def test_ldac_conversion(self):
dtm = self.dtm
N, V = dtm.shape
doclines = list(utils.dtm2ldac(self.dtm))
nd_unique = np.sum(dtm > 0, axis=1)
for n, docline in zip(nd_unique, doclines):
self.assertEqual(n, int(docline.split(' ')[0]))
self.assertEqual(len(doclines), N)
f = io.StringIO('\n'.join(doclines))
dtm_new = utils.ldac2dtm(f)
self.assertTrue(np.all(dtm == dtm_new))
def test_ldac_conversion(self):
dtm = self.dtm
N, V = dtm.shape
doclines = list(utils.dtm2ldac(self.dtm))
nd_unique = np.sum(dtm > 0, axis=1)
for n, docline in zip(nd_unique, doclines):
self.assertEqual(n, int(docline.split(' ')[0]))
self.assertEqual(len(doclines), N)
f = io.StringIO('\n'.join(doclines))
dtm_new = utils.ldac2dtm(f)
self.assertTrue(np.all(dtm == dtm_new))
def test_ldac2dtm(self):
test_dir = os.path.dirname(__file__)
reuters_ldac_fn = os.path.join(test_dir, 'reuters.ldac')
dtm = utils.ldac2dtm(open(reuters_ldac_fn))
self.assertEqual(dtm.shape, (395, 4258))
self.assertEqual(dtm.sum(), 84010)
def test_ldac2dtm(self):
test_dir = os.path.dirname(__file__)
reuters_ldac_fn = os.path.join(test_dir, 'reuters.ldac')
dtm = utils.ldac2dtm(open(reuters_ldac_fn))
self.assertEqual(dtm.shape, (395, 4258))
self.assertEqual(dtm.sum(), 84010)
def load_ldac(file_path):
return ldac2dtm(open(file_path), offset=0)
