Ticket machine project

Library for download list of stations and give interface for UI for select one based on typed letters

Full documentation is in refman.pdf and can be regenerated with doxygen

Goals

The main goal is return as fast as possible acceptable chars for complete search and possible stations based on input query.

Result can be used to get price of ticket by using station code and buy selected choice. Pricing should be implemented in another projects as is no information and requirements for that.

Requirements:

1. Typing a search string will return:
   1. All stations that start with the search string.
   2. All valid next characters for each matched station.
2. Space is a valid character when returning a list of next characters.
3. The service response will be used for the machine UI and for the routing and pricing purposes.

Operational requirements:

1. Runtime speed is very important, loading time is not.

2. Make no assumptions about the data source in real life.

3. Expected Scenarios:

   • Given a list of stations DARTFORD, DARTON, TOWER HILL, DERBY
     1. When input DART
     2. Then should return:
        • The characters of F O
        • The stations DARTFORD, DARTON
   • Given a list of stations LIVERPOOL, LIVERPOOL LIME STREET, PADDINGTON
     1. When input LIVERPOOL
     2. Then should return:
        • The characters of ' ' (space)
        • The stations LIVERPOOL, LIVERPOOL LIME STREET
   • Given a list of stations EUSTON, LONDON BRIDGE, VICTORIA
     1. When input KINGS CROSS
     2. Then should return:
        • no next characters
        • no stations

   These scenarios are tested in project:

   • TicketMachine.DB.Tests with tests in OperationalRequirementsTests/ScenariosTest.cs
   • TicketMachine.Controllers.Tests with tests in OperationalRequirementsTests/ControllerScenariosTest.cs

Others requirements

User can input only ASCII letters (A-Z) and space.

Example applications

In the projects TicketMachine.md and TicketMachine.WebApi.md there is an example application that uses a library for searching stations.

Used external libraries

Libraries what not come from Microsoft company:

• Palmer, licence MIT:
  • simple library providing retry logic
  • source:
    • https://github.com/mitchdenny/palmer/tree/master
    • https://www.nuget.org/packages/Palmer
• MockHttp, licence MIT
  • Testing layer for Microsoft HttpClient providing mock
  • source:
    • https://github.com/richardszalay/mockhttp
    • https://www.nuget.org/packages/RichardSzalay.MockHttp
• Json.Net, licence MIT:
  • Json parser
    • source:
      • https://www.nuget.org/packages/Newtonsoft.Json
      • https://www.newtonsoft.com/json

Others:

• NUnit (with NUnit3TestAdapter, NUnit.Analyzer)
• Moq (with Castle.Core)
• coverlet.collector

Additional tools

Documentation generation

Documentation are generated by doxygen and plantuml:

• https://doxygen.nl/
• https://plantuml.com/

For doxygen there is configuration file: doxygen.conf For simplify process there is script written in bash: make_doc.sh

Status

Work is split into two stages:

• stage 1 - most important thinks to make in given time
• stage 2 - future improvements

Stage 1 [DONE]

Status	Task
[DONE]	Working code
[DONE]	Documentation: * Function and code description * Graphs * Project description
[DONE]	Tests

Stage 2 (future improvements)

Tasks:

• Add logger (NLog)
• Tests on big data
• Improve speed:
  • Add letters indexing
  • Check other search algorithm
    • Check if mixed algorithms can improve speed
      • Search params by genetic or linear algorithm
  • Retest
• Check different memory access and allocating

Ticket Machine Licence

Licence MIT is in file LICENSE.md