Skip to content

mfe-/DataStructures.Algorithms

Repository files navigation

DataStructures and algorithms

A collection of basic algorithms and data structures (algodat)

DataStructures (Experimental)

  • AvlTree
  • BstTree
  • The graphs datastructure with it vertices and edges is implemented as "linked list".
    • The idea is that it is not neccessary to load the entire graph to execute an algorithm.
    • Each vertex can save a generic Value.
    • The graph is fully (de)serializeable.
  • LinkedList (will be reimplemented see)
  • Documentation

nugets

nuget.org
Abstract.DataStructures
Abstract.DataStructures.Algorithms
Abstract.DataStructures.Algorithms.Graph
Abstract.DataStructures.UI
Azure DevOps
Pipeline
Doc Release Pipeline

DataStructures.UI

Wpf control to visualize, create and edit a graph. See here.

Live Samples

A* search algorithm

alt text

BFS - breadth first search

alt text

Kruskal

alt text

Introduction

When taking a beginners class at the technical university of vienna, I started to implement and adopt some lessions from the algorithms and data structures course. I thought it would be fun to put my recently gained knowledge into practice.

Of course some frameworks regarding this topic already exist. For example linked list and various tree classes. When applying these I ran into several issues like limited extensibility, missing features and all of them came from diffrent sources. Therefore I wasn't able to combine the tree classes, especially the data structure classes.

So I started to create my own algorithms and data structures "framework".

Goal

With the old version of the framework I mainly focused on visually representing the graph data structure with wpf. (see UI\DataStructures.Demo)

Some algo related links

Graph control

When creating the graph with the ui (graph control) the proper model will be created in the background. alt tag

Another approach would be to overgive a graph data structure to the graph visualization control which created the proper ui graph. One of the benefits of the old implemention is that the kruskal algorithm creats a copy of the graph, therefore the vertices dont't stay in the same position.

alt tag

When modifying a graph during runtime, the graph visualization control updates the ui.

alt tag

The aforementioned features (graph visualization control) will be implemented into the new framework as well.