bct.community_louvain adjmat must not contain negative weights about bctpy HOT 5 CLOSED

I saw a previous post commenting the problem but nobody answer him:
#57

from bctpy.

I have checked the function, in line 110-111 you have this error message if the adjacency matrix has negative values:

if np.min(W) < -1e-10:
    raise BCTParamError('adjmat must not contain negative weights')

But then for lines 122 to 138 if we select the negative_sym option or the_asym the program should continue:
if B in ('negative_sym', 'negative_asym'):
W0 = W * (W > 0)
s0 = np.sum(W0)
B0 = W0 - gamma * np.outer(np.sum(W0, axis=1), np.sum(W, axis=0)) / s0

    W1 = W * (W < 0)
    s1 = np.sum(W1)
    if s1:
        B1 = (W1 - gamma * np.outer(np.sum(W1, axis=1), np.sum(W1, axis=0))
            / s1)
    else:
        B1 = 0

elif np.min(W) < -1e-10:
    raise BCTParamError("Input connection matrix contains negative "
        'weights but objective function dealing with negative weights '
        'was not selected')

However since you impose first the condition in lines 110-111 the function doesn't execute at all, I think that the condition from line 110-111 should be eliminated, since in lines 122 to 138 this problem is already dealt if the proper objective function isn't chosen.

from bctpy.

I believe you can happily eliminate the lines:

if np.min(W) < -1e-10:
    raise BCTParamError('adjmat must not contain negative weights')

Moreover you need to fix a bit the code, there are a few errors using 'negative_asym' or 'negative_sym'. The value of modularity is normalized by s twice in this case and the calculation of both B0 and B1 are wrong. I'll open a pull request with the fixed code.

from bctpy.

from bctpy.

Fixed in PR #65

from bctpy.