Second working commit.

1 files changed, 138 insertions, 0 deletions

diff --git a/extractor/precomputation_test.cc b/extractor/precomputation_test.cc
new file mode 100644
index 00000000..9edb29db
--- /dev/null
+++ b/extractor/precomputation_test.cc
@@ -0,0 +1,138 @@
+#include <gtest/gtest.h>
+
+#include <memory>
+#include <vector>
+
+#include "mocks/mock_data_array.h"
+#include "mocks/mock_suffix_array.h"
+#include "precomputation.h"
+
+using namespace std;
+using namespace ::testing;
+
+namespace {
+
+class PrecomputationTest : public Test {
+ protected:
+  virtual void SetUp() {
+    data = {4, 2, 3, 5, 7, 2, 3, 5, 2, 3, 4, 2, 1};
+    data_array = make_shared<MockDataArray>();
+    EXPECT_CALL(*data_array, GetData()).WillRepeatedly(ReturnRef(data));
+
+    vector<int> suffixes{12, 8, 5, 1, 9, 6, 2, 0, 10, 7, 3, 4, 13};
+    vector<int> lcp{-1, 0, 2, 3, 1, 0, 1, 2, 0, 2, 0, 1, 0, 0};
+    suffix_array = make_shared<MockSuffixArray>();
+    EXPECT_CALL(*suffix_array, GetData()).WillRepeatedly(Return(data_array));
+    for (size_t i = 0; i < suffixes.size(); ++i) {
+      EXPECT_CALL(*suffix_array,
+                  GetSuffix(i)).WillRepeatedly(Return(suffixes[i]));
+    }
+    EXPECT_CALL(*suffix_array, BuildLCPArray()).WillRepeatedly(Return(lcp));
+  }
+
+  vector<int> data;
+  shared_ptr<MockDataArray> data_array;
+  shared_ptr<MockSuffixArray> suffix_array;
+};
+
+TEST_F(PrecomputationTest, TestInvertedIndex) {
+  Precomputation precomputation(suffix_array, 100, 3, 10, 5, 1, 4, 2);
+  Index inverted_index = precomputation.GetInvertedIndex();
+
+  EXPECT_EQ(8, inverted_index.size());
+  vector<int> key = {2};
+  vector<int> expected_value = {1, 5, 8, 11};
+  EXPECT_EQ(expected_value, inverted_index[key]);
+  key = {3};
+  expected_value = {2, 6, 9};
+  EXPECT_EQ(expected_value, inverted_index[key]);
+  key = {4};
+  expected_value = {0, 10};
+  EXPECT_EQ(expected_value, inverted_index[key]);
+  key = {5};
+  expected_value = {3, 7};
+  EXPECT_EQ(expected_value, inverted_index[key]);
+  key = {4, 2};
+  expected_value = {0, 10};
+  EXPECT_EQ(expected_value, inverted_index[key]);
+  key = {2, 3};
+  expected_value = {1, 5, 8};
+  EXPECT_EQ(expected_value, inverted_index[key]);
+  key = {3, 5};
+  expected_value = {2, 6};
+  EXPECT_EQ(expected_value, inverted_index[key]);
+  key = {2, 3, 5};
+  expected_value = {1, 5};
+  EXPECT_EQ(expected_value, inverted_index[key]);
+
+  key = {2, 4};
+  EXPECT_EQ(0, inverted_index.count(key));
+}
+
+TEST_F(PrecomputationTest, TestCollocations) {
+  Precomputation precomputation(suffix_array, 3, 3, 10, 5, 1, 4, 2);
+  Index collocations = precomputation.GetCollocations();
+
+  EXPECT_EQ(-1, precomputation.NON_TERMINAL);
+  vector<int> key = {2, 3, -1, 2};
+  vector<int> expected_value = {1, 5, 1, 8, 5, 8, 5, 11, 8, 11};
+  EXPECT_EQ(expected_value, collocations[key]);
+  key = {2, 3, -1, 2, 3};
+  expected_value = {1, 5, 1, 8, 5, 8};
+  EXPECT_EQ(expected_value, collocations[key]);
+  key = {2, 3, -1, 3};
+  expected_value = {1, 6, 1, 9, 5, 9};
+  EXPECT_EQ(expected_value, collocations[key]);
+  key = {3, -1, 2};
+  expected_value = {2, 5, 2, 8, 2, 11, 6, 8, 6, 11, 9, 11};
+  EXPECT_EQ(expected_value, collocations[key]);
+  key = {3, -1, 3};
+  expected_value = {2, 6, 2, 9, 6, 9};
+  EXPECT_EQ(expected_value, collocations[key]);
+  key = {3, -1, 2, 3};
+  expected_value = {2, 5, 2, 8, 6, 8};
+  EXPECT_EQ(expected_value, collocations[key]);
+  key = {2, -1, 2};
+  expected_value = {1, 5, 1, 8, 5, 8, 5, 11, 8, 11};
+  EXPECT_EQ(expected_value, collocations[key]);
+  key = {2, -1, 2, 3};
+  expected_value = {1, 5, 1, 8, 5, 8};
+  EXPECT_EQ(expected_value, collocations[key]);
+  key = {2, -1, 3};
+  expected_value = {1, 6, 1, 9, 5, 9};
+  EXPECT_EQ(expected_value, collocations[key]);
+
+  key = {2, -1, 2, -1, 2};
+  expected_value = {1, 5, 8, 5, 8, 11};
+  EXPECT_EQ(expected_value, collocations[key]);
+  key = {2, -1, 2, -1, 3};
+  expected_value = {1, 5, 9};
+  EXPECT_EQ(expected_value, collocations[key]);
+  key = {2, -1, 3, -1, 2};
+  expected_value = {1, 6, 8, 5, 9, 11};
+  EXPECT_EQ(expected_value, collocations[key]);
+  key = {2, -1, 3, -1, 3};
+  expected_value = {1, 6, 9};
+  EXPECT_EQ(expected_value, collocations[key]);
+  key = {3, -1, 2, -1, 2};
+  expected_value = {2, 5, 8, 2, 5, 11, 2, 8, 11, 6, 8, 11};
+  EXPECT_EQ(expected_value, collocations[key]);
+  key = {3, -1, 2, -1, 3};
+  expected_value = {2, 5, 9};
+  EXPECT_EQ(expected_value, collocations[key]);
+  key = {3, -1, 3, -1, 2};
+  expected_value = {2, 6, 8, 2, 6, 11, 2, 9, 11, 6, 9, 11};
+  EXPECT_EQ(expected_value, collocations[key]);
+  key = {3, -1, 3, -1, 3};
+  expected_value = {2, 6, 9};
+  EXPECT_EQ(expected_value, collocations[key]);
+
+  // Exceeds max_rule_symbols.
+  key = {2, -1, 2, -1, 2, 3};
+  EXPECT_EQ(0, collocations.count(key));
+  // Contains non frequent pattern.
+  key = {2, -1, 5};
+  EXPECT_EQ(0, collocations.count(key));
+}
+
+}  // namespace