Merge pull request #841 from jmartisk/tavily

Tavily web search engine integration
quarkiverse · Aug 30, 2024 · 1bb873c · 1bb873c
2 parents b9c9cfc + 4eb8947
commit 1bb873c
Show file tree

Hide file tree

Showing 34 changed files with 1,688 additions and 0 deletions.
diff --git a/core/deployment/src/main/java/io/quarkiverse/langchain4j/deployment/ToolProcessor.java b/core/deployment/src/main/java/io/quarkiverse/langchain4j/deployment/ToolProcessor.java
@@ -92,6 +92,7 @@ public void telemetry(Capabilities capabilities, BuildProducer<AdditionalBeanBui
     @Record(ExecutionTime.STATIC_INIT)
     public void handleTools(CombinedIndexBuildItem indexBuildItem,
             ToolsRecorder recorder,
+            BuildProducer<AdditionalBeanBuildItem> additionalBeanProducer,
             RecorderContext recorderContext,
             BuildProducer<BytecodeTransformerBuildItem> transformerProducer,
             BuildProducer<GeneratedClassBuildItem> generatedClassProducer,
@@ -176,6 +177,14 @@ public void handleTools(CombinedIndexBuildItem indexBuildItem,
                     boolean ignoreToolMethod = ignoreToolMethod(toolMethod, index);
                     if (ignoreToolMethod) {
                         continue;
+                    } else {
+                        // The WebSearchTool class isn't a CDI bean, so if
+                        // we consider it as a tool, we have to also turn it into one
+                        if (LangChain4jDotNames.WEB_SEARCH_TOOL.equals(className)) {
+                            additionalBeanProducer.produce(AdditionalBeanBuildItem.builder()
+                                    .addBeanClass(className.toString())
+                                    .setUnremovable().build());
+                        }
                     }
 
                     AnnotationValue nameValue = instance.value("name");

diff --git a/docs/modules/ROOT/nav.adoc b/docs/modules/ROOT/nav.adoc
@@ -31,6 +31,7 @@
 ** xref:easy-rag.adoc[Easy RAG]
 ** xref:dev-ui.adoc[Dev UI]
 ** xref:reranking.adoc[Reranking]
+** xref:web-search.adoc[Web search]
 
 * Advanced topics
 ** xref:fault-tolerance.adoc[Fault Tolerance]

diff --git a/docs/modules/ROOT/pages/includes/quarkus-langchain4j-tavily.adoc b/docs/modules/ROOT/pages/includes/quarkus-langchain4j-tavily.adoc
@@ -0,0 +1,220 @@
+
+:summaryTableId: quarkus-langchain4j-tavily
+[.configuration-legend]
+icon:lock[title=Fixed at build time] Configuration property fixed at build time - All other configuration properties are overridable at runtime
+[.configuration-reference.searchable, cols="80,.^10,.^10"]
+|===
+
+h|[[quarkus-langchain4j-tavily_configuration]]link:#quarkus-langchain4j-tavily_configuration[Configuration property]
+
+h|Type
+h|Default
+
+a| [[quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-base-url]]`link:#quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-base-url[quarkus.langchain4j.tavily.base-url]`
+
+
+[.description]
+--
+Base URL of the Tavily API
+
+ifdef::add-copy-button-to-env-var[]
+Environment variable: env_var_with_copy_button:+++QUARKUS_LANGCHAIN4J_TAVILY_BASE_URL+++[]
+endif::add-copy-button-to-env-var[]
+ifndef::add-copy-button-to-env-var[]
+Environment variable: `+++QUARKUS_LANGCHAIN4J_TAVILY_BASE_URL+++`
+endif::add-copy-button-to-env-var[]
+--|string 
+|`https://api.tavily.com`
+
+
+a| [[quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-api-key]]`link:#quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-api-key[quarkus.langchain4j.tavily.api-key]`
+
+
+[.description]
+--
+API key for the Tavily API
+
+ifdef::add-copy-button-to-env-var[]
+Environment variable: env_var_with_copy_button:+++QUARKUS_LANGCHAIN4J_TAVILY_API_KEY+++[]
+endif::add-copy-button-to-env-var[]
+ifndef::add-copy-button-to-env-var[]
+Environment variable: `+++QUARKUS_LANGCHAIN4J_TAVILY_API_KEY+++`
+endif::add-copy-button-to-env-var[]
+--|string 
+|required icon:exclamation-circle[title=Configuration property is required]
+
+
+a| [[quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-max-results]]`link:#quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-max-results[quarkus.langchain4j.tavily.max-results]`
+
+
+[.description]
+--
+Maximum number of results to return
+
+ifdef::add-copy-button-to-env-var[]
+Environment variable: env_var_with_copy_button:+++QUARKUS_LANGCHAIN4J_TAVILY_MAX_RESULTS+++[]
+endif::add-copy-button-to-env-var[]
+ifndef::add-copy-button-to-env-var[]
+Environment variable: `+++QUARKUS_LANGCHAIN4J_TAVILY_MAX_RESULTS+++`
+endif::add-copy-button-to-env-var[]
+--|int 
+|`5`
+
+
+a| [[quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-timeout]]`link:#quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-timeout[quarkus.langchain4j.tavily.timeout]`
+
+
+[.description]
+--
+The timeout duration for Tavily requests.
+
+ifdef::add-copy-button-to-env-var[]
+Environment variable: env_var_with_copy_button:+++QUARKUS_LANGCHAIN4J_TAVILY_TIMEOUT+++[]
+endif::add-copy-button-to-env-var[]
+ifndef::add-copy-button-to-env-var[]
+Environment variable: `+++QUARKUS_LANGCHAIN4J_TAVILY_TIMEOUT+++`
+endif::add-copy-button-to-env-var[]
+--|link:https://docs.oracle.com/javase/8/docs/api/java/time/Duration.html[Duration]
+  link:#duration-note-anchor-{summaryTableId}[icon:question-circle[title=More information about the Duration format]]
+|`60S`
+
+
+a| [[quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-log-requests]]`link:#quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-log-requests[quarkus.langchain4j.tavily.log-requests]`
+
+
+[.description]
+--
+Whether requests to Tavily should be logged
+
+ifdef::add-copy-button-to-env-var[]
+Environment variable: env_var_with_copy_button:+++QUARKUS_LANGCHAIN4J_TAVILY_LOG_REQUESTS+++[]
+endif::add-copy-button-to-env-var[]
+ifndef::add-copy-button-to-env-var[]
+Environment variable: `+++QUARKUS_LANGCHAIN4J_TAVILY_LOG_REQUESTS+++`
+endif::add-copy-button-to-env-var[]
+--|boolean 
+|`false`
+
+
+a| [[quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-log-responses]]`link:#quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-log-responses[quarkus.langchain4j.tavily.log-responses]`
+
+
+[.description]
+--
+Whether responses from Tavily should be logged
+
+ifdef::add-copy-button-to-env-var[]
+Environment variable: env_var_with_copy_button:+++QUARKUS_LANGCHAIN4J_TAVILY_LOG_RESPONSES+++[]
+endif::add-copy-button-to-env-var[]
+ifndef::add-copy-button-to-env-var[]
+Environment variable: `+++QUARKUS_LANGCHAIN4J_TAVILY_LOG_RESPONSES+++`
+endif::add-copy-button-to-env-var[]
+--|boolean 
+|`false`
+
+
+a| [[quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-search-depth]]`link:#quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-search-depth[quarkus.langchain4j.tavily.search-depth]`
+
+
+[.description]
+--
+The search depth to use. This can be "basic" or "advanced". Basic is the default.
+
+ifdef::add-copy-button-to-env-var[]
+Environment variable: env_var_with_copy_button:+++QUARKUS_LANGCHAIN4J_TAVILY_SEARCH_DEPTH+++[]
+endif::add-copy-button-to-env-var[]
+ifndef::add-copy-button-to-env-var[]
+Environment variable: `+++QUARKUS_LANGCHAIN4J_TAVILY_SEARCH_DEPTH+++`
+endif::add-copy-button-to-env-var[]
+-- a|
+`basic`, `advanced` 
+|`basic`
+
+
+a| [[quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-include-answer]]`link:#quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-include-answer[quarkus.langchain4j.tavily.include-answer]`
+
+
+[.description]
+--
+Include a short answer to original query. Default is false.
+
+ifdef::add-copy-button-to-env-var[]
+Environment variable: env_var_with_copy_button:+++QUARKUS_LANGCHAIN4J_TAVILY_INCLUDE_ANSWER+++[]
+endif::add-copy-button-to-env-var[]
+ifndef::add-copy-button-to-env-var[]
+Environment variable: `+++QUARKUS_LANGCHAIN4J_TAVILY_INCLUDE_ANSWER+++`
+endif::add-copy-button-to-env-var[]
+--|boolean 
+|`false`
+
+
+a| [[quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-include-raw-content]]`link:#quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-include-raw-content[quarkus.langchain4j.tavily.include-raw-content]`
+
+
+[.description]
+--
+Include the cleaned and parsed HTML content of each search result. Default is false.
+
+ifdef::add-copy-button-to-env-var[]
+Environment variable: env_var_with_copy_button:+++QUARKUS_LANGCHAIN4J_TAVILY_INCLUDE_RAW_CONTENT+++[]
+endif::add-copy-button-to-env-var[]
+ifndef::add-copy-button-to-env-var[]
+Environment variable: `+++QUARKUS_LANGCHAIN4J_TAVILY_INCLUDE_RAW_CONTENT+++`
+endif::add-copy-button-to-env-var[]
+--|boolean 
+|`false`
+
+
+a| [[quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-include-domains]]`link:#quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-include-domains[quarkus.langchain4j.tavily.include-domains]`
+
+
+[.description]
+--
+A list of domains to specifically include in the search results. Default is ++[]++, which includes all domains.
+
+ifdef::add-copy-button-to-env-var[]
+Environment variable: env_var_with_copy_button:+++QUARKUS_LANGCHAIN4J_TAVILY_INCLUDE_DOMAINS+++[]
+endif::add-copy-button-to-env-var[]
+ifndef::add-copy-button-to-env-var[]
+Environment variable: `+++QUARKUS_LANGCHAIN4J_TAVILY_INCLUDE_DOMAINS+++`
+endif::add-copy-button-to-env-var[]
+--|list of string 
+|`[]`
+
+
+a| [[quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-exclude-domains]]`link:#quarkus-langchain4j-tavily_quarkus-langchain4j-tavily-exclude-domains[quarkus.langchain4j.tavily.exclude-domains]`
+
+
+[.description]
+--
+A list of domains to specifically exclude from the search results. Default is ++[]++, which doesn't exclude any domains.
+
+ifdef::add-copy-button-to-env-var[]
+Environment variable: env_var_with_copy_button:+++QUARKUS_LANGCHAIN4J_TAVILY_EXCLUDE_DOMAINS+++[]
+endif::add-copy-button-to-env-var[]
+ifndef::add-copy-button-to-env-var[]
+Environment variable: `+++QUARKUS_LANGCHAIN4J_TAVILY_EXCLUDE_DOMAINS+++`
+endif::add-copy-button-to-env-var[]
+--|list of string 
+|`[]`
+
+|===
+ifndef::no-duration-note[]
+[NOTE]
+[id='duration-note-anchor-{summaryTableId}']
+.About the Duration format
+====
+To write duration values, use the standard `java.time.Duration` format.
+See the link:https://docs.oracle.com/en/java/javase/17/docs/api/java.base/java/time/Duration.html#parse(java.lang.CharSequence)[Duration#parse() Java API documentation] for more information.
+
+You can also use a simplified format, starting with a number:
+
+* If the value is only a number, it represents time in seconds.
+* If the value is a number followed by `ms`, it represents time in milliseconds.
+
+In other cases, the simplified format is translated to the `java.time.Duration` format for parsing:
+
+* If the value is a number followed by `h`, `m`, or `s`, it is prefixed with `PT`.
+* If the value is a number followed by `d`, it is prefixed with `P`.
+====
+endif::no-duration-note[]
diff --git a/docs/modules/ROOT/pages/web-search.adoc b/docs/modules/ROOT/pages/web-search.adoc
@@ -0,0 +1,67 @@
+= Using web search
+
+Quarkus LangChain4j currently supports the https://tavily.com/[Tavily] search engine.
+To use it, add the `quarkus-langchain4j-tavily` extension to your project. You'll need to specify the API key, this is done by the `quarkus.langchain4j.tavily.api-key` property.
+
+After this, you can inject the search engine into your application using
+
+[source,java]
+----
+@Inject
+WebSearchEngine engine;
+----
+
+and then use it by calling its `search` method.
+
+If you want to let an chat model use web search by itself, there are
+generally two recommended ways to accomplish this: either by implementing a
+tool that uses it, or as a content retriever inside a RAG pipeline. The
+https://github.com/quarkiverse/quarkus-langchain4j/tree/main/samples/chatbot-web-search[chatbot-web-search]
+example in the `quarkus-langchain4j` repository demonstrates using web
+search as a tool.
+
+== Using Web search as a tool
+
+To use web search as a tool that the LLM can decide to execute (and the
+relevant search results will be the return value of the tool execution), you
+can either use the provided tool from the upstream LangChain4j project,
+in class `dev.langchain4j.web.search.WebSearchTool`, or implement your own tool
+if that one does not fit your requirements. The `samples/chatbot-web-search`
+example demonstrates how to use the provided tool.
+
+== Using Web search in a RAG pipeline
+
+There is also a provided content retriever, `dev.langchain4j.rag.content.retriever.WebSearchContentRetriever` that uses
+a web search engine to retrieve relevant documents.
+For inspiration, the retrieval augmentor that wraps it may look like this:
+
+[source,java]
+----
+@ApplicationScoped
+public class WebSearchRetrievalAugmentor implements Supplier<RetrievalAugmentor> {
+
+    @Inject
+    WebSearchEngine webSearchEngine;
+
+    @Inject
+    ChatLanguageModel chatModel;
+
+    @Override
+    public RetrievalAugmentor get() {
+        return DefaultRetrievalAugmentor.builder()
+                .queryTransformer((question) -> {
+                    // before actually querying the engine, we need to transform the
+                    // user's question into a suitable search query
+                    String query = chatModel.generate("Transform the user's question into a suitable query for the " +
+                            "Tavily search engine. The query should yield the results relevant to answering the user's question." +
+                            "User's question: " + question.text());
+                    return Collections.singleton(Query.from(query));
+                }).contentRetriever(new WebSearchContentRetriever(webSearchEngine, 10))
+                .build();
+    }
+}
+----
+
+== Tavily configuration reference
+
+include::includes/quarkus-langchain4j-tavily.adoc[leveloffset=+1,opts=optional]
diff --git a/docs/pom.xml b/docs/pom.xml
@@ -76,6 +76,11 @@
             <artifactId>quarkus-langchain4j-watsonx</artifactId>
             <version>${project.version}</version>
         </dependency>
+        <dependency>
+            <groupId>io.quarkiverse.langchain4j</groupId>
+            <artifactId>quarkus-langchain4j-tavily</artifactId>
+            <version>${project.version}</version>
+        </dependency>
         <dependency>
             <groupId>io.quarkiverse.langchain4j</groupId>
             <artifactId>quarkus-langchain4j-easy-rag</artifactId>
@@ -279,6 +284,19 @@
                 </exclusion>
             </exclusions>
         </dependency>
+        <dependency>
+            <groupId>io.quarkiverse.langchain4j</groupId>
+            <artifactId>quarkus-langchain4j-tavily-deployment</artifactId>
+            <version>${project.version}</version>
+            <type>pom</type>
+            <scope>test</scope>
+            <exclusions>
+                <exclusion>
+                    <groupId>*</groupId>
+                    <artifactId>*</artifactId>
+                </exclusion>
+            </exclusions>
+        </dependency>
     </dependencies>
 
     <build>
@@ -343,6 +361,7 @@
                                     <include>quarkus-langchain4j-watsonx.adoc</include>
                                     <include>quarkus-langchain4j-mistralai.adoc</include>
                                     <include>quarkus-langchain4j-neo4j.adoc</include>
+                                    <include>quarkus-langchain4j-tavily.adoc</include>
                                     <filtering>false</filtering>
                                 </resource>
                                 <resource>

diff --git a/docs/src/main/resources/application.properties b/docs/src/main/resources/application.properties
@@ -9,6 +9,7 @@ quarkus.langchain4j.pinecone.environment=abc
 quarkus.langchain4j.pinecone.index-name=abc
 quarkus.langchain4j.pinecone.project-id=abc
 quarkus.langchain4j.pinecone.api-key=abc
+quarkus.langchain4j.tavily.api-key=abc
 quarkus.langchain4j.redis.dimension=180
 quarkus.langchain4j.easy-rag.path=abc
 quarkus.langchain4j.easy-rag.ingestion-strategy=off

diff --git a/pom.xml b/pom.xml
@@ -46,6 +46,8 @@
 
     <module>quarkus-integrations/websockets-next</module>
 
+    <module>web-search-engines/tavily</module>
+
     <module>rag/easy-rag</module>
     <module>rag/parsers-base</module>
 
@@ -205,6 +207,7 @@
         <module>samples/chatbot</module>
         <module>samples/chatbot-easy-rag</module>
         <module>samples/sql-chatbot</module>
+        <module>samples/chatbot-web-search</module>
       </modules>
     </profile>
   </profiles>