summaryrefslogtreecommitdiff
path: root/extractor/matchings_sampler_test.cc
diff options
context:
space:
mode:
authorPaul Baltescu <pauldb89@gmail.com>2013-11-27 14:33:36 +0000
committerPaul Baltescu <pauldb89@gmail.com>2013-11-27 14:33:36 +0000
commitd389d25b78e5c99366f49cdcaf788693f3c01c40 (patch)
tree03f6b880eebf2981d5ea8a9a2ec15b10eb812b97 /extractor/matchings_sampler_test.cc
parent7b7167248cd605c9f5bd91b3c87e2826d2336a9f (diff)
Unify sampling backoff strategy.
Diffstat (limited to 'extractor/matchings_sampler_test.cc')
-rw-r--r--extractor/matchings_sampler_test.cc118
1 files changed, 118 insertions, 0 deletions
diff --git a/extractor/matchings_sampler_test.cc b/extractor/matchings_sampler_test.cc
new file mode 100644
index 00000000..bc927152
--- /dev/null
+++ b/extractor/matchings_sampler_test.cc
@@ -0,0 +1,118 @@
+#include <gtest/gtest.h>
+
+#include <memory>
+
+#include "mocks/mock_data_array.h"
+#include "matchings_sampler.h"
+#include "phrase_location.h"
+
+using namespace std;
+using namespace ::testing;
+
+namespace extractor {
+namespace {
+
+class MatchingsSamplerTest : public Test {
+ protected:
+ virtual void SetUp() {
+ vector<int> locations = {0, 1, 2, 3, 4, 5, 6, 7, 8, 9};
+ location = PhraseLocation(locations, 2);
+
+ data_array = make_shared<MockDataArray>();
+ for (int i = 0; i < 10; ++i) {
+ EXPECT_CALL(*data_array, GetSentenceId(i)).WillRepeatedly(Return(i / 2));
+ }
+ }
+
+ unordered_set<int> blacklisted_sentence_ids;
+ PhraseLocation location;
+ shared_ptr<MockDataArray> data_array;
+ shared_ptr<MatchingsSampler> sampler;
+};
+
+TEST_F(MatchingsSamplerTest, TestSample) {
+ sampler = make_shared<MatchingsSampler>(data_array, 1);
+ vector<int> expected_locations = {0, 1};
+ EXPECT_EQ(PhraseLocation(expected_locations, 2),
+ sampler->Sample(location, blacklisted_sentence_ids));
+
+ sampler = make_shared<MatchingsSampler>(data_array, 2);
+ expected_locations = {0, 1, 6, 7};
+ EXPECT_EQ(PhraseLocation(expected_locations, 2),
+ sampler->Sample(location, blacklisted_sentence_ids));
+
+ sampler = make_shared<MatchingsSampler>(data_array, 3);
+ expected_locations = {0, 1, 4, 5, 6, 7};
+ EXPECT_EQ(PhraseLocation(expected_locations, 2),
+ sampler->Sample(location, blacklisted_sentence_ids));
+
+ sampler = make_shared<MatchingsSampler>(data_array, 7);
+ expected_locations = {0, 1, 2, 3, 4, 5, 6, 7, 8, 9};
+ EXPECT_EQ(PhraseLocation(expected_locations, 2),
+ sampler->Sample(location, blacklisted_sentence_ids));
+}
+
+TEST_F(MatchingsSamplerTest, TestBackoffSample) {
+ sampler = make_shared<MatchingsSampler>(data_array, 1);
+ blacklisted_sentence_ids = {0};
+ vector<int> expected_locations = {2, 3};
+ EXPECT_EQ(PhraseLocation(expected_locations, 2),
+ sampler->Sample(location, blacklisted_sentence_ids));
+
+ blacklisted_sentence_ids = {0, 1, 2, 3};
+ expected_locations = {8, 9};
+ EXPECT_EQ(PhraseLocation(expected_locations, 2),
+ sampler->Sample(location, blacklisted_sentence_ids));
+
+ blacklisted_sentence_ids = {0, 1, 2, 3, 4};
+ expected_locations = {};
+ EXPECT_EQ(PhraseLocation(expected_locations, 2),
+ sampler->Sample(location, blacklisted_sentence_ids));
+
+ sampler = make_shared<MatchingsSampler>(data_array, 2);
+ blacklisted_sentence_ids = {0, 3};
+ expected_locations = {2, 3, 4, 5};
+ EXPECT_EQ(PhraseLocation(expected_locations, 2),
+ sampler->Sample(location, blacklisted_sentence_ids));
+
+ sampler = make_shared<MatchingsSampler>(data_array, 3);
+ blacklisted_sentence_ids = {0, 3};
+ expected_locations = {2, 3, 4, 5, 8, 9};
+ EXPECT_EQ(PhraseLocation(expected_locations, 2),
+ sampler->Sample(location, blacklisted_sentence_ids));
+
+ blacklisted_sentence_ids = {0, 2, 3};
+ expected_locations = {2, 3, 8, 9};
+ EXPECT_EQ(PhraseLocation(expected_locations, 2),
+ sampler->Sample(location, blacklisted_sentence_ids));
+
+ sampler = make_shared<MatchingsSampler>(data_array, 4);
+ blacklisted_sentence_ids = {0, 1, 2, 3};
+ expected_locations = {8, 9};
+ EXPECT_EQ(PhraseLocation(expected_locations, 2),
+ sampler->Sample(location, blacklisted_sentence_ids));
+
+ blacklisted_sentence_ids = {1, 3};
+ expected_locations = {0, 1, 4, 5, 8, 9};
+ EXPECT_EQ(PhraseLocation(expected_locations, 2),
+ sampler->Sample(location, blacklisted_sentence_ids));
+
+ sampler = make_shared<MatchingsSampler>(data_array, 7);
+ blacklisted_sentence_ids = {0, 1, 2, 3, 4};
+ expected_locations = {};
+ EXPECT_EQ(PhraseLocation(expected_locations, 2),
+ sampler->Sample(location, blacklisted_sentence_ids));
+
+ blacklisted_sentence_ids = {0, 2, 4};
+ expected_locations = {2, 3, 6, 7};
+ EXPECT_EQ(PhraseLocation(expected_locations, 2),
+ sampler->Sample(location, blacklisted_sentence_ids));
+
+ blacklisted_sentence_ids = {1, 3};
+ expected_locations = {0, 1, 4, 5, 8, 9};
+ EXPECT_EQ(PhraseLocation(expected_locations, 2),
+ sampler->Sample(location, blacklisted_sentence_ids));
+}
+
+}
+} // namespace extractor