Merge pull request #1112 from glaforge:main

PiperOrigin-RevId: 895777413

Author: copybara-github

contrib/langchain4j/README.md

@@ -0,0 +1,203 @@
+# ADK LangChain4j Integration Library
+
+## Overview
+
+The ADK LangChain4j Integration library provides a bridge between the Agent Development Kit (ADK) and 
+[LangChain4j](https://docs.langchain4j.dev/). 
+The main purpose of this module is to allow ADK to have access to all the LLM providers supported 
+by the LangChain4j ecosystem (e.g., Anthropic, OpenAI, Ollama, Google Gemini, and many more). 
+
+This library supports multiple AI providers, function calling (tools), streaming responses, 
+and automatically maps ADK models and mechanisms to their LangChain4j counterparts.
+
+## Getting Started
+
+### Maven Dependencies
+
+To use ADK Java with the LangChain4j integration in your application, 
+add the following dependencies to your `pom.xml`:
+
+#### Basic Setup
+
+```xml
+<dependencies>
+    <!-- ADK Core -->
+    <dependency>
+        <groupId>com.google.adk</groupId>
+        <artifactId>google-adk</artifactId>
+        <version>1.0.0</version> <!-- Use the most recent ADK version -->
+    </dependency>
+
+    <!-- ADK LangChain4j Integration -->
+    <dependency>
+        <groupId>com.google.adk</groupId>
+        <artifactId>google-adk-langchain4j</artifactId>
+        <version>1.0.0</version> <!-- Use the most recent ADK version -->
+    </dependency>
+</dependencies>
+```
+
+#### Provider-Specific Dependencies
+
+You'll also need to add the LangChain4j provider dependencies for the AI services you want to use.
+Refer to the [full list](https://docs.langchain4j.dev/category/language-models)
+of supported models that you want to use in your project.
+
+**Anthropic**  (Claude):
+```xml
+<dependency>
+    <groupId>dev.langchain4j</groupId>
+    <artifactId>langchain4j-anthropic</artifactId>
+</dependency>
+```
+
+**OpenAI** and compatible models:
+```xml
+<dependency>
+    <groupId>dev.langchain4j</groupId>
+    <artifactId>langchain4j-open-ai</artifactId>
+</dependency>
+```
+
+**Ollama** (local models):
+```xml
+<dependency>
+    <groupId>dev.langchain4j</groupId>
+    <artifactId>langchain4j-ollama</artifactId>
+</dependency>
+```
+
+*(You can use any other `langchain4j-*` module supported by LangChain4j).*
+
+## Quick Start Examples
+
+Once you have the dependencies set up, you can create a simple ADK agent using any LangChain4j chat model. 
+You just need to wrap the `ChatModel` into the ADK `LangChain4j` model builder.
+
+### Anthropic Example
+
+```java
+import com.google.adk.agents.LlmAgent;
+import com.google.adk.models.langchain4j.LangChain4j;
+import dev.langchain4j.model.anthropic.AnthropicChatModel;
+
+// 1. Initialize LangChain4j model
+AnthropicChatModel claudeModel = AnthropicChatModel.builder()
+    .apiKey(System.getenv("ANTHROPIC_API_KEY"))
+    .modelName("claude-sonnet-4-6")
+    .build();
+
+// 2. Wrap the LangChain4j model for ADK
+LangChain4j adkModel = LangChain4j.builder()
+    .chatModel(claudeModel)
+    .modelName("claude-sonnet-4-6")
+    .build();
+
+// 3. Create your agent
+LlmAgent agent = LlmAgent.builder()
+    .name("science-teacher")
+    .description("A helpful science teacher")
+    .model(adkModel)
+    .instruction("You are a helpful science teacher that explains concepts clearly.")
+    .build();
+```
+
+### OpenAI Example
+
+```java
+import com.google.adk.agents.LlmAgent;
+import com.google.adk.models.langchain4j.LangChain4j;
+import dev.langchain4j.model.openai.OpenAiChatModel;
+
+// 1. Initialize LangChain4j model
+OpenAiChatModel gptModel = OpenAiChatModel.builder()
+    .apiKey(System.getenv("OPENAI_API_KEY"))
+    .modelName("gpt-4o-mini")
+    .build();
+
+// 2. Wrap the LangChain4j model for ADK
+LangChain4j adkModel = LangChain4j.builder()
+    .chatModel(gptModel)
+    .modelName("gpt-4o-mini")
+    .build();
+
+// 3. Create your agent
+LlmAgent agent = LlmAgent.builder()
+    .name("friendly-assistant")
+    .description("A friendly assistant")
+    .model(adkModel)
+    .instruction("You are a friendly assistant.")
+    .build();
+```
+
+### Ollama Example (Local Models)
+
+```java
+import com.google.adk.agents.LlmAgent;
+import com.google.adk.models.langchain4j.LangChain4j;
+import dev.langchain4j.model.ollama.OllamaChatModel;
+
+// 1. Initialize LangChain4j model for Ollama
+OllamaChatModel ollamaModel = OllamaChatModel.builder()
+    .baseUrl("http://localhost:11434")
+    .modelName("llama3")
+    .build();
+
+// 2. Wrap the LangChain4j model for ADK
+LangChain4j adkModel = LangChain4j.builder()
+    .chatModel(ollamaModel)
+    .modelName("llama3")
+    .build();
+
+// 3. Create your agent
+LlmAgent agent = LlmAgent.builder()
+    .name("local-agent")
+    .description("A local assistant running on Ollama")
+    .model(adkModel)
+    .instruction("You are an assistant running locally.")
+    .build();
+```
+
+## Advanced Usage
+
+### Streaming Responses
+
+The integration fully supports streaming models from LangChain4j using the `streamingChatModel` property of the `LangChain4j` builder. 
+Make sure to use a `StreamingChatModel` instead of a regular `ChatModel` from LangChain4j.
+
+```java
+import dev.langchain4j.model.anthropic.AnthropicStreamingChatModel;
+
+// 1. Initialize a LangChain4j STREAMING chat model
+AnthropicStreamingChatModel streamingModel = AnthropicStreamingChatModel.builder()
+    .apiKey(System.getenv("ANTHROPIC_API_KEY"))
+    .modelName("claude-sonnet-4-6")
+    .build();
+
+// 2. Wrap it for ADK using `streamingChatModel` config
+LangChain4j adkStreamingModel = LangChain4j.builder()
+    .streamingChatModel(streamingModel)
+    .modelName("claude-sonnet-4-6")
+    .build();
+
+// 3. Create your agent as usual
+LlmAgent agent = LlmAgent.builder()
+    .name("streaming-agent")
+    .model(adkStreamingModel)
+    .instruction("You answer questions as fast as possible.")
+    .build();
+```
+
+### Function Calling (Tools)
+
+ADK tools are automatically mapped to LangChain4j tools under the hood. 
+You can configure them as you usually do in ADK:
+
+```java
+LlmAgent agent = LlmAgent.builder()
+    .name("weather-agent")
+    .model(adkModel)
+    .instruction("If asked about weather, you MUST call the `getWeather` function.")
+    .tools(FunctionTool.create(ToolExample.class, "getWeather"))
+    .build();
+```