Skip to content

mimuw-jnp2-rust/project-mikolaj-wasiak-mikolaj-pirog

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Graphs

Authors

  • Mikołaj Piróg (@aetn23)
  • Mikołaj Wasiak (@RudyMis)

Overview

A Rust application to edit graphs and run selected algorithms on them.

Inspiration: this editor

Features

Done in part 1

  • adding and removing nodes,
  • adding and deleting directed edged between nodes,
  • moving the graph around the window screen,
  • gravity between nodes,
  • basic camera transformations: moving (on WSAD), rotating (on Q,E), zooming (on mouse wheel),
  • running dfs visualization on the graph,
  • creating a clique out of spawned nodes,

Done in part 2

  • removing of edges
  • adjustments to camera work
  • small ui rework
  • bfs
  • algorithms now require the user to choose a starting point (by clicking on a node and then on algorithm button)
  • ability to write in node on RMB
  • generation of random graph
  • gravity slides now work correctly
  • bugfixes and code simplification/refactoring - that constitutes majority of commits in this part
  • small unit tests for graphs
  • big refactor of handling of the algorithm part of the app - it is now easier to implement new algorithms that work on generic graphs and offer big flexibility of what can they do

Code walkthrough

The three meaningful parts of our application are graphs, step algorithms and the whole drawing to the screen using egui_tetra.

Graph module implements internal graph representation using petgraph, as well as utility function on graph, gravity and drawing in particular.

Step algorithms now implements a trait which requires a single function returning a vector of algorithm steps, which is also a trait. That allows a gradual control over what an algorithm does on a graph. It is also easier to actually see what is required to implement a working algorithm.

Drawing utilizes egui for gui (duh), which is pretty self-explanatory in the code itself. As of now, UI drawing awaits some better modularization, instead of simple function call to draw everything. Tetra code is mostly handled in game_state.rs, which calls appropriate functions (drawing, updating, handling events) to handle drawing the screen. Important thing to note is the details of updating/drawing given thing are delegated to thing's struct, i.e. graph drawing is implemented in the graph module and called in the game_state.

We encountered little problems of rusty nature - the entire process of producing the code was surprisingly pleasant. Sometimes we had to dust off our linear algebra skills, though. The biggest challenge was how to approach the modularization. As of now, we do not claim that our approach as of now is the best and/or final one.

The biggest change to code structure is rework of step_algorithm and packing of some GameState fields into a separate struct to ease argument passing. Things outside gamestate.rs implement custom trait that requires update and draw function to be present (the tetra one requires too much).

Libraries

Petgraph for graphs structures, egui_tetra for graphics. egui_tetra is a wrapper for egui, a gui library, and tetra, a library for game development.

We also use dyn_partial_eq because of this.

We must say, handling the libraries went surprisingly smooth. Petgraph is widely used and pretty mature, so we expected no problems and it delivered. On the other hand, egui_tetra was looking suspicious - it's managed by Literally One Guytm, so we feared buggy behaviour - especially after Hello World! refused to even compile on the newest version, and when it did on downgraded one, it produced extremely buggy results. Fortunately, everything went smooth, and no problems were encountered during the development itself.

Installation

Petgraph and egui install their dependencies' from crates - no work required on our part. Tetra has some dependencies that need to be installed manually - see this.

About

project-mikolaj-wasiak-mikolaj-pirog created by GitHub Classroom

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages