[폐업추정] String 비교 알고리즘 변경 (#416)

chore: update string similarity comparison algorithm
Stair-Crusher-Club · Nov 6, 2024 · 975e5c7 · 975e5c7
1 parent dd097eb
commit 975e5c7
Show file tree

Hide file tree

Showing 6 changed files with 16 additions and 124 deletions.
diff --git a/app-server/gradle.properties b/app-server/gradle.properties
@@ -25,6 +25,7 @@ guavaVersion=31.1-jre
 sentryVersion=6.7.0
 logbackJsonVersion=0.1.5
 firebaseAdminVersion=9.3.0
+apacheCommonsVersion=1.11.0
 
 # gradle config
 org.gradle.jvmargs=-Xmx8g

diff --git a/app-server/subprojects/bounded_context/external_accessibility/infra/build.gradle.kts b/app-server/subprojects/bounded_context/external_accessibility/infra/build.gradle.kts
@@ -4,19 +4,19 @@ plugins {
 }
 
 dependencies {
+    val awsSdkVersion: String by project
+    val apacheCommonsVersion: String by project
+
     api(projects.apiSpecification.api)
     api(projects.apiSpecification.domainEvent)
     implementation(projects.crossCuttingConcern.infra.persistenceModel)
     implementation("org.springframework.boot:spring-boot-starter-web")
     implementation("io.projectreactor.netty:reactor-netty")
     implementation("org.springframework:spring-webflux")
-    implementation("org.apache.commons:commons-csv:1.11.0")
-
-    integrationTestImplementation(projects.crossCuttingConcern.test.springIt)
-
-    val awsSdkVersion: String by project
+    implementation("org.apache.commons:commons-csv:$apacheCommonsVersion")
     implementation("software.amazon.awssdk:s3:$awsSdkVersion")
     runtimeOnly("software.amazon.awssdk:sts:$awsSdkVersion") // IRSA를 사용하기 위해서 필요함
 
+    integrationTestImplementation(projects.crossCuttingConcern.test.springIt)
     testImplementation(projects.boundedContext.externalAccessibility.application)
 }
diff --git a/.../kotlin/club/staircrusher/place/application/port/in/CreateClosedPlaceCandidatesUseCase.kt b/.../kotlin/club/staircrusher/place/application/port/in/CreateClosedPlaceCandidatesUseCase.kt
@@ -5,6 +5,7 @@ import club.staircrusher.place.application.port.out.web.OpenDataService
 import club.staircrusher.place.domain.model.ClosedPlaceCandidate
 import club.staircrusher.stdlib.persistence.TransactionManager
 import club.staircrusher.stdlib.time.toStartOfDay
+import club.staircrusher.stdlib.util.string.getSimilarityWith
 import mu.KotlinLogging
 import org.springframework.stereotype.Service
 import java.util.UUID
@@ -26,7 +27,7 @@ class CreateClosedPlaceCandidatesUseCase(
             if (nearbyPlaces.isEmpty()) return@mapNotNull null
 
             val placeToSimilarity = nearbyPlaces
-                .associateWith { StringSimilarityComparator.getSimilarity(it.name, closedPlace.name) }
+                .associateWith { it.name.getSimilarityWith(closedPlace.name) }
             val similarPlace = placeToSimilarity
                 .filter { it.value < SIMILARITY_THRESHOLD }
                 .minByOrNull { it.value }
@@ -58,6 +59,6 @@ class CreateClosedPlaceCandidatesUseCase(
 
     companion object {
         private const val SEARCH_RADIUS = 30
-        private const val SIMILARITY_THRESHOLD = 0.3
+        private const val SIMILARITY_THRESHOLD = 0.2
     }
 }
diff --git a/...src/main/kotlin/club/staircrusher/place/application/port/in/StringSimilarityComparator.kt b/...src/main/kotlin/club/staircrusher/place/application/port/in/StringSimilarityComparator.kt
diff --git a/app-server/subprojects/cross_cutting_concern/stdlib/build.gradle.kts b/app-server/subprojects/cross_cutting_concern/stdlib/build.gradle.kts
@@ -7,13 +7,15 @@ dependencies {
     val jtsVersion: String by project
     val proj4jVersion: String by project
     val kopyKatVersion: String by project
+    val apacheCommonsVersion: String by project
     implementation("org.geotools:gt-referencing:$geoToolsVersion")
     implementation("org.locationtech.jts:jts-core:$jtsVersion")
     implementation("org.locationtech.proj4j:proj4j:$proj4jVersion")
     implementation("org.locationtech.proj4j:proj4j-epsg:$proj4jVersion")
     implementation("com.auth0:java-jwt:3.18.1")
     implementation("com.fasterxml.jackson.module:jackson-module-kotlin:2.11.0")
     implementation("org.jetbrains.kotlinx:kotlinx-coroutines-core")
+    implementation("org.apache.commons:commons-text:$apacheCommonsVersion")
 
     api("jakarta.persistence:jakarta.persistence-api")
     api("org.springframework.data:spring-data-commons")

diff --git a/...cutting_concern/stdlib/src/main/kotlin/club/staircrusher/stdlib/util/string/StringUtil.kt b/...cutting_concern/stdlib/src/main/kotlin/club/staircrusher/stdlib/util/string/StringUtil.kt
@@ -1,5 +1,6 @@
 package club.staircrusher.stdlib.util.string
 
+import org.apache.commons.text.similarity.JaroWinklerDistance
 import java.util.*
 
 fun String.emptyToNull() = this.ifBlank { null }
@@ -11,6 +12,10 @@ fun String.isSimilarWith(pattern: String): Boolean {
     )
 }
 
+fun String.getSimilarityWith(other: String): Double {
+    return JaroWinklerDistance().apply(this, other)
+}
+
 private fun simpleMatch(text: String, pattern: String): Boolean {
     var patternIndex = 0
     for (char in text) {