From 671c21451542e2dd20e45b4033d44d8e8735f87b Mon Sep 17 00:00:00 2001 From: Chris Dyer Date: Thu, 3 Dec 2009 16:33:55 -0500 Subject: initial check in --- vest/scorer_test.cc | 178 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 178 insertions(+) create mode 100644 vest/scorer_test.cc (limited to 'vest/scorer_test.cc') diff --git a/vest/scorer_test.cc b/vest/scorer_test.cc new file mode 100644 index 00000000..fca219cc --- /dev/null +++ b/vest/scorer_test.cc @@ -0,0 +1,178 @@ +#include +#include +#include +#include + +#include "tdict.h" +#include "scorer.h" + +using namespace std; + +class ScorerTest : public testing::Test { + protected: + virtual void SetUp() { + refs0.resize(4); + refs1.resize(4); + TD::ConvertSentence("export of high-tech products in guangdong in first two months this year reached 3.76 billion us dollars", &refs0[0]); + TD::ConvertSentence("guangdong's export of new high technology products amounts to us $ 3.76 billion in first two months of this year", &refs0[1]); + TD::ConvertSentence("guangdong exports us $ 3.76 billion worth of high technology products in the first two months of this year", &refs0[2]); + TD::ConvertSentence("in the first 2 months this year , the export volume of new hi-tech products in guangdong province reached 3.76 billion us dollars .", &refs0[3]); + TD::ConvertSentence("xinhua news agency , guangzhou , march 16 ( reporter chen ji ) the latest statistics show that from january through february this year , the export of high-tech products in guangdong province reached 3.76 billion us dollars , up 34.8 \% over the same period last year and accounted for 25.5 \% of the total export in the province .", &refs1[0]); + TD::ConvertSentence("xinhua news agency , guangzhou , march 16 ( reporter : chen ji ) -- latest statistic indicates that guangdong's export of new high technology products amounts to us $ 3.76 billion , up 34.8 \% over corresponding period and accounts for 25.5 \% of the total exports of the province .", &refs1[1]); + TD::ConvertSentence("xinhua news agency report of march 16 from guangzhou ( by staff reporter chen ji ) - latest statistics indicate guangdong province exported us $ 3.76 billion worth of high technology products , up 34.8 percent from the same period last year , which account for 25.5 percent of the total exports of the province .", &refs1[2]); + TD::ConvertSentence("guangdong , march 16 , ( xinhua ) -- ( chen ji reports ) as the newest statistics shows , in january and feberuary this year , the export volume of new hi-tech products in guangdong province reached 3.76 billion us dollars , up 34.8 \% than last year , making up 25.5 \% of the province's total .", &refs1[3]); + TD::ConvertSentence("one guangdong province will next export us $ 3.76 high-tech product two months first this year 3.76 billion us dollars", &hyp1); + TD::ConvertSentence("xinhua news agency , guangzhou , 16th of march ( reporter chen ) -- latest statistics suggest that guangdong exports new advanced technology product totals $ 3.76 million , 34.8 percent last corresponding period and accounts for 25.5 percent of the total export province .", &hyp2); + } + + virtual void TearDown() { } + + vector > refs0; + vector > refs1; + vector hyp1; + vector hyp2; +}; + +TEST_F(ScorerTest, TestCreateFromFiles) { + vector files; + files.push_back("test_data/re.txt.0"); + files.push_back("test_data/re.txt.1"); + files.push_back("test_data/re.txt.2"); + files.push_back("test_data/re.txt.3"); + DocScorer ds(IBM_BLEU, files); +} + +TEST_F(ScorerTest, TestBLEUScorer) { + SentenceScorer* s1 = SentenceScorer::CreateSentenceScorer(IBM_BLEU, refs0); + SentenceScorer* s2 = SentenceScorer::CreateSentenceScorer(IBM_BLEU, refs1); + Score* b1 = s1->ScoreCandidate(hyp1); + EXPECT_FLOAT_EQ(0.23185077, b1->ComputeScore()); + Score* b2 = s2->ScoreCandidate(hyp2); + EXPECT_FLOAT_EQ(0.38101241, b2->ComputeScore()); + b1->PlusEquals(*b2); + EXPECT_FLOAT_EQ(0.348854, b1->ComputeScore()); + EXPECT_FALSE(b1->IsAdditiveIdentity()); + string details; + b1->ScoreDetails(&details); + EXPECT_EQ("BLEU = 34.89, 81.5|50.8|29.5|18.6 (brev=0.898)", details); + cerr << details << endl; + string enc; + b1->Encode(&enc); + Score* b3 = SentenceScorer::CreateScoreFromString(IBM_BLEU, enc); + details.clear(); + cerr << "Encoded BLEU score size: " << enc.size() << endl; + b3->ScoreDetails(&details); + cerr << details << endl; + EXPECT_FALSE(b3->IsAdditiveIdentity()); + EXPECT_EQ("BLEU = 34.89, 81.5|50.8|29.5|18.6 (brev=0.898)", details); + Score* bz = b3->GetZero(); + EXPECT_TRUE(bz->IsAdditiveIdentity()); + delete bz; + delete b1; + delete s1; + delete b2; + delete s2; +} + +TEST_F(ScorerTest, TestTERScorer) { + SentenceScorer* s1 = SentenceScorer::CreateSentenceScorer(TER, refs0); + SentenceScorer* s2 = SentenceScorer::CreateSentenceScorer(TER, refs1); + string details; + Score* t1 = s1->ScoreCandidate(hyp1); + t1->ScoreDetails(&details); + cerr << "DETAILS: " << details << endl; + cerr << t1->ComputeScore() << endl; + Score* t2 = s2->ScoreCandidate(hyp2); + t2->ScoreDetails(&details); + cerr << "DETAILS: " << details << endl; + cerr << t2->ComputeScore() << endl; + t1->PlusEquals(*t2); + cerr << t1->ComputeScore() << endl; + t1->ScoreDetails(&details); + cerr << "DETAILS: " << details << endl; + EXPECT_EQ("TER = 44.16, 4| 8| 16| 6 (len=77)", details); + string enc; + t1->Encode(&enc); + Score* t3 = SentenceScorer::CreateScoreFromString(TER, enc); + details.clear(); + t3->ScoreDetails(&details); + EXPECT_EQ("TER = 44.16, 4| 8| 16| 6 (len=77)", details); + EXPECT_FALSE(t3->IsAdditiveIdentity()); + Score* tz = t3->GetZero(); + EXPECT_TRUE(tz->IsAdditiveIdentity()); + delete tz; + delete t3; + delete t1; + delete s1; + delete t2; + delete s2; +} + +TEST_F(ScorerTest, TestTERScorerSimple) { + vector > ref(1); + TD::ConvertSentence("1 2 3 A B", &ref[0]); + vector hyp; + TD::ConvertSentence("A B 1 2 3", &hyp); + SentenceScorer* s1 = SentenceScorer::CreateSentenceScorer(TER, ref); + string details; + Score* t1 = s1->ScoreCandidate(hyp); + t1->ScoreDetails(&details); + cerr << "DETAILS: " << details << endl; + delete t1; + delete s1; +} + +TEST_F(ScorerTest, TestSERScorerSimple) { + vector > ref(1); + TD::ConvertSentence("A B C D", &ref[0]); + vector hyp1; + TD::ConvertSentence("A B C", &hyp1); + vector hyp2; + TD::ConvertSentence("A B C D", &hyp2); + SentenceScorer* s1 = SentenceScorer::CreateSentenceScorer(SER, ref); + string details; + Score* t1 = s1->ScoreCandidate(hyp1); + t1->ScoreDetails(&details); + cerr << "DETAILS: " << details << endl; + Score* t2 = s1->ScoreCandidate(hyp2); + t2->ScoreDetails(&details); + cerr << "DETAILS: " << details << endl; + t2->PlusEquals(*t1); + t2->ScoreDetails(&details); + cerr << "DETAILS: " << details << endl; + delete t1; + delete t2; + delete s1; +} + +TEST_F(ScorerTest, TestCombiScorer) { + SentenceScorer* s1 = SentenceScorer::CreateSentenceScorer(BLEU_minus_TER_over_2, refs0); + string details; + Score* t1 = s1->ScoreCandidate(hyp1); + t1->ScoreDetails(&details); + cerr << "DETAILS: " << details << endl; + cerr << t1->ComputeScore() << endl; + string enc; + t1->Encode(&enc); + Score* t2 = SentenceScorer::CreateScoreFromString(BLEU_minus_TER_over_2, enc); + details.clear(); + t2->ScoreDetails(&details); + cerr << "DETAILS: " << details << endl; + Score* cz = t2->GetZero(); + EXPECT_FALSE(t2->IsAdditiveIdentity()); + EXPECT_TRUE(cz->IsAdditiveIdentity()); + cz->PlusEquals(*t2); + EXPECT_FALSE(cz->IsAdditiveIdentity()); + string d2; + cz->ScoreDetails(&d2); + EXPECT_EQ(d2, details); + delete cz; + delete t2; + delete t1; +} + +int main(int argc, char **argv) { + testing::InitGoogleTest(&argc, argv); + return RUN_ALL_TESTS(); +} + -- cgit v1.2.3