def calcHrespForMergePairs(resp, Data, mPairIDs, returnVec=1):
''' Calculate resp entropy terms for all candidate merge pairs
Returns
---------
Hresp : 2D array, size K x K
or
Hresp : 1D array, size M
where each entry corresponds to one merge pair in mPairIDs
'''
if hasattr(Data, 'word_count') and resp.shape[0] == Data.word_count.size:
if mPairIDs is None:
Hmat = -1 * calcRlogRdotv_allpairs(resp, Data.word_count)
else:
Hmat = -1 * calcRlogRdotv_specificpairs(
resp, Data.word_count, mPairIDs)
else:
if mPairIDs is None:
Hmat = -1 * calcRlogR_allpairs(resp)
else:
Hmat = -1 * calcRlogR_specificpairs(resp, mPairIDs)
if returnVec:
if Hmat.ndim == 2:
Hvec = np.zeros(len(mPairIDs))
for ii, (kA, kB) in enumerate(mPairIDs):
Hvec[ii] = Hmat[kA, kB]
else:
Hvec = Hmat
return Hvec
else:
return Hmat
def calcHrespForSpecificMergePairs(LP, Data, mPairIDs):
''' Calculate resp entropy terms for all candidate merge pairs
Returns
---------
Hresp : 1D array, size M
where each entry corresponds to one merge pair in mPairIDs
'''
assert mPairIDs is not None
if 'resp' in LP:
N = LP['resp'].shape[0]
if hasattr(Data, 'word_count') and N == Data.word_count.size:
m_Hresp = -1 * calcRlogRdotv_specificpairs(
LP['resp'], Data.word_count, mPairIDs)
else:
m_Hresp = -1 * calcRlogR_specificpairs(LP['resp'], mPairIDs)
else:
if LP['nnzPerRow'] == 1:
return None
N = LP['spR'].shape[0]
if hasattr(Data, 'word_count') and N == Data.word_count.size:
m_Hresp = calcSparseMergeRlogRdotv(
spR_csr=LP['spR'],
nnzPerRow=LP['nnzPerRow'],
v=Data.word_count,
mPairIDs=mPairIDs)
else:
m_Hresp = calcSparseMergeRlogR(
spR_csr=LP['spR'],
nnzPerRow=LP['nnzPerRow'],
mPairIDs=mPairIDs)
assert m_Hresp.size == len(mPairIDs)
return m_Hresp
def calcElogqZForMergePairs(self, resp, Data, mPairIDs):
''' Calculate resp entropy terms for all candidate merge pairs
Returns
---------
ElogqZ : 2D array, size K x K
'''
if hasattr(Data, 'word_count'):
if mPairIDs is None:
ElogqZMat = calcRlogRdotv_allpairs(resp, Data.word_count)
else:
ElogqZMat = calcRlogRdotv_specificpairs(
resp, Data.word_count, mPairIDs)
else:
if mPairIDs is None:
ElogqZMat = calcRlogR_allpairs(resp)
else:
ElogqZMat = calcRlogR_specificpairs(resp, mPairIDs)
return ElogqZMat