Merge branch 'develop'

CHANGELOG.md

@@ -6,6 +6,18 @@ Most recent releases are shown at the top. Each release shows:
 - **Changed**: Additional parameters, changes to inputs or outputs, etc
 - **Fixed**: Bug fixes that don't change documented behaviour
 
+## 0.19.9 (2020-08-17)
+
+### New:
+- N/A
+
+### Changed
+- N/A
+
+### Fixed:
+- In `texts_from_array`, check `class_names` only after preprocessing before printing classification vs. regression status.
+
+
 ## 0.19.8 (2020-08-17)
 
 ### New:

ktrain/tests/test_linkpred.py

@@ -10,7 +10,7 @@
 from ktrain import graph as gr
 from ktrain.imports import ACC_NAME, VAL_ACC_NAME
 
-class TestNodeClassification(TestCase):
+class TestLinkPrediction(TestCase):
 
 
     def test_cora(self):

ktrain/tests/test_transformers.py

@@ -30,9 +30,16 @@ def setUp(self):
         y_train = train_b.target
         x_test = test_b.data
         y_test = test_b.target
+
+        # convert to string labels
+        y_train = [train_b.target_names[y] for y in y_train]
+        y_test = [train_b.target_names[y] for y in y_test]
+
+        # setup
         self.trn = (x_train, y_train)
         self.val = (x_test, y_test)
-        self.classes = train_b.target_names
+        #self.classes = train_b.target_names
+        self.classes = []  # discover from string labels
 
 
 

ktrain/text/data.py

@@ -330,13 +330,6 @@ def texts_from_array(x_train, y_train, x_test=None, y_test=None,
         verbose (boolean): verbosity
     """
     U.check_array(x_train,  y=y_train, X_name='x_train', y_name='y_train')
-    if not class_names and verbose:
-        #classes =  list(set(y_train))
-        #classes.sort()
-        #class_names = ["%s" % (c) for c in classes]
-        print('task: text regression (supply class_names argument if this is supposed to be classification task)')
-    else:
-        print('task: text classification')
 
     if x_test is None or y_test is None:
         x_train, x_test, y_train, y_test = train_test_split(x_train, y_train, 
@@ -369,6 +362,10 @@ def texts_from_array(x_train, y_train, x_test=None, y_test=None,
                            lang=lang, ngram_range=ngram_range)
     trn = preproc.preprocess_train(x_train, y_train, verbose=verbose)
     val = preproc.preprocess_test(x_test,  y_test, verbose=verbose)
+    if not preproc.get_classes() and verbose:
+        print('task: text regression (supply class_names argument if this is supposed to be classification task)')
+    else:
+        print('task: text classification')
     return (trn, val, preproc)
 
 

ktrain/version.py

@@ -1,2 +1,2 @@
 __all__ = ['__version__']
-__version__ = '0.19.8'
+__version__ = '0.19.9'