FisherExact

Fisher exact test for mxn contingency table

Installation

FisherExact should be python2/3 compatible. You can install it with pip : pip install FisherExact

If you get an error about builtins module, install "future" with pip install future

This package use fortran sources, so you need to have a fortran compiler (gfortran) installed. See here ==> https://gcc.gnu.org/wiki/GFortranBinaries.

The source code was tested on Linux and Mac (thanks to @fomightez)

Binary Usage

A binary is provided to use FisherExact from the terminal

usage: fexact [-h] [--simulate [SIMULATE]] [--hybrid] [--midP]
              [--retry ATTEMPT] [--workspace WORKSPACE] [--version]
              table

Fisher's Exact test for mxn contingency table

positional arguments:
  table                 Contingency table in a file, without header

optional arguments:
  -h, --help            show this help message and exit
  --simulate [SIMULATE]
                        Simulate p-values with n replicates
  --hybrid              Use hybrid mode
  --midP                Use midP correction
  --retry ATTEMPT       Number of attempt to made if execution fail
  --workspace WORKSPACE
                        Workspace size to use, Increase this if the program
                        crash
  --version             show program's version number and exit

Contingency table format if fexact is used as binary

The accepted format is space/tab or comma (or both) separated values with an optionnal first line starting with a "#" that specified the number of rows and column:

For example, the following format are accepted

# 2 3
8        2       12
1        5       2

8 2 12
1 5 2

#2, 3
8	2	12
1	5	2

8,2,12
1,5,2

Use as a module

fisher_exact(table, alternative='two-sided', hybrid=False, midP=False, simulate_pval=False, replicate=2000, workspace=300, attempt=2, seed=None)
    Performs a Fisher exact test on a mxn contingency table.
    
    Parameters
    ----------
    table : array_like of ints
        A 2x2 contingency table.  Elements should be non-negative integers.
    alternative : {'two-sided', 'less', 'greater'}, optional
        Which alternative hypothesis to the null hypothesis the test uses.
        Default is 'two-sided'.  Only used in the 2 x 2 case (with the scipy function).
        In every other case, the two-sided pval is returned.
    mult : int 
        Specify the size of the workspace used in the network algorithm.  
        Only used for non-simulated p-values larger than 2 x 2 table. 
        You might want to increase this if the p-value failed
    hybrid : bool
        Only used for larger than 2 x 2 tables, in which cases it indicates
        whether the exact probabilities (default) or a hybrid approximation 
        thereof should be computed.
    midP : bool
        Use this to enable mid-P correction. Could lead to slow computation.
        This is not applicable for simulation p-values. `alternative` cannot 
        be used if you enable midpoint correction.
    simulate_pval : bool 
        Indicate whether to compute p-values by Monte Carlo simulation,
         in larger than 2 x 2 tables.
    replicate : int
        An integer specifying the number of replicates used in the Monte Carlo test.
    workspace : int
        An integer specifying the workspace size. Default value is 300. 
    attempt : int
        Number of attempts to try, if the workspace size is not enough. 
        On each attempt, the workspace size is doubled. 
    seed : int
        Random number to use as seed. If a seed isn't provided. 4 bytes will be read 
        from os.urandom. If this fail, getrandbits of the random module 
        (with 32 random bits) will be used. In the particular case where both failed,
        the current time will be used
    
    Returns
    -------
    p_value : float
        The probability of a more extreme table, where 'extreme' is in a 
        probabilistic sense.