superzer0 / tda.java

Part of TDA Thesis

TDA.JAVA

Console application that can be used to generate persistence diagrams based on points cloud by means of TDA.
Program generates diagrams from collection of points in JSON and writes results files to the specified folder.
Program produces persistence diagrams (png), intervals (txt) and diagrams entropy values (json array in separate file).
Program execution for dimensions larger than 2 might be very memory consuming.
Program is based on Javaplex library for TDA related computing (https://github.com/appliedtopology/javaplex).

Building from sources

TDA.JAVA is maven based so building a package is as simple as:

mvn clean compile package assembly:single -e

Executing / Input

java -jar tdathesis-1.0-SNAPSHOT.jar <input_file_name> <max_dimension> <max_filtration> <use_landmarks> <max_landmarks> <output_folder>

e.g. java -jar C:\tda.java\tdathesis-1.0-SNAPSHOT.jar "C:\program_data\points_cloud.json" 2 118 True 200 "C:\program_data_output"

• <input_file_name> - path to JSON file with array of coordinates (points).
• <max_dimension> - integer value setting max value for dimensions to compute
• <max_filtration> - max value of parameter p (circle diameter) in which simplicial complexes are constructed
• <use_landmarks> - allows to replace large amount of input points by their landmarks reducing compute time
• <max_landmarks> - integer value setting how many points should be taken by landmarks
• <output_folder> - folder in which program should write its computing results

Output

Program produces serveral files in specified output folder

• Barcodes_<num_dimension>.png where <num_dimension> is number of the dimension processed
• persistent_homology_summary.txt - all discovered persistence intervals saved in text file
• entropy.json - JSON array with all barcodes entropy. (http://www.sciencedirect.com/science/article/pii/S0031320314002453)