ERROR: Cannot have test within function name

[mortonjt]

I have written up a bunch of tests. The tests pass just fine if I run the following

python test_permutation.py

Now if I were to run nosetests as follows

nosetests .

I get the following error

======================================================================
ERROR: Conducts a fishers test on a contingency table.
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/Users/mortonjt/miniconda3/envs/canvas/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/Users/mortonjt/miniconda3/envs/canvas/lib/python2.7/site-packages/nose/util.py", line 620, in newfunc
    return func(*arg, **kw)
TypeError: fisher_mean_test() takes at least 2 arguments (0 given)

----------------------------------------------------------------------

Now the function that I'm trying to test is under permutation.py, the function signature+docstring is as follows

def fisher_mean_test(table, grouping, permutations=1000, random_state=None):
    """ Conducts a fishers test on a contingency table.

    This module will conduct a mean permutation test using
    numpy matrix algebra.

    table: pd.DataFrame
        Contingency table of where columns correspond to features
        and rows correspond to samples.
    grouping : pd.Series
        Vector indicating the assignment of samples to groups.  For example,
        these could be strings or integers denoting which group a sample
        belongs to.  It must be the same length as the samples in `table`.
        The index must be the same on `table` and `grouping` but need not be
        in the same order.
    permutations: int
         Number of permutations to calculate
    random_state : int or RandomState, optional
        Pseudo number generator state used for random sampling.

    Return
    ------
    pd.DataFrame
        A table of features, their t-statistics and p-values
        `"m"` is the t-statistic.
        `"pvalue"` is the p-value calculated from the permutation test.

    Examples
    --------
    >>> from canvas.stats.permutation import fisher_mean
    >>> import pandas as pd
    >>> table = pd.DataFrame([[12, 11, 10, 10, 10, 10, 10],
    ...                       [9,  11, 12, 10, 10, 10, 10],
    ...                       [1,  11, 10, 11, 10, 5,  9],
    ...                       [22, 21, 9,  10, 10, 10, 10],
    ...                       [20, 22, 10, 10, 13, 10, 10],
    ...                       [23, 21, 14, 10, 10, 10, 10]],
    ...                      index=['s1','s2','s3','s4','s5','s6'],
    ...                      columns=['b1','b2','b3','b4','b5','b6','b7'])
    >>> grouping = pd.Series([0, 0, 0, 1, 1, 1],
    ...                      index=['s1','s2','s3','s4','s5','s6'])
    >>> results = fisher_mean(table, grouping,
    ...                       permutations=100, random_state=0)
    >>> results
                m    pvalue
    b1  14.333333  0.108910
    b2  10.333333  0.108910
    b3   0.333333  1.000000
    b4   0.333333  1.000000
    b5   1.000000  1.000000
    b6   1.666667  0.108910
    b7   0.333333  1.000000

    Notes
    -----
    Only works on binary classes.
    """

It took me a few hours to realize that nosetest was treating this function as a unittest, and was trying to execute it. Just because the word test is inside the function name. The nosetests worked when I renamed the function the fisher_mean

I'm not sure what the best way around this problem is, but it took a while to hunt down this problem and it would be nice to at least document this issue.

[jszakmeister]

It is documented: http://nose.readthedocs.io/en/latest/writing_tests.html#writing-tests

You can use the @nottest decorator to mark it as not being a test, if you want to keep the old name:

from nose.tools import *

@nottest
def fisher_mean_test(table, grouping, permutations=1000, random_state=None):
    # ...

[mortonjt]

Thanks!

On Tue, Jun 21, 2016 at 12:21 PM, John Szakmeister <notifications@github.com

wrote:

It is documented:
http://nose.readthedocs.io/en/latest/writing_tests.html#writing-tests

You can use the @nottest decorator to mark it as not being a test, if you
want to keep the old name:

from nose.tools import *
@nottestdef fisher_mean_test(table, grouping, permutations=1000, random_state=None):
# ...

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
#1003 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/AD_a3Y4UyYtUcKvPV_axVcMJOOc1TlFRks5qODnKgaJpZM4I7DaE
.

[eggachecat]

@mortonjt @jszakmeister you saved my life!!