feat: remove duplicates

Informasjonsforvaltning · Aug 19, 2024 · 6567820 · 6567820
1 parent 8f67ed0
commit 6567820
Show file tree

Hide file tree

Showing 10 changed files with 349 additions and 45 deletions.
diff --git a/pom.xml b/pom.xml
@@ -131,6 +131,12 @@
             <version>3.9.1</version>
             <scope>test</scope>
         </dependency>
+        <dependency>
+            <groupId>org.apache.httpcomponents.client5</groupId>
+            <artifactId>httpclient5</artifactId>
+            <version>5.3.1</version>
+            <scope>test</scope>
+        </dependency>
         <dependency>
             <groupId>org.testcontainers</groupId>
             <artifactId>testcontainers</artifactId>

diff --git a/...no/digdir/informasjonsforvaltning/fdk_dataservice_harvester/adapter/DataServiceAdapter.kt b/...no/digdir/informasjonsforvaltning/fdk_dataservice_harvester/adapter/DataServiceAdapter.kt
@@ -7,7 +7,6 @@ import org.springframework.stereotype.Service
 import java.io.BufferedReader
 import java.net.HttpURLConnection
 import java.net.URI
-import java.net.URL
 
 private const val TEN_MINUTES = 600000
 

diff --git a/...ir/informasjonsforvaltning/fdk_dataservice_harvester/controller/DataservicesController.kt b/...ir/informasjonsforvaltning/fdk_dataservice_harvester/controller/DataservicesController.kt
@@ -1,5 +1,6 @@
 package no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.controller
 
+import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.model.DuplicateIRI
 import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.rdf.jenaTypeFromAcceptHeader
 import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.service.DataServiceService
 import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.service.EndpointPermissions
@@ -54,4 +55,14 @@ open class DataServicesController(
             ResponseEntity(HttpStatus.NO_CONTENT)
         } else ResponseEntity(HttpStatus.FORBIDDEN)
 
+    @PostMapping("/duplicates")
+    fun removeDuplicates(
+        @AuthenticationPrincipal jwt: Jwt,
+        @RequestBody duplicates: List<DuplicateIRI>
+    ): ResponseEntity<Void> =
+        if (endpointPermissions.hasAdminPermission(jwt)) {
+            dataServiceService.removeDuplicates(duplicates)
+            ResponseEntity(HttpStatus.OK)
+        } else ResponseEntity(HttpStatus.FORBIDDEN)
+
 }
diff --git a/...igdir/informasjonsforvaltning/fdk_dataservice_harvester/harvester/DataServiceHarvester.kt b/...igdir/informasjonsforvaltning/fdk_dataservice_harvester/harvester/DataServiceHarvester.kt
@@ -12,9 +12,6 @@ import org.apache.jena.riot.Lang
 import org.slf4j.LoggerFactory
 import org.springframework.data.repository.findByIdOrNull
 import org.springframework.stereotype.Service
-import java.time.ZoneId
-import java.time.ZonedDateTime
-import java.time.format.DateTimeFormatter
 import java.util.*
 
 private val LOGGER = LoggerFactory.getLogger(DataServiceHarvester::class.java)

diff --git a/.../kotlin/no/digdir/informasjonsforvaltning/fdk_dataservice_harvester/model/DuplicateIRI.kt b/.../kotlin/no/digdir/informasjonsforvaltning/fdk_dataservice_harvester/model/DuplicateIRI.kt
@@ -0,0 +1,7 @@
+package no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.model
+
+data class DuplicateIRI(
+    val iriToRetain: String,
+    val iriToRemove: String,
+    val keepRemovedFdkId: Boolean = true
+)
diff --git a/...no/digdir/informasjonsforvaltning/fdk_dataservice_harvester/service/DataServiceService.kt b/...no/digdir/informasjonsforvaltning/fdk_dataservice_harvester/service/DataServiceService.kt
@@ -1,13 +1,15 @@
 package no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.service
 
 import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.harvester.formatNowWithOsloTimeZone
+import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.model.DuplicateIRI
 import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.model.FdkIdAndUri
 import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.model.HarvestReport
 import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.rabbit.RabbitMQPublisher
 import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.rdf.*
 import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.repository.DataServiceRepository
 import org.apache.jena.rdf.model.ModelFactory
 import org.apache.jena.riot.Lang
+import org.springframework.data.repository.findByIdOrNull
 import org.springframework.http.HttpStatus
 import org.springframework.stereotype.Service
 import org.springframework.web.server.ResponseStatusException
@@ -65,4 +67,45 @@ class DataServiceService(
         }
     }
 
+    fun removeDuplicates(duplicates: List<DuplicateIRI>) {
+        val start = formatNowWithOsloTimeZone()
+        val reportAsRemoved: MutableList<FdkIdAndUri> = mutableListOf()
+
+        duplicates.flatMap { duplicate ->
+            val remove = dataServiceRepository.findByIdOrNull(duplicate.iriToRemove)
+                ?: throw ResponseStatusException(HttpStatus.BAD_REQUEST, "No data service connected to IRI ${duplicate.iriToRemove}")
+
+            val retain = dataServiceRepository.findByIdOrNull(duplicate.iriToRetain)
+                ?.let { if (it.issued > remove.issued) it.copy(issued = remove.issued) else it } // keep earliest issued
+                ?.let { if (it.modified < remove.modified) it.copy(modified = remove.modified) else it } // keep latest modified
+                ?.let {
+                    if (duplicate.keepRemovedFdkId) {
+                        if (it.removed) throw ResponseStatusException(HttpStatus.BAD_REQUEST, "Data service with IRI ${it.uri} has already been removed")
+                        reportAsRemoved.add(FdkIdAndUri(fdkId = it.fdkId, uri = it.uri))
+                        it.copy(fdkId = remove.fdkId)
+                    } else {
+                        if (remove.removed) throw ResponseStatusException(HttpStatus.BAD_REQUEST, "Data service with IRI ${remove.uri} has already been removed")
+                        reportAsRemoved.add(FdkIdAndUri(fdkId = remove.fdkId, uri = remove.uri))
+                        it
+                    }
+                }
+                ?: remove.copy(uri = duplicate.iriToRetain)
+
+            listOf(remove.copy(removed = true), retain.copy(removed = false))
+        }.run { dataServiceRepository.saveAll(this) }
+
+        if (reportAsRemoved.isNotEmpty()) {
+            rabbitPublisher.send(listOf(
+                HarvestReport(
+                    id = "duplicate-delete",
+                    url = "https://fellesdatakatalog.digdir.no/duplicates",
+                    harvestError = false,
+                    startTime = start,
+                    endTime = formatNowWithOsloTimeZone(),
+                    removedResources = reportAsRemoved
+                )
+            ))
+        }
+    }
+
 }
diff --git a/src/main/resources/application.yml b/src/main/resources/application.yml
@@ -6,6 +6,7 @@ logging:
   level.org.apache.jena.riot: ERROR
 server:
   port: 8080
+  error.include-message: always
 application:
   dataserviceUri: ${FDK_DATASERVICE_HARVESTER_URI:https://dataservices.staging.fellesdatakatalog.digdir.no}/dataservices
   catalogUri: ${FDK_DATASERVICE_HARVESTER_URI:https://dataservices.staging.fellesdatakatalog.digdir.no}/catalogs

diff --git a/...digdir/informasjonsforvaltning/fdk_dataservice_harvester/contract/DataServicesContract.kt b/...digdir/informasjonsforvaltning/fdk_dataservice_harvester/contract/DataServicesContract.kt
@@ -1,7 +1,11 @@
 package no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.contract
 
+import com.fasterxml.jackson.module.kotlin.jacksonObjectMapper
+import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.model.DuplicateIRI
 import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.utils.ApiTestContext
 import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.utils.DATASERVICE_ID_0
+import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.utils.DATA_SERVICE_DBO_0
+import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.utils.DATA_SERVICE_DBO_1
 import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.utils.TestResponseReader
 import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.utils.apiGet
 import no.digdir.informasjonsforvaltning.fdk_dataservice_harvester.utils.authorizedRequest
@@ -15,18 +19,21 @@ import org.junit.jupiter.api.Tag
 import org.junit.jupiter.api.Test
 import org.junit.jupiter.api.TestInstance
 import org.springframework.boot.test.context.SpringBootTest
+import org.springframework.http.HttpMethod
 import org.springframework.http.HttpStatus
 import org.springframework.test.context.ContextConfiguration
 import kotlin.test.assertTrue
 
 @TestInstance(TestInstance.Lifecycle.PER_CLASS)
 @SpringBootTest(
     properties = ["spring.profiles.active=contract-test"],
-    webEnvironment = SpringBootTest.WebEnvironment.RANDOM_PORT)
+    webEnvironment = SpringBootTest.WebEnvironment.RANDOM_PORT
+)
 @ContextConfiguration(initializers = [ApiTestContext.Initializer::class])
 @Tag("contract")
 class DataServicesContract : ApiTestContext() {
     private val responseReader = TestResponseReader()
+    private val mapper = jacksonObjectMapper()
 
     @Test
     fun findSpecific() {
@@ -61,7 +68,12 @@ class DataServicesContract : ApiTestContext() {
 
         @Test
         fun unauthorizedForNoToken() {
-            val response = authorizedRequest(port, "/dataservices/$DATASERVICE_ID_0", null, "DELETE")
+            val response = authorizedRequest(
+                port,
+                "/dataservices/$DATASERVICE_ID_0",
+                null,
+                HttpMethod.DELETE
+            )
             assertEquals(HttpStatus.UNAUTHORIZED.value(), response["status"])
         }
 
@@ -71,15 +83,19 @@ class DataServicesContract : ApiTestContext() {
                 port,
                 "/dataservices/$DATASERVICE_ID_0",
                 JwtToken(Access.ORG_WRITE).toString(),
-                "DELETE"
+                HttpMethod.DELETE
             )
             assertEquals(HttpStatus.FORBIDDEN.value(), response["status"])
         }
 
         @Test
         fun notFoundWhenIdNotInDB() {
-            val response =
-                authorizedRequest(port, "/dataservices/123", JwtToken(Access.ROOT).toString(), "DELETE")
+            val response = authorizedRequest(
+                port,
+                "/dataservices/123",
+                JwtToken(Access.ROOT).toString(),
+                HttpMethod.DELETE
+            )
             assertEquals(HttpStatus.NOT_FOUND.value(), response["status"])
         }
 
@@ -89,10 +105,67 @@ class DataServicesContract : ApiTestContext() {
                 port,
                 "/dataservices/$DATASERVICE_ID_0",
                 JwtToken(Access.ROOT).toString(),
-                "DELETE"
+                HttpMethod.DELETE
             )
             assertEquals(HttpStatus.NO_CONTENT.value(), response["status"])
         }
     }
 
+    @Nested
+    internal inner class RemoveDuplicates {
+
+        @Test
+        fun unauthorizedForNoToken() {
+            val body = listOf(DuplicateIRI(iriToRemove = DATA_SERVICE_DBO_0.uri, iriToRetain = DATA_SERVICE_DBO_1.uri))
+            val response = authorizedRequest(
+                port,
+                "/dataservices/duplicates",
+                null,
+                HttpMethod.POST,
+                mapper.writeValueAsString(body)
+            )
+            assertEquals(HttpStatus.UNAUTHORIZED.value(), response["status"])
+        }
+
+        @Test
+        fun forbiddenWithNonSysAdminRole() {
+            val body = listOf(DuplicateIRI(iriToRemove = DATA_SERVICE_DBO_0.uri, iriToRetain = DATA_SERVICE_DBO_1.uri))
+            val response = authorizedRequest(
+                port,
+                "/dataservices/duplicates",
+                JwtToken(Access.ORG_WRITE).toString(),
+                HttpMethod.POST,
+                mapper.writeValueAsString(body)
+            )
+            assertEquals(HttpStatus.FORBIDDEN.value(), response["status"])
+        }
+
+        @Test
+        fun badRequestWhenRemoveIRINotInDB() {
+            val body = listOf(DuplicateIRI(iriToRemove = "https://123.no", iriToRetain = DATA_SERVICE_DBO_1.uri))
+            val response =
+                authorizedRequest(
+                    port,
+                    "/dataservices/duplicates",
+                    JwtToken(Access.ROOT).toString(),
+                    HttpMethod.POST,
+                    mapper.writeValueAsString(body)
+                )
+            assertEquals(HttpStatus.BAD_REQUEST.value(), response["status"])
+        }
+
+        @Test
+        fun okWithSysAdminRole() {
+            val body = listOf(DuplicateIRI(iriToRemove = DATA_SERVICE_DBO_0.uri, iriToRetain = DATA_SERVICE_DBO_1.uri))
+            val response = authorizedRequest(
+                port,
+                "/dataservices/duplicates",
+                JwtToken(Access.ROOT).toString(),
+                HttpMethod.POST,
+                mapper.writeValueAsString(body)
+            )
+            assertEquals(HttpStatus.OK.value(), response["status"])
+        }
+    }
+
 }