Fix Condalock fetching for cached build layers (#727)

Signed-off-by: munishchouhan <hrma017@gmail.com>
Signed-off-by: Paolo Di Tommaso <paolo.ditommaso@gmail.com>
Co-authored-by: Paolo Di Tommaso <paolo.ditommaso@gmail.com>

src/main/groovy/io/seqera/wave/service/logs/BuildLogServiceImpl.groovy

@@ -152,7 +152,17 @@ class BuildLogServiceImpl implements BuildLogService {
         if( !logs ) return
         try {
             String condaLock = extractCondaLockFile(logs)
-            if (condaLock){
+              /* When a container image is cached, dockerfile does not get executed.
+                 In that case condalock file will contain "cat environment.lock" because its not been executed.
+                 So wave will check the previous builds of that container image
+                 and render the condalock file from latest successful build
+                 and replace with the current build's condalock file.
+               */
+            if( condaLock && condaLock.contains('cat environment.lock') ){
+                condaLock = fetchValidCondaLock(buildId)
+            }
+
+            if ( condaLock ){
                 log.debug "Storing conda lock for buildId: $buildId"
                 final uploadRequest = UploadRequest.fromBytes(condaLock.bytes, condaLockKey(buildId))
                 objectStorageOperations.upload(uploadRequest)
@@ -198,4 +208,32 @@ class BuildLogServiceImpl implements BuildLogService {
                     .replaceAll(/#\d+ \d+\.\d+\s*/, '')
     }
 
+    String fetchValidCondaLock(String buildId) {
+        try {
+            final result = fetchValidCondaLock0(buildId)
+            if( result )
+                log.debug "Container Image is already cached for buildId: $buildId - uploading build's condalock file from buildId: $result"
+            else
+                log.warn "Container Image is already cached for buildId: $buildId - Unable to find condalock file from previous build"
+            return result
+        }
+        catch (Throwable t) {
+            log.error "Unable to determine condalock content for buildId: ${buildId} - cause: ${t.message}", t
+            return null
+        }
+    }
+
+    private String fetchValidCondaLock0(String buildId) {
+        def builds = persistenceService.allBuilds(buildId.split('-')[1].split('_')[0])
+        for (def build : builds) {
+            if ( build.succeeded() ){
+                def curCondaLock = fetchCondaLockString(build.buildId)
+                if( curCondaLock && !curCondaLock.contains('cat environment.lock') ){
+                    return curCondaLock
+                }
+            }
+        }
+        return null
+    }
+
 }

src/test/groovy/io/seqera/wave/service/logs/BuildLogsServiceTest.groovy

@@ -22,13 +22,19 @@ import spock.lang.Specification
 import spock.lang.Unroll
 
 import io.micronaut.objectstorage.InputStreamMapper
+import io.micronaut.objectstorage.ObjectStorageOperations
 import io.micronaut.objectstorage.aws.AwsS3Configuration
+import io.micronaut.objectstorage.aws.AwsS3ObjectStorageEntry
 import io.micronaut.objectstorage.aws.AwsS3Operations
+import io.seqera.wave.service.persistence.PersistenceService
+import io.seqera.wave.service.persistence.WaveBuildRecord
 import io.seqera.wave.test.AwsS3TestContainer
 import software.amazon.awssdk.auth.credentials.AwsBasicCredentials
 import software.amazon.awssdk.auth.credentials.StaticCredentialsProvider
+import software.amazon.awssdk.core.ResponseInputStream
 import software.amazon.awssdk.regions.Region
 import software.amazon.awssdk.services.s3.S3Client
+import software.amazon.awssdk.services.s3.model.GetObjectResponse
 
 /**
  *
@@ -167,4 +173,74 @@ class BuildLogsServiceTest extends Specification implements AwsS3TestContainer {
         noExceptionThrown()
     }
 
+    def 'should return valid conda lock from previous successful build'() {
+        given:
+        def persistenceService = Mock(PersistenceService)
+        def objectStorageOperations = Mock(ObjectStorageOperations)
+        def service = new BuildLogServiceImpl(persistenceService: persistenceService, objectStorageOperations: objectStorageOperations)
+        def build1 = Mock(WaveBuildRecord) {
+            succeeded() >> false
+            buildId >> 'bd-abc_1'
+        }
+        def build2 = Mock(WaveBuildRecord) {
+            succeeded() >> true
+            buildId >> 'bd-abc_2'
+        }
+        def build3 = Mock(WaveBuildRecord) {
+            succeeded() >> true
+            buildId >> 'bd-abc_3'
+        }
+        def responseMetadata = GetObjectResponse.builder()
+                .contentLength(1024L)
+                .contentType("text/plain")
+                .build()
+        def contentStream = new ByteArrayInputStream("valid conda lock".bytes);
+        def responseInputStream = new ResponseInputStream<>(responseMetadata, contentStream);
+        persistenceService.allBuilds(_) >> [build1, build2]
+        objectStorageOperations.retrieve(service.condaLockKey('bd-abc_2')) >> Optional.of(new AwsS3ObjectStorageEntry('bd-abc_2', responseInputStream))
+
+        expect:
+        service.fetchValidCondaLock('bd-abc_3') == 'valid conda lock'
+    }
+
+    def 'should return null when no successful build has valid conda lock'() {
+        given:
+        def persistenceService = Mock(PersistenceService)
+        def objectStorageOperations = Mock(ObjectStorageOperations)
+        def service = new BuildLogServiceImpl(persistenceService: persistenceService, objectStorageOperations: objectStorageOperations)
+        def build1 = Mock(WaveBuildRecord) {
+            succeeded() >> false
+            buildId >> 'bd-abc_1'
+        }
+        def build2 = Mock(WaveBuildRecord) {
+            succeeded() >> true
+            buildId >> 'bd-abc_2'
+        }
+        def build3 = Mock(WaveBuildRecord) {
+            succeeded() >> true
+            buildId >> 'bd-abc_3'
+        }
+        def responseMetadata = GetObjectResponse.builder()
+                .contentLength(1024L)
+                .contentType("text/plain")
+                .build()
+        def contentStream = new ByteArrayInputStream("cat environment.lock".bytes);
+        def responseInputStream = new ResponseInputStream<>(responseMetadata, contentStream);
+        persistenceService.allBuilds(_) >> [build1, build2]
+        objectStorageOperations.retrieve(service.condaLockKey('bd-abc_2')) >> Optional.of(new AwsS3ObjectStorageEntry('bd-abc_2', responseInputStream))
+
+        expect:
+        service.fetchValidCondaLock('bd-abc_3') == null
+    }
+
+    def 'should return null when no builds are available'() {
+        given:
+        def persistenceService = Mock(PersistenceService)
+        def service = new BuildLogServiceImpl(persistenceService: persistenceService)
+        persistenceService.allBuilds(_) >> []
+
+        expect:
+        service.fetchValidCondaLock('bd-abc_1') == null
+    }
+
 }