dummy test for svm

README.md

@@ -44,5 +44,5 @@ python src/hogwild/coordinator.py
 #### Run tests
 ```bash
 source activate hogwild-python 
-py.test --color=yes
+py.test --color=yes -v
 ```

src/hogwild/settings.py

@@ -15,4 +15,4 @@
 test_percentage = 0.1
 epochs = 100
 batch_size = 100
-lambda_reg=1e-5
+lambda_reg = 1e-5

src/hogwild/svm.py

@@ -1,4 +1,4 @@
-from utils import dotproduct, sign
+from hogwild.utils import dotproduct, sign
 
 
 class SVM:

tests/test_svm.py

@@ -0,0 +1,32 @@
+from hogwild.svm import SVM
+
+mock_data = [{1: 0.1, 2: 0.2},
+             {1: 0.2, 4: 0.9},
+             {3: 0.9, 8: 1},
+             {4: 0.4, 5: 0.7}]
+mock_labels = [1, -1, 1, -1]
+
+mock_delta_w = {1: 0.01, 2: 0.02, 3: 0.03, 4: 0.04, 5: 0.05, 6: 0, 7: 0, 8: 0.08}
+
+
+def test_fit():
+    svm = SVM(1, 1e-5, 9)
+    expected_result = {1: -0.100001,
+                       2: 0.2,
+                       3: 0.9,
+                       4: -1.29999199999,
+                       5: -0.6999909999899999,
+                       8: 1.0}
+
+    result = svm.fit(mock_data, mock_labels)
+
+    assert expected_result == result
+
+
+def test_predict():
+    svm = SVM(1, 1e-5, 9)
+    svm.fit(mock_data, mock_labels)
+    expected_result = [1]
+    result = svm.predict([{2: 0.8, 3: 0.9}])
+
+    assert expected_result == result