gfpop Time-Dependent Constraints Project

Overview

This project extends the gfpop package to support time-dependent constraints in changepoint detection models. The implementation allows constraints to vary over time, enabling implementations like Labeled Optimal Partitioning and unconstrained optimal partitioning.

Project Structure

• easy_test/ - Implementation of the Easy Test (visualizing gfpop graphs)
  • plotModel.R - Main function for graph visualization
  • test_plotModel.R - Examples showing how to use plotModel
  • simple_gfpop_example.R - Beginner-friendly example script
  • README.md - Instructions specific to the easy test
• problem.md - Original problem statement and tests description

Completed Tests

Easy Test ✅

We've implemented a solution for the easy test:

Write an R function `plotModel` which takes a `gfpop::graph` as above and draws a matrix of nodes (one row for each state, one column for each rule) and edges.

Our implementation visualizes gfpop graphs as a matrix with:

• Each row representing a state
• Each column representing a rule ID
• Nodes at each state/rule intersection
• Directed edges showing valid transitions

Environment Setup

Prerequisites

• R (version 3.6.0 or higher)
• Rtools (for Windows users building packages from source)

Installing Required Packages

Run these commands in an R console:

install.packages("igraph")     # For graph visualization
install.packages("remotes")    # For installing packages from GitHub
remotes::install_github("vrunge/gfpop")  # Install gfpop from GitHub

Setting Up a Reproducible Environment with renv

For a fully reproducible environment, we use renv:

1. Start an R Console: Open RStudio or VS Code with R extension

2. Install renv:

   install.packages("renv")

3. Navigate to the project directory:

   setwd("path/to/gfpop-time-dependent-constraints")

4. Initialize the environment:

   renv::init()

5. Install required packages in the isolated environment:

   renv::install("igraph")
   renv::install("remotes")
   remotes::install_github("vrunge/gfpop")
   
   # For testing
   renv::install("testthat")

<<<<<<< HEAD

6. Save the environment state:

   renv::snapshot()

======= 2. Clone the Repository:

git clone https://github.com/Transcendental-Programmer/gfpop-time-dependent-constraints.git
>>>>>>> 7eb01425ac58e93e3e3d2978611808a48225b131

7. When returning to the project later:

   renv::restore()  # Reinstall all packages from the lockfile

Running Tests

We use testthat for testing the implementation:

Running Easy Test Visually

1. Navigate to the easy_test directory:

   setwd("path/to/gfpop-time-dependent-constraints/easy_test")

2. Run the example script:

   source("test_plotModel.R")

3. What you'll observe: Three graph visualizations will appear:

   • Simple Graph Example (2 states, 2 rules)
   • LOPART Graph Example (complex graph with multiple transitions)
   • Complex Graph Example (up/down states with multiple rules)

4. For a simpler example:

   source("simple_gfpop_example.R")

Implementation Approach

Easy Test Solution

The plotModel function visualizes gfpop graphs using these steps:

1. Extract unique states and rules from the graph
2. Create a matrix layout with states as rows and rules as columns
3. Build an igraph object with nodes at each state/rule intersection
4. Add edges according to the transitions defined in the graph
5. Render the visualization with igraph's plotting functions

File Overview

• easy_test/plotModel.R: Contains the main plotModel function implementation
• easy_test/test_plotModel.R: Demonstrates usage with multiple example graphs
• easy_test/simple_gfpop_example.R: A beginner-friendly example that works with any gfpop version

Usage Examples

library(gfpop)
library(igraph)

# Source the plotModel function
source("easy_test/plotModel.R")

# Create a graph (using the approach that works with all gfpop versions)
g <- list()
g$edges <- data.frame(
  state1 = c("normal", "normal", "noChange", "noChange"),
  state2 = c("normal", "noChange", "noChange", "normal"),
  type = c("null", "std", "null", "null"),
  penalty = c(0, 5.5, 0, 0),
  rule = c(1, 2, 2, 3),
  stringsAsFactors = FALSE
)
class(g) <- "graph"

# Visualize the graph
plotModel(g, title="Example Graph")

Contributors

• Priyansh Saxena - Implemented the Easy Test solution

Mentors

• Primary: Vincent Runge - [email protected]
• Co-Mentor: Toby Dylan Hocking - [email protected]

License

<<<<<<< HEAD This project is licensed under the terms of the R license.

For distribution details, run in an R session:

license()

Acknowledgements

This work extends established changepoint detection methods by incorporating time-sensitive constraints, with contributions from both the gfpop community and the broader changepoint research community.

7eb01425ac58e93e3e3d2978611808a48225b131