implications that has
* update todo.orgstuebinm2022-08-271-25/+14
|
* controlroom: lots of pretty little knobsstuebinm2022-08-279-123/+204
| | | | (also some database schema changes, for good measure)
* controlroom: some stylingstuebinm2022-08-264-44/+95
|
* ControlRoom: default layoutstuebinm2022-08-142-18/+31
|
* control room: translations & form inputstuebinm2022-08-148-201/+185
| | | | still basic, but works pretty well overall
* controlroom: replace servant/lucid with yesodstuebinm2022-08-146-272/+214
| | | | aka use something meant for webapps to write the webapp
* better Show instance for Timestuebinm2022-08-141-3/+7
|
* a little more control room workstuebinm2022-07-202-9/+39
|
* some work on the control roomstuebinm2022-07-173-32/+80
| | | | (it can do useful stuff! but it also still breaks on some things!)
* fix an operator precedence bugstuebinm2022-07-171-6/+6
|
* add an explanatory glossarystuebinm2022-07-111-0/+63
| | | | (also a reminder to me to stay at least somewhat consistent with names)
* somewhat functioning control roomstuebinm2022-07-118-36/+118
| | | | | | tbh i've kinda lost track at what has all been changed in this, but the control room form handling now works, and i can write announcements into the database. Now on to making it do useful things!
* generic html forms for ADTsstuebinm2022-07-073-88/+123
| | | | this does some GHC generics witchery, now in its own module!
* don't have the ControlRoom API in Swaggerstuebinm2022-07-073-13/+3
| | | | | | just causes pain and isn't really useful for anything since it's a webform API meant to be used by the web frontend, not by other applications.
* wrote some generic html form codestuebinm2022-07-061-12/+97
| | | | | this automatically derives lucid html versions of ADTs which are hopefully compatible with the FormUrlEncoded type class
* barebones webform API & frontendstuebinm2022-07-036-18/+104
|
* remove some unused importsstuebinm2022-07-033-89/+63
|
* add service monad (with built-in logging)stuebinm2022-07-033-20/+35
|
* websockets and better (empty) response messagesstuebinm2022-07-024-58/+86
| | | | (tough mostly untested)
* update todo.orgstuebinm2022-07-021-2/+4
|
* better module namesstuebinm2022-07-023-3/+4
|
* rudimentary admin api implementationstuebinm2022-07-024-10/+26
|
* gtfs realtime: add vehicle positionsstuebinm2022-07-023-17/+71
| | | | | | (or at least something that looks like it might be correct — though it carries very little information, and lacks things like what the next stop along the route is)
* more or less functional servicealerts for gtfs rtstuebinm2022-07-028-379/+232
| | | | (kinda barebones, but the important things should be there)
* gtfs realtime proof of conceptstuebinm2022-07-0239-11/+4479
| | | | | | | | this adds a package for protobuf stuff, generated via hprotoc. Seems to work kinda fine? (the generated API is horrible though, will have to write some wrappers for that)
* add a todo filestuebinm2022-06-271-0/+18
|
* optional date parameter for the timetable endpointstuebinm2022-06-272-16/+31
| | | | (mostly to make debugging easier, but also more generially useful i guess)
* fix hie.yamlstuebinm2022-06-271-2/+2
|
* handle GTFS shapes (& send them out along trips)stuebinm2022-06-193-14/+65
| | | | | unfortunately, this makes some responses pretty large. Let's see if it works this way or if i have to find some way to break them up instead.
* make connstring configurable at runtimestuebinm2022-06-182-1/+7
|
* generate javascript functionsstuebinm2022-06-172-0/+36
| | | | | | (this is commented out in the cabal file, since unfortunately the servant-js package is broken in nixpkgs and therefore otherwise breaks the nix build process … but it can be used "by hand")