trainline-eu / csa-challenge Goto Github PK

Since https://github.com/BliksemLabs/rrrr/tree/tdata4 has a GTFS import, and the data format you have described is trivial. Would a utility to go from GTFS to your intermediate format something you wish to add to this repo?

Do you have any guidance as to how to address any of the limitations? Specifically I'm looking to address:

it computes the earliest arrival. However, a better solution might leaver later and arrive at the same time

Hi,

I have this issue running ruby test.rb *

ruby test.rb ./csa_cpp
Warning: you should require 'minitest/autorun' instead.
Warning: or add 'gem "minitest"' before 'require "minitest/autorun"'
From:
  /Users/poc/.rvm/rubies/ruby-2.1.5/lib/ruby/2.1.0/test/unit.rb:1:in `<top (required)>'
  test.rb:3:in `<main>'
MiniTest::Unit::TestCase is now Minitest::Test. From /Users/poc/.rvm/rubies/ruby-2.1.5/lib/ruby/2.1.0/test/unit/testcase.rb:8:in `<module:Unit>'
/Users/poc/.rvm/rubies/ruby-2.1.5/lib/ruby/2.1.0/test/unit.rb:676:in `<class:Runner>': undefined method `_run_suite' for class `Test::Unit::Runner' (NameError)
    from /Users/poc/.rvm/rubies/ruby-2.1.5/lib/ruby/2.1.0/test/unit.rb:261:in `<module:Unit>'
    from /Users/poc/.rvm/rubies/ruby-2.1.5/lib/ruby/2.1.0/test/unit.rb:15:in `<module:Test>'
    from /Users/poc/.rvm/rubies/ruby-2.1.5/lib/ruby/2.1.0/test/unit.rb:7:in `<top (required)>'
    from /Users/poc/.rvm/rubies/ruby-2.1.5/lib/ruby/site_ruby/2.1.0/rubygems/core_ext/kernel_require.rb:54:in `require'
    from /Users/poc/.rvm/rubies/ruby-2.1.5/lib/ruby/site_ruby/2.1.0/rubygems/core_ext/kernel_require.rb:54:in `require'
    from test.rb:3:in `<main>'

I can run ruby bench.rb *

Hello. I enjoyed this, thanks for sharing it.

I refactored the Rust version to be easier for me to read. That's a subjective goal, and most of the changes I made are departures from the reference implementation. I still wanted to share, though, so I'm posting a link to a gist instead of a pull request.

https://gist.github.com/m-n/8a15d864099778154529

I shortened some of the variable names and made an explicit Query struct. I use a corresponding parse function to move the query parsing logic out of the control loop. In my version (and the official Rust version), the timetable is not its own struct, though -- just a vector of connections.

The biggest change was refactoring the CSA. I kept two functions: query, which takes a timetable and a query and returns the best route as a vector of connections, and print_result, which iterates over the vector of connections and prints them.

Does the code has really to be the the "most unreadable" (see README)?

https://github.com/linkedconnections/csa.js

;-)

Hello,
I am trying to do "ruby test.rb ./csa_hs" on my comuter but I have a message "File not found" whereas both files: test.rb and csa_hs are in the folder for which I am calling them. Can somebody explain me what I am doing wrong? Thanks in advance.

The example in the readme works fine

#input.txt
1 2 3600 7200
2 3 7800 9000

1 3 3000

$ cat input.txt | ./csa_cpp
1 2 3600 7200
2 3 7800 9000

but if the input connections are inverted, the algorithm does not output a solution.

#input2.txt
2 3 7800 9000
1 2 3600 7200

1 3 3000

$ cat input2.txt | ./csa_cpp
NO_SOLUTION

The algorithm seems very similar to the Bellman–Ford shortest path algorithm. I think to fix the algorithm you have to repeat the computation of computing the earliest arrival for the number of stations - 1, so something like:

For #station - 1
    For each connection c
        if arrival_timestamp[c.departure_station] < c.departure_timestamp and arrival_timestamp[c.arrival_station] > c.arrival_timestamp
            arrival_timestamp[c.arrival_station] ← c.arrival_timestamp
            in_connection[c.arrival_station] = c

In many implementations the pruning condition is
if connection.arrivalTime > earliestArrivalTime then break loop
This is incorrect since we can't make assumptions on arrival times.

The correct pruning condition is
if connection.departureTime >= earliestArrivalTime then break loop

Pull request that addresses the issue: #32