jFit (deprecated) (see jfit2)

jFit is a Java utility that automates the generation of Python code for empirical complexity analysis by producing plots of time against input size (n) using Matplotlib, and fitting an appropriate curve

Features

• Automated python code generation

Prerequisites

• jFit relies on the Java Runtime Environment (JRE) for its proper execution. To run jFit, a compatible version of JDK and a compatible JVM must be installed on the system. Please refer to the official documentation for specific version requirements
• To execute the generated Python code, install the following packages using pip

  pip install scipy numpy matplotlib

Installation

• Clone the git repository:

  git clone https://github.com/fringewidth/jfit.git

• Open the package in a Java IDE of your choice(e.g., VSCode):

  code jfit/jFit

Using jFit

• Replace the body of public static void method(long[] array, long n) with the body of the java method of your choice. The array[] parameter represents the array to be manipulated, and n signifies the array's length.

• Important: Limitation on array analysis: Please note that jFit currently supports the analysis of algorithms that work with a single, one dimensional array only.

• Modify the parameters of ComplexityAnalyzer.generateMatPlotLib() in the main method. Adjust the parameters of:

  • startRange: Start of the input size n
  • endRange: End (non-inclusive) of the size n
  • incrementValue: Step size of input size n
  • efficiencyClass: Specify the efficiency class for your method. jFit generates empirical data on the running time across the specified n range and fits a corresponding curve.

• The following efficiency classes are currently avaliable:

  • EFFICIENCY_CLASS.CONSTANT for $O(1)$ complexities
  • EFFICIENCY_CLASS.LOGARITHMIC for $O(\log n)$ complexities
  • EFFICIENCY_CLASS.LINEAR for $O(n)$ complexities
  • EFFICIENCY_CLASS.POLYNOMIAL for $O(n^2)$ complexities
  • EFFICIENCY_CLASS.LINEARITHMIC for $O(n\log n)$ complexities
  • EFFICIENCY_CLASS.EXPONENTIAL for $O(2^n)$ complexities
  • EFFICIENCY_CLASS.FACTORIAL for $O(n!)$ complexities
    Support for $O(n^k)$ is planned.
    Note: Attempting to analyse linear time curves over large ranges of n results in a plot with several segmented lines, most likely due to caching effects. jFit currently can't fit a good curve to such data, but support for segmented regression is planned.

• Run the script.

Examples

• See examples.md for examples and specific use cases.