Show how OIDC ModelAuthProvider can be used with Azure OpenAI

core/runtime/src/main/java/io/quarkiverse/langchain4j/auth/ModelAuthProvider.java

@@ -22,15 +22,25 @@ interface Input {
     }
 
     static Optional<ModelAuthProvider> resolve(String modelName) {
-        Instance<ModelAuthProvider> beanInstance = modelName == null
-                ? CDI.current().select(ModelAuthProvider.class)
-                : CDI.current().select(ModelAuthProvider.class, ModelName.Literal.of(modelName));
-
-        //get the first one without causing a bean1 resolution exception
+        // This will likely need to be refactored again.
+        // ModelAuthProvider should return a set of supported models (empty by default),
+        // otherwise the resolution on the main branch does not work for more than one OIDC provider
         ModelAuthProvider authorizer = null;
-        for (var handle : beanInstance.handles()) {
-            authorizer = handle.get();
-            break;
+        if (modelName != null) {
+            Instance<ModelAuthProvider> beanInstance = CDI.current().select(ModelAuthProvider.class,
+                    ModelName.Literal.of(modelName));
+
+            for (var handle : beanInstance.handles()) {
+                authorizer = handle.get();
+                break;
+            }
+        }
+        if (authorizer == null) {
+            Instance<ModelAuthProvider> beanInstance = CDI.current().select(ModelAuthProvider.class);
+            for (var handle : beanInstance.handles()) {
+                authorizer = handle.get();
+                break;
+            }
         }
         return Optional.ofNullable(authorizer);
     }

embedding-stores/pgvector/deployment/src/main/java/io/quarkiverse/langchain4j/pgvector/deployment/DevServicesConfigBuilderCustomizer.java

@@ -12,9 +12,10 @@
 public class DevServicesConfigBuilderCustomizer implements SmallRyeConfigBuilderCustomizer {
     @Override
     public void configBuilder(final SmallRyeConfigBuilder builder) {
-        // use a priority of 50 to make sure that this is overridable by any of the standard methods
-        builder.withSources(
-                new PropertiesConfigSource(Map.of("quarkus.datasource.devservices.image-name", "pgvector/pgvector:pg17"),
-                        "quarkus-langchain4j-pgvector", 50));
+        // use a priority of 50 to make sure that this is overridable by any of the
+        // standard methods
+        builder.withSources(new PropertiesConfigSource(
+                Map.of("quarkus.datasource.devservices.image-name", "pgvector/pgvector:pg17"),
+                "quarkus-langchain4j-pgvector", 50));
     }
 }

embedding-stores/pgvector/deployment/src/main/java/io/quarkiverse/langchain4j/pgvector/deployment/PgVectorEmbeddingStoreBuildTimeConfig.java

@@ -12,8 +12,8 @@
 public interface PgVectorEmbeddingStoreBuildTimeConfig {
 
     /**
-     * The name of the configured Postgres datasource to use for this store. If not set,
-     * the default datasource from the Agroal extension will be used.
+     * The name of the configured Postgres datasource to use for this store. If not
+     * set, the default datasource from the Agroal extension will be used.
      */
     Optional<String> datasource();
 

embedding-stores/pgvector/deployment/src/main/java/io/quarkiverse/langchain4j/pgvector/deployment/PgVectorEmbeddingStoreProcessor.java

@@ -33,7 +33,8 @@ class PgVectorEmbeddingStoreProcessor {
 
     private static final DotName PG_VECTOR_EMBEDDING_STORE = DotName.createSimple(PgVectorEmbeddingStore.class);
     private static final DotName AGROAL_POOL_INTERCEPTOR = DotName.createSimple(AgroalPoolInterceptor.class);
-    private static final DotName PG_VECTOR_AGROAL_POOL_INTERCEPTOR = DotName.createSimple(PgVectorAgroalPoolInterceptor.class);
+    private static final DotName PG_VECTOR_AGROAL_POOL_INTERCEPTOR = DotName
+            .createSimple(PgVectorAgroalPoolInterceptor.class);
 
     private static final String FEATURE = "langchain4j-pgvector";
 
@@ -50,39 +51,27 @@ void indexDependencies(BuildProducer<IndexDependencyBuildItem> producer) {
 
     @BuildStep
     @Record(ExecutionTime.RUNTIME_INIT)
-    public void createBean(
-            BuildProducer<SyntheticBeanBuildItem> beanProducer,
-            PgVectorEmbeddingStoreRecorder recorder,
-            PgVectorEmbeddingStoreConfig config,
-            PgVectorEmbeddingStoreBuildTimeConfig buildTimeConfig,
+    public void createBean(BuildProducer<SyntheticBeanBuildItem> beanProducer, PgVectorEmbeddingStoreRecorder recorder,
+            PgVectorEmbeddingStoreConfig config, PgVectorEmbeddingStoreBuildTimeConfig buildTimeConfig,
             BuildProducer<EmbeddingStoreBuildItem> embeddingStoreProducer) {
 
         AnnotationInstance datasourceQualifier = buildTimeConfig.datasource()
                 .map(dn -> AnnotationInstance.builder(DataSource.class).add("value", dn).build())
                 .orElse(AnnotationInstance.builder(Default.class).build());
 
-        beanProducer.produce(SyntheticBeanBuildItem
-                .configure(PG_VECTOR_EMBEDDING_STORE)
+        beanProducer.produce(SyntheticBeanBuildItem.configure(PG_VECTOR_EMBEDDING_STORE)
                 .types(ClassType.create(dev.langchain4j.store.embedding.pgvector.PgVectorEmbeddingStore.class),
                         ClassType.create(EmbeddingStore.class),
                         ParameterizedType.create(EmbeddingStore.class, ClassType.create(TextSegment.class)))
-                .setRuntimeInit()
-                .defaultBean()
-                .unremovable()
-                .scope(ApplicationScoped.class)
+                .setRuntimeInit().defaultBean().unremovable().scope(ApplicationScoped.class)
                 .createWith(recorder.embeddingStoreFunction(config, buildTimeConfig.datasource().orElse(null)))
                 .addInjectionPoint(ClassType.create(DotName.createSimple(AgroalDataSource.class)), datasourceQualifier)
                 .done());
 
-        beanProducer.produce(SyntheticBeanBuildItem
-                .configure(PG_VECTOR_AGROAL_POOL_INTERCEPTOR)
-                .types(ClassType.create(AGROAL_POOL_INTERCEPTOR))
-                .setRuntimeInit()
-                .unremovable()
-                .scope(ApplicationScoped.class)
-                .supplier(recorder.pgVectorAgroalPoolInterceptor())
-                .qualifiers(datasourceQualifier)
-                .done());
+        beanProducer.produce(SyntheticBeanBuildItem.configure(PG_VECTOR_AGROAL_POOL_INTERCEPTOR)
+                .types(ClassType.create(AGROAL_POOL_INTERCEPTOR)).setRuntimeInit().unremovable()
+                .scope(ApplicationScoped.class).supplier(recorder.pgVectorAgroalPoolInterceptor())
+                .qualifiers(datasourceQualifier).done());
 
         embeddingStoreProducer.produce(new EmbeddingStoreBuildItem());
     }

embedding-stores/pgvector/deployment/src/test/java/io/quarkiverse/langchain4j/pgvector/test/ColumnsTest.java

@@ -13,15 +13,14 @@ class ColumnsTest extends LangChain4jPgVectorBaseTest {
     @RegisterExtension
     static final QuarkusUnitTest test = new QuarkusUnitTest()
             .setArchiveProducer(() -> ShrinkWrap.create(JavaArchive.class)
-                    .addAsResource(new StringAsset(
-                            "quarkus.langchain4j.pgvector.dimension=384\n" +
-                                    "quarkus.langchain4j.pgvector.drop-table-first=true\n" +
-                                    "quarkus.class-loading.parent-first-artifacts=ai.djl.huggingface:tokenizers\n" +
-                                    "quarkus.log.category.\"io.quarkiverse.langchain4j.pgvector\".level=DEBUG\n\n" +
-                                    "quarkus.langchain4j.pgvector.metadata.storage-mode=COLUMN_PER_KEY\n" +
-                                    "quarkus.langchain4j.pgvector.metadata.column-definitions=key text NULL, name text NULL, " +
-                                    "age float NULL, city varchar null, country varchar null\n" +
-                                    "quarkus.langchain4j.pgvector.metadata.indexes=key, name, age"),
+                    .addAsResource(new StringAsset("quarkus.langchain4j.pgvector.dimension=384\n"
+                            + "quarkus.langchain4j.pgvector.drop-table-first=true\n"
+                            + "quarkus.class-loading.parent-first-artifacts=ai.djl.huggingface:tokenizers\n"
+                            + "quarkus.log.category.\"io.quarkiverse.langchain4j.pgvector\".level=DEBUG\n\n"
+                            + "quarkus.langchain4j.pgvector.metadata.storage-mode=COLUMN_PER_KEY\n"
+                            + "quarkus.langchain4j.pgvector.metadata.column-definitions=key text NULL, name text NULL, "
+                            + "age float NULL, city varchar null, country varchar null\n"
+                            + "quarkus.langchain4j.pgvector.metadata.indexes=key, name, age"),
                             "application.properties"));
 
     @Test

embedding-stores/pgvector/deployment/src/test/java/io/quarkiverse/langchain4j/pgvector/test/JSONBMultiIndexTest.java

@@ -10,18 +10,15 @@
 public class JSONBMultiIndexTest extends LangChain4jPgVectorBaseTest {
 
     @RegisterExtension
-    static final QuarkusUnitTest test = new QuarkusUnitTest()
-            .setArchiveProducer(() -> ShrinkWrap.create(JavaArchive.class)
-                    .addAsResource(new StringAsset(
-                            "quarkus.langchain4j.pgvector.dimension=384\n" +
-                                    "quarkus.langchain4j.pgvector.drop-table-first=true\n" +
-                                    "quarkus.class-loading.parent-first-artifacts=ai.djl.huggingface:tokenizers\n" +
-                                    "quarkus.log.category.\"io.quarkiverse.langchain4j.pgvector\".level=DEBUG\n\n" +
-                                    "quarkus.langchain4j.pgvector.metadata.storage-mode=COMBINED_JSONB\n" +
-                                    "quarkus.langchain4j.pgvector.metadata.column-definitions=metadata_b JSONB NULL\n" +
-                                    "quarkus.langchain4j.pgvector.metadata.indexes=(metadata_b->'key'), (metadata_b->'name'), (metadata_b->'age')\n"
-                                    +
-                                    "quarkus.langchain4j.pgvector.metadata.index-type=GIN"),
-                            "application.properties"));
+    static final QuarkusUnitTest test = new QuarkusUnitTest().setArchiveProducer(() -> ShrinkWrap
+            .create(JavaArchive.class)
+            .addAsResource(new StringAsset("quarkus.langchain4j.pgvector.dimension=384\n"
+                    + "quarkus.langchain4j.pgvector.drop-table-first=true\n"
+                    + "quarkus.class-loading.parent-first-artifacts=ai.djl.huggingface:tokenizers\n"
+                    + "quarkus.log.category.\"io.quarkiverse.langchain4j.pgvector\".level=DEBUG\n\n"
+                    + "quarkus.langchain4j.pgvector.metadata.storage-mode=COMBINED_JSONB\n"
+                    + "quarkus.langchain4j.pgvector.metadata.column-definitions=metadata_b JSONB NULL\n"
+                    + "quarkus.langchain4j.pgvector.metadata.indexes=(metadata_b->'key'), (metadata_b->'name'), (metadata_b->'age')\n"
+                    + "quarkus.langchain4j.pgvector.metadata.index-type=GIN"), "application.properties"));
 
 }

embedding-stores/pgvector/deployment/src/test/java/io/quarkiverse/langchain4j/pgvector/test/JSONBTest.java

@@ -10,16 +10,14 @@
 public class JSONBTest extends LangChain4jPgVectorBaseTest {
 
     @RegisterExtension
-    static final QuarkusUnitTest test = new QuarkusUnitTest()
-            .setArchiveProducer(() -> ShrinkWrap.create(JavaArchive.class)
-                    .addAsResource(new StringAsset(
-                            "quarkus.langchain4j.pgvector.dimension=384\n" +
-                                    "quarkus.langchain4j.pgvector.drop-table-first=true\n" +
-                                    "quarkus.class-loading.parent-first-artifacts=ai.djl.huggingface:tokenizers\n" +
-                                    "quarkus.log.category.\"io.quarkiverse.langchain4j.pgvector\".level=DEBUG\n\n" +
-                                    "quarkus.langchain4j.pgvector.metadata.storage-mode=COMBINED_JSONB\n" +
-                                    "quarkus.langchain4j.pgvector.metadata.column-definitions=metadata JSONB NULL\n" +
-                                    "quarkus.langchain4j.pgvector.metadata.indexes=metadata"),
-                            "application.properties"));
+    static final QuarkusUnitTest test = new QuarkusUnitTest().setArchiveProducer(() -> ShrinkWrap
+            .create(JavaArchive.class)
+            .addAsResource(new StringAsset("quarkus.langchain4j.pgvector.dimension=384\n"
+                    + "quarkus.langchain4j.pgvector.drop-table-first=true\n"
+                    + "quarkus.class-loading.parent-first-artifacts=ai.djl.huggingface:tokenizers\n"
+                    + "quarkus.log.category.\"io.quarkiverse.langchain4j.pgvector\".level=DEBUG\n\n"
+                    + "quarkus.langchain4j.pgvector.metadata.storage-mode=COMBINED_JSONB\n"
+                    + "quarkus.langchain4j.pgvector.metadata.column-definitions=metadata JSONB NULL\n"
+                    + "quarkus.langchain4j.pgvector.metadata.indexes=metadata"), "application.properties"));
 
 }

embedding-stores/pgvector/deployment/src/test/java/io/quarkiverse/langchain4j/pgvector/test/JSONTest.java

@@ -11,13 +11,12 @@ public class JSONTest extends LangChain4jPgVectorBaseTest {
 
     @RegisterExtension
     static final QuarkusUnitTest test = new QuarkusUnitTest()
-            .setArchiveProducer(() -> ShrinkWrap.create(JavaArchive.class)
-                    .addAsResource(new StringAsset(
-                            "quarkus.langchain4j.pgvector.dimension=384\n" +
-                                    "quarkus.langchain4j.pgvector.drop-table-first=true\n" +
-                                    "quarkus.class-loading.parent-first-artifacts=ai.djl.huggingface:tokenizers\n" +
-                                    "quarkus.log.category.\"io.quarkiverse.langchain4j.pgvector\".level=DEBUG\n\n"),
-                            "application.properties"));
+            .setArchiveProducer(() -> ShrinkWrap.create(JavaArchive.class).addAsResource(
+                    new StringAsset("quarkus.langchain4j.pgvector.dimension=384\n"
+                            + "quarkus.langchain4j.pgvector.drop-table-first=true\n"
+                            + "quarkus.class-loading.parent-first-artifacts=ai.djl.huggingface:tokenizers\n"
+                            + "quarkus.log.category.\"io.quarkiverse.langchain4j.pgvector\".level=DEBUG\n\n"),
+                    "application.properties"));
 
     // Default behavior
 }

embedding-stores/pgvector/deployment/src/test/java/io/quarkiverse/langchain4j/pgvector/test/LangChain4jPgVectorBaseTest.java

@@ -47,9 +47,8 @@ protected EmbeddingStore<TextSegment> embeddingStore() {
     }
 
     /**
-     * Just for information, not real benchmark.
-     * JSONTest: Ingesting time 50849 ms. Query average 10 ms
-     * JSONBTest: Ingesting time 56035 ms. Query average 6 ms.
+     * Just for information, not real benchmark. JSONTest: Ingesting time 50849 ms.
+     * Query average 10 ms JSONBTest: Ingesting time 56035 ms. Query average 6 ms.
      * JSONBMultiIndexTest: Ingesting time 47344 ms. Query average 6 ms.
      * ColumnsTest: Ingesting time 49752 ms. Query average 3 ms.
      */

embedding-stores/pgvector/deployment/src/test/java/io/quarkiverse/langchain4j/pgvector/test/PgVectorDataSourceTest.java

@@ -27,15 +27,14 @@ public class PgVectorDataSourceTest {
 
     @RegisterExtension
     static final QuarkusUnitTest test = new QuarkusUnitTest()
-            .setArchiveProducer(() -> ShrinkWrap.create(JavaArchive.class)
-                    .addAsResource(new StringAsset(
-                            // DevServicesConfigBuilderCustomizer overrides the image-name only
-                            // for the default DS, so in this case we have to override it manually
-                            "quarkus.datasource.embeddings-ds.devservices.image-name=pgvector/pgvector:pg16\n" +
-                                    "quarkus.langchain4j.pgvector.datasource=embeddings-ds\n" +
-                                    "quarkus.langchain4j.pgvector.dimension=1536\n" +
-                                    "quarkus.datasource.embeddings-ds.db-kind=postgresql\n"),
-                            "application.properties"));
+            .setArchiveProducer(() -> ShrinkWrap.create(JavaArchive.class).addAsResource(new StringAsset(
+                    // DevServicesConfigBuilderCustomizer overrides the image-name only
+                    // for the default DS, so in this case we have to override it manually
+                    "quarkus.datasource.embeddings-ds.devservices.image-name=pgvector/pgvector:pg16\n"
+                            + "quarkus.langchain4j.pgvector.datasource=embeddings-ds\n"
+                            + "quarkus.langchain4j.pgvector.dimension=1536\n"
+                            + "quarkus.datasource.embeddings-ds.db-kind=postgresql\n"),
+                    "application.properties"));
 
     @io.quarkus.agroal.DataSource("embeddings-ds")
     DataSource ds;

model-providers/openai/azure-openai/runtime/src/main/java/io/quarkiverse/langchain4j/azure/openai/runtime/AzureOpenAiRecorder.java

@@ -24,6 +24,7 @@
 import dev.langchain4j.model.embedding.EmbeddingModel;
 import dev.langchain4j.model.image.DisabledImageModel;
 import dev.langchain4j.model.image.ImageModel;
+import io.quarkiverse.langchain4j.auth.ModelAuthProvider;
 import io.quarkiverse.langchain4j.azure.openai.AzureOpenAiChatModel;
 import io.quarkiverse.langchain4j.azure.openai.AzureOpenAiEmbeddingModel;
 import io.quarkiverse.langchain4j.azure.openai.AzureOpenAiImageModel;
@@ -35,6 +36,8 @@
 import io.quarkiverse.langchain4j.openai.common.QuarkusOpenAiClient;
 import io.quarkiverse.langchain4j.openai.common.runtime.AdditionalPropertiesHack;
 import io.quarkiverse.langchain4j.runtime.NamedConfigUtil;
+import io.quarkus.arc.Arc;
+import io.quarkus.arc.ArcContainer;
 import io.quarkus.arc.SyntheticCreationalContext;
 import io.quarkus.runtime.ShutdownContext;
 import io.quarkus.runtime.annotations.Recorder;
@@ -295,6 +298,12 @@ private LangChain4jAzureOpenAiConfig.AzureAiConfig correspondingAzureOpenAiConfi
 
     private void throwIfApiKeysNotConfigured(String apiKey, String adToken, String configName) {
         if ((apiKey != null) == (adToken != null)) {
+            ArcContainer container = Arc.container();
+            if (container != null && container.instance(ModelAuthProvider.class).isAvailable()) {
+                // Perhaps ModelAuthProvider can provide a method with a default implementation returning a value like `ALL`
+                // to indicate that it applies to all or a specific model only
+                return;
+            }
             throw new ConfigValidationException(createKeyMisconfigurationProblem(configName));
         }
     }

samples/pom.xml

@@ -19,7 +19,7 @@
         <module>fraud-detection</module>
         <module>review-triage</module>
         <module>secure-fraud-detection</module>
-        <module>secure-vertex-ai-gemini-poem</module>
+        <module>secure-poem</module>
         <module>sql-chatbot</module>
     </modules>