def qm_result_matrices(matrices, tmp_mean, multiprocessing=True):
"""builds the resulting matrices by looking at the rank of their
original values and retrieving the means at the specified position"""
if multiprocessing:
# parallelized ranking
with util.get_mp_pool() as pool:
results = pool.map(rank_fun, [(matrix.values, matrix.row_names,
matrix.column_names, tmp_mean)
for matrix in matrices])
return results
else:
# non-parallelized
result = []
for i in xrange(len(matrices)):
matrix = matrices[i]
values = matrix.values
num_rows, num_cols = values.shape
rankvals = util.rrank_matrix(values)
values = np.reshape(tmp_mean[rankvals], (num_rows, num_cols))
outmatrix = DataMatrix(num_rows,
num_cols,
matrix.row_names,
matrix.column_names,
values=values)
result.append(outmatrix)
return result
def qm_result_matrices(matrices, tmp_mean, multiprocessing=True):
"""builds the resulting matrices by looking at the rank of their
original values and retrieving the means at the specified position"""
if multiprocessing:
# parallelized ranking
with util.get_mp_pool() as pool:
results = pool.map(rank_fun,
[(matrix.values, matrix.row_names, matrix.column_names, tmp_mean)
for matrix in matrices])
return results
else:
# non-parallelized
result = []
for i in xrange(len(matrices)):
matrix = matrices[i]
values = matrix.values
num_rows, num_cols = values.shape
rankvals = util.rrank_matrix(values)
values = np.reshape(tmp_mean[rankvals], (num_rows, num_cols))
outmatrix = DataMatrix(num_rows,
num_cols,
matrix.row_names,
matrix.column_names,
values=values)
result.append(outmatrix)
return result
def rank_fun(mat_mean):
"""ranking function that is run within Pool.map()"""
values, row_names, column_names, tmp_mean = mat_mean
num_rows, num_cols = values.shape
rankvals = util.rrank_matrix(values)
values = np.reshape(tmp_mean[rankvals], (num_rows, num_cols))
return DataMatrix(num_rows, num_cols, row_names, column_names,
values=values)
def rank_fun(mat_mean):
"""ranking function that is run within Pool.map()"""
values, row_names, column_names, tmp_mean = mat_mean
num_rows, num_cols = values.shape
rankvals = util.rrank_matrix(values)
values = np.reshape(tmp_mean[rankvals], (num_rows, num_cols))
return DataMatrix(num_rows,
num_cols,
row_names,
column_names,
values=values)