Merge pull request #38 from codingpot/fix-floating-point-comparisons

fix(test): use go-cmp for robust floating-point comparisons

Author: mrchypark

.github/workflows/ci.yaml

@@ -9,16 +9,17 @@ jobs:
     strategy:
       matrix:
         go-version:
-          - '1.17'
+          - '1.23'
         os:
           - ubuntu-latest
           - macos-latest
     runs-on: ${{ matrix.os }}
     steps:
       - name: Install Go ${{ matrix.go-version }}
-        uses: actions/setup-go@v2
+        uses: actions/setup-go@v5
         with:
           go-version: ${{ matrix.go-version }}
-      - uses: actions/checkout@v2
+      - uses: actions/checkout@v4
       - run: make install-kiwi
       - run: make test
+

.gitignore

@@ -20,4 +20,6 @@ __debug_bin
 base/
 .idea/
 models/
-include/
+include/
+ModelGenerator/
+

Makefile

@@ -23,3 +23,4 @@ clean:
 format:
 	# go install mvdan.cc/gofumpt@latest
 	gofumpt -l -w .
+

go.mod

@@ -1,11 +1,15 @@
 module github.com/codingpot/kiwigo
 
-go 1.17
+go 1.23
 
-require github.com/stretchr/testify v1.7.0
+require (
+	github.com/google/go-cmp v0.6.0
+	github.com/stretchr/testify v1.7.0
+)
 
 require (
 	github.com/davecgh/go-spew v1.1.0 // indirect
 	github.com/pmezard/go-difflib v1.0.0 // indirect
 	gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c // indirect
 )
+

go.sum

@@ -1,5 +1,7 @@
 github.com/davecgh/go-spew v1.1.0 h1:ZDRjVQ15GmhC3fiQ8ni8+OwkZQO4DARzQgrnXU1Liz8=
 github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
+github.com/google/go-cmp v0.6.0 h1:ofyhxvXcZhMsU5ulbFiLKl/XBFqE1GSq7atu8tAmTRI=
+github.com/google/go-cmp v0.6.0/go.mod h1:17dUlkBOakJ0+DkrSSNjCkIjxS6bF9zb3elmeNGIjoY=
 github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
 github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
 github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=

kiwi.go

@@ -153,7 +153,7 @@ type SplitResult struct {
 
 // SplitSentence returns the line of sentences.
 func (k *Kiwi) SplitSentence(text string, options AnalyzeOption) ([]SplitResult, error) {
-	var cText = C.CString(text)
+	cText := C.CString(text)
 	defer C.free(unsafe.Pointer(cText))
 
 	kiwiSsH := C.kiwi_split_into_sents(k.handler, cText, C.int(options), nil)

kiwi_example_test.go

@@ -14,7 +14,11 @@ func Example() {
 	defer k.Close() // don't forget to Close()!
 
 	results, _ := k.Analyze("안녕하세요 코딩냄비입니다. 부글부글.", 1 /*=topN*/, kiwi.KIWI_MATCH_ALL)
-	fmt.Println(results)
+
+	// Print tokens without the score to avoid floating-point output issues
+	if len(results) > 0 {
+		fmt.Printf("Tokens: %v\n", results[0].Tokens)
+	}
 	// Output:
-	// [{[{0 NNG 안녕} {2 XSA 하} {3 EF 세요} {6 NNP 코딩냄비} {10 VCP 이} {10 EF ᆸ니다} {13 SF .} {15 MAG 부글부글} {19 SF .}] -55.869953}]
+	// Tokens: [{0 NNG 안녕} {2 XSA 하} {3 EF 세요} {6 NNP 코딩냄비} {10 VCP 이} {10 EF ᆸ니다} {13 SF .} {15 MAG 부글부글} {19 SF .}]
 }

kiwi_test.go

@@ -5,9 +5,16 @@ import (
 	"strings"
 	"testing"
 
+	"github.com/google/go-cmp/cmp"
+	"github.com/google/go-cmp/cmp/cmpopts"
 	"github.com/stretchr/testify/assert"
 )
 
+// floatComparer returns a cmp.Option for floating-point comparisons with tolerance.
+func floatComparer() cmp.Option {
+	return cmpopts.EquateApprox(0, 1e-5)
+}
+
 func TestKiwiVersion(t *testing.T) {
 	assert.Equal(t, KiwiVersion(), "0.21.0")
 }
@@ -59,7 +66,9 @@ func TestAnalyze(t *testing.T) {
 		},
 	}
 
-	assert.Equal(t, expected, res)
+	if diff := cmp.Diff(expected, res, floatComparer()); diff != "" {
+		t.Errorf("Analyze result mismatch (-want +got):\n%s", diff)
+	}
 	assert.Equal(t, 0, kiwi.Close())
 }
 
@@ -143,7 +152,9 @@ func TestAddWord(t *testing.T) {
 		},
 	}
 
-	assert.Equal(t, expected, res)
+	if diff := cmp.Diff(expected, res, floatComparer()); diff != "" {
+		t.Errorf("AddWord result mismatch (-want +got):\n%s", diff)
+	}
 	assert.Equal(t, 0, kiwi.Close())
 }
 
@@ -202,7 +213,9 @@ func TestLoadDict(t *testing.T) {
 		},
 	}
 
-	assert.Equal(t, expected, res)
+	if diff := cmp.Diff(expected, res, floatComparer()); diff != "" {
+		t.Errorf("LoadDict result mismatch (-want +got):\n%s", diff)
+	}
 	assert.Equal(t, 0, kiwi.Close())
 }
 
@@ -242,7 +255,9 @@ func TestLoadDict2(t *testing.T) {
 		},
 	}
 
-	assert.Equal(t, expected, res)
+	if diff := cmp.Diff(expected, res, floatComparer()); diff != "" {
+		t.Errorf("LoadDict2 result mismatch (-want +got):\n%s", diff)
+	}
 	assert.Equal(t, 0, kiwi.Close())
 }
 
@@ -258,7 +273,7 @@ func TestExtractWord(t *testing.T) {
 가능성이 크다. 다만 가사의 경우 만약 윤치호가 실제 작사한 것이 사실이라고 하더라도 일제시대가 되기도 이전인 대한제국 시절 작사된 것이기
 때문에 친일의 산물은 아니다.`)
 	wordInfos, _ := kb.ExtractWords(rs, 3 /*=minCnt*/, 3 /*=maxWordLen*/, 0.0 /*=minScore*/, -3.0 /*=posThreshold*/)
-	assert.Equal(t, []WordInfo{
+	expected := []WordInfo{
 		{
 			Form:     "안익",
 			Freq:     3,
@@ -271,7 +286,10 @@ func TestExtractWord(t *testing.T) {
 			POSScore: -0.23702252,
 			Score:    0,
 		},
-	}, wordInfos)
+	}
+	if diff := cmp.Diff(expected, wordInfos, floatComparer()); diff != "" {
+		t.Errorf("ExtractWord result mismatch (-want +got):\n%s", diff)
+	}
 	assert.Equal(t, 0, kb.Close())
 }
 
@@ -280,8 +298,11 @@ func TestExtractWordwithFile(t *testing.T) {
 	file, _ := os.Open("./example/test.txt")
 
 	wordInfos, _ := kb.ExtractWords(file, 10 /*=minCnt*/, 5 /*=maxWordLen*/, 0.0 /*=minScore*/, -25.0 /*=posThreshold*/)
-	assert.Equal(t, WordInfo{
+	expectedWordInfo := WordInfo{
 		Form: "무위원", Freq: 17, POSScore: -1.7342134, Score: 0.69981515,
-	}, wordInfos[0])
+	}
+	if diff := cmp.Diff(expectedWordInfo, wordInfos[0], floatComparer()); diff != "" {
+		t.Errorf("ExtractWordwithFile result mismatch (-want +got):\n%s", diff)
+	}
 	assert.Equal(t, 0, kb.Close())
 }