Usage

Installation

Use the package manager [pip](https://pip.pypa.io/en/stable/) to install CHESSBOARD and the API.

$ pip install git+https://bitbucket.org/biociphers/chessboard

Usage

CHESSBOARD is a command line tool that can be run as follows using the provided arguments. Note that the input is a CHESSBOARD .cb file which can be generated using our API. The API allows the user to import raw data for processing prior to running CHESSBOARD and explore the algorithm output for downstream analsysis.

usage: chessboard [-h] [-A CONC] [-L REG] [-I ITER] [-B BURNIN] [-T THIN]
                     [-S BLOCKSIZE] [-M minsize] [-K kinit] [-V verbose] [--version]
                     input out_dir

positional arguments:
  input                 Path to the chessboard input file.
  out_dir               Path to the output directory; will be created if it
                        does not already exist.

optional arguments:
  -h, --help            show this help message and exit
  -A CONC, --conc CONC  Concentration parameter. Larger values increase the
                        number of learned clusters. (default: 1e-100)
  -L REG, --reg REG     Regularization parameter. (default: 50)
  -I ITER, --iter ITER  Number of iterations to run the Gibbs sampler.
                        (default: 1000)
  -B BURNIN, --burnin BURNIN
                        Number of burn-in iterations. (default: 200)
  -T THIN, --thin THIN  Number of samples to discard between iterations to
                        reduce autocorrelation. (default: 2)
  -S BLOCKSIZE, --blocksize BLOCKSIZE
                        Maximum block size for blocked Gibbs sampling
                        procedure. (default: 10)
  -M MINSIZE, --minsize MINSIZE
            Minimum tile size. (default: 20)

  -K KINIT, --kinit KINIT
            Number of clusters for k-means initialization. (default: 5)

  -V, --verbose
            Verbose mode. (default: False)
  --version             show program's version number and exit