-
Notifications
You must be signed in to change notification settings - Fork 35
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Merge pull request #28 from awslabs/chain-agent
Chain agent
- Loading branch information
Showing
21 changed files
with
1,904 additions
and
35 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
276 changes: 276 additions & 0 deletions
276
docs/src/content/docs/agents/built-in/bedrock-translator-agent.mdx
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,276 @@ | ||
| --- | ||
| title: Bedrock Translator Agent | ||
| description: Documentation for the Bedrock Translator Agent in the Multi-Agent Orchestrator System | ||
| --- | ||
|
|
||
| The `BedrockTranslatorAgent` uses Amazon Bedrock's language models to translate text between different languages. | ||
|
|
||
| ## Key Features | ||
|
|
||
| - Utilizes Amazon Bedrock's language models | ||
| - Supports translation between multiple languages | ||
| - Allows dynamic setting of source and target languages | ||
| - Can be used standalone or as part of a [ChainAgent](/multi-agent-orchestrator/agents/built-in/chain-agent) | ||
| - Configurable inference parameters for fine-tuned control | ||
|
|
||
| ## Creating a Bedrock Translator Agent | ||
|
|
||
| ### Basic Example | ||
|
|
||
| To create a new `BedrockTranslatorAgent` with minimal configuration: | ||
|
|
||
| import { Tabs, TabItem } from '@astrojs/starlight/components'; | ||
|
|
||
| <Tabs syncKey="runtime"> | ||
| <TabItem label="TypeScript" icon="seti:typescript" color="blue"> | ||
| ```typescript | ||
| import { BedrockTranslatorAgent, BedrockTranslatorAgentOptions } from 'multi-agent-orchestrator'; | ||
|
|
||
| const agent = new BedrockTranslatorAgent({ | ||
| name: 'BasicTranslator', | ||
| description: 'Translates text to English', | ||
| targetLanguage: 'English' | ||
| }); | ||
| ``` | ||
| </TabItem> | ||
| <TabItem label="Python" icon="seti:python"> | ||
| ```python | ||
| from multi_agent_orchestrator.agents import BedrockTranslatorAgent, BedrockTranslatorAgentOptions | ||
| agent = BedrockTranslatorAgent(BedrockTranslatorAgentOptions( | ||
| name='BasicTranslator', | ||
| description='Translates text to English', | ||
| target_language='English' | ||
| )) | ||
| ``` | ||
| </TabItem> | ||
| </Tabs> | ||
|
|
||
| ### Advanced Example | ||
|
|
||
| For more complex use cases, you can create a BedrockTranslatorAgent with custom settings: | ||
|
|
||
| <Tabs syncKey="runtime"> | ||
| <TabItem label="TypeScript" icon="seti:typescript" color="blue"> | ||
| ```typescript | ||
| import { BedrockTranslatorAgent, BedrockTranslatorAgentOptions, BEDROCK_MODEL_ID_CLAUDE_3_SONNET } from 'multi-agent-orchestrator'; | ||
| const options: BedrockTranslatorAgentOptions = { | ||
| name: 'AdvancedTranslator', | ||
| description: 'Advanced translator with custom settings', | ||
| sourceLanguage: 'French', | ||
| targetLanguage: 'German', | ||
| modelId: BEDROCK_MODEL_ID_CLAUDE_3_SONNET, | ||
| region: 'us-west-2', | ||
| inferenceConfig: { | ||
| maxTokens: 2000, | ||
| temperature: 0.1, | ||
| topP: 0.95, | ||
| stopSequences: ['###'] | ||
| } | ||
| }; | ||
| const agent = new BedrockTranslatorAgent(options); | ||
| ``` | ||
| </TabItem> | ||
| <TabItem label="Python" icon="seti:python"> | ||
| ```python | ||
| from multi_agent_orchestrator.agents import BedrockTranslatorAgent, BedrockTranslatorAgentOptions | ||
| from multi_agent_orchestrator.types import BEDROCK_MODEL_ID_CLAUDE_3_SONNET | ||
| options = BedrockTranslatorAgentOptions( | ||
| name='AdvancedTranslator', | ||
| description='Advanced translator with custom settings', | ||
| source_language='French', | ||
| target_language='German', | ||
| model_id=BEDROCK_MODEL_ID_CLAUDE_3_SONNET, | ||
| region='us-west-2', | ||
| inference_config={ | ||
| 'maxTokens': 2000, | ||
| 'temperature': 0.1, | ||
| 'topP': 0.95, | ||
| 'stopSequences': ['###'] | ||
| } | ||
| ) | ||
| agent = BedrockTranslatorAgent(options) | ||
| ``` | ||
| </TabItem> | ||
| </Tabs> | ||
|
|
||
| ## Dynamic Language Setting | ||
|
|
||
| To set the language during the invocation: | ||
|
|
||
| <Tabs syncKey="runtime"> | ||
| <TabItem label="TypeScript" icon="seti:typescript" color="blue"> | ||
| ```typescript | ||
| import { MultiAgentOrchestrator, BedrockTranslatorAgent } from 'multi-agent-orchestrator'; | ||
| const translator = new BedrockTranslatorAgent({ | ||
| name: 'DynamicTranslator', | ||
| description: 'Translator with dynamically set languages' | ||
| }); | ||
| const orchestrator = new MultiAgentOrchestrator(); | ||
| orchestrator.addAgent(translator); | ||
| async function translateWithDynamicLanguages(text: string, fromLang: string, toLang: string) { | ||
| translator.setSourceLanguage(fromLang); | ||
| translator.setTargetLanguage(toLang); | ||
| const response = await orchestrator.routeRequest( | ||
| text, | ||
| 'user123', | ||
| 'session456' | ||
| ); | ||
| console.log(`Translated from ${fromLang} to ${toLang}:`, response); | ||
| } | ||
| // Usage | ||
| translateWithDynamicLanguages("Hello, world!", "English", "French"); | ||
| translateWithDynamicLanguages("Bonjour le monde!", "French", "Spanish"); | ||
| ``` | ||
| </TabItem> | ||
| <TabItem label="Python" icon="seti:python"> | ||
| ```python | ||
| from multi_agent_orchestrator.orchestrator import MultiAgentOrchestrator | ||
| from multi_agent_orchestrator.agents import BedrockTranslatorAgent, BedrockTranslatorAgentOptions | ||
| translator = BedrockTranslatorAgent(BedrockTranslatorAgentOptions( | ||
| name='DynamicTranslator', | ||
| description='Translator with dynamically set languages' | ||
| )) | ||
| orchestrator = MultiAgentOrchestrator() | ||
| orchestrator.add_agent(translator) | ||
| async def translate_with_dynamic_languages(text: str, from_lang: str, to_lang: str): | ||
| translator.set_source_language(from_lang) | ||
| translator.set_target_language(to_lang) | ||
| response = await orchestrator.route_request( | ||
| text, | ||
| 'user123', | ||
| 'session456' | ||
| ) | ||
| print(f"Translated from {from_lang} to {to_lang}:", response) | ||
| # Usage | ||
| import asyncio | ||
| asyncio.run(translate_with_dynamic_languages("Hello, world!", "English", "French")) | ||
| asyncio.run(translate_with_dynamic_languages("Bonjour le monde!", "French", "Spanish")) | ||
| ``` | ||
| </TabItem> | ||
| </Tabs> | ||
|
|
||
| ## Usage with ChainAgent | ||
|
|
||
| The `BedrockTranslatorAgent` can be effectively used within a `ChainAgent` for complex multilingual processing workflows. Here's an example that demonstrates translating user input and processing it: | ||
|
|
||
| <Tabs syncKey="runtime"> | ||
| <TabItem label="TypeScript" icon="seti:typescript" color="blue"> | ||
| ```typescript | ||
| import { MultiAgentOrchestrator, ChainAgent, BedrockTranslatorAgent, BedrockLLMAgent } from 'multi-agent-orchestrator'; | ||
| // Create translator agents | ||
| const translatorToEnglish = new BedrockTranslatorAgent({ | ||
| name: 'TranslatorToEnglish', | ||
| description: 'Translates input to English', | ||
| targetLanguage: 'English' | ||
| }); | ||
| // Create a processing agent (e.g., a BedrockLLMAgent) | ||
| const processor = new BedrockLLMAgent({ | ||
| name: 'EnglishProcessor', | ||
| description: 'Processes text in English' | ||
| }); | ||
| // Create a ChainAgent | ||
| const chainAgent = new ChainAgent({ | ||
| name: 'TranslateProcessTranslate', | ||
| description: 'Translates, processes, and translates back', | ||
| agents: [translatorToEnglish, processor] | ||
| }); | ||
| const orchestrator = new MultiAgentOrchestrator(); | ||
| orchestrator.addAgent(chainAgent); | ||
| // Function to handle user input | ||
| async function handleMultilingualInput(input: string, sourceLanguage: string) { | ||
| translatorToEnglish.setSourceLanguage(sourceLanguage); | ||
| const response = await orchestrator.routeRequest( | ||
| input, | ||
| 'user123', | ||
| 'session456' | ||
| ); | ||
| console.log('Response:', response); | ||
| } | ||
| // Usage | ||
| handleMultilingualInput("Hola, ¿cómo estás?", "Spanish"); | ||
| ``` | ||
| </TabItem> | ||
| <TabItem label="Python" icon="seti:python"> | ||
| ```python | ||
| from multi_agent_orchestrator.orchestrator import MultiAgentOrchestrator | ||
| from multi_agent_orchestrator.agents import ChainAgent, BedrockTranslatorAgent, BedrockLLMAgent | ||
| from multi_agent_orchestrator.agents import ChainAgentOptions, BedrockTranslatorAgentOptions, BedrockLLMAgentOptions | ||
| # Create translator agents | ||
| translator_to_english = BedrockTranslatorAgent(BedrockTranslatorAgentOptions( | ||
| name='TranslatorToEnglish', | ||
| description='Translates input to English', | ||
| target_language='English' | ||
| )) | ||
| # Create a processing agent (e.g., a BedrockLLMAgent) | ||
| processor = BedrockLLMAgent(BedrockLLMAgentOptions( | ||
| name='EnglishProcessor', | ||
| description='Processes text in English' | ||
| )) | ||
| # Create a ChainAgent | ||
| chain_agent = ChainAgent(ChainAgentOptions( | ||
| name='TranslateProcessTranslate', | ||
| description='Translates, processes, and translates back', | ||
| agents=[translator_to_english, processor] | ||
| )) | ||
| orchestrator = MultiAgentOrchestrator() | ||
| orchestrator.add_agent(chain_agent) | ||
| # Function to handle user input | ||
| async def handle_multilingual_input(input_text: str, source_language: str): | ||
| translator_to_english.set_source_language(source_language) | ||
| response = await orchestrator.route_request( | ||
| input_text, | ||
| 'user123', | ||
| 'session456' | ||
| ) | ||
| print('Response:', response) | ||
| # Usage | ||
| import asyncio | ||
| asyncio.run(handle_multilingual_input("Hola, ¿cómo estás?", "Spanish")) | ||
| ``` | ||
| </TabItem> | ||
| </Tabs> | ||
|
|
||
| In this example: | ||
| 1. The first translator agent converts the input to English. | ||
| 2. The processor agent (e.g., a `BedrockLLMAgent`) processes the English text. | ||
|
|
||
| This setup allows for seamless multilingual processing, where the core logic can be implemented in English while supporting input and output in various languages. | ||
|
|
||
| --- | ||
|
|
||
| By leveraging the `BedrockTranslatorAgent`, you can create sophisticated multilingual applications and workflows, enabling seamless communication and processing across language barriers in your Multi-Agent Orchestrator system. |
Oops, something went wrong.