LangChain Utilities
Components provided by the LangChain Utilities integration.
CSV Agent
Construct a CSV agent from a CSV and tools.
| Input | Type | Required |
|---|---|---|
| File Path | file | Yes |
| Allow Dangerous Code | bool | Yes |
| watsonx API Endpoint | str | No |
| Text | str | Yes |
| Language Model | model | Yes |
| watsonx Project ID | str | No |
Advanced options: Agent Description [Deprecated], Agent Type, API Key, Handle Parse Errors, Max Iterations, Pandas Kwargs, Verbose
| Output | Type |
|---|---|
| Response | Message |
| Agent | AgentExecutor |
Character Text Splitter
Split text by number of characters.
| Input | Type | Required |
|---|---|---|
| Input | other | Yes |
| Chunk Overlap | int | No |
| Chunk Size | int | No |
| Separator | str | No |
| Output | Type |
|---|---|
| JSON | JSON |
ConversationChain
Chain to have a conversation and load context from memory.
| Input | Type | Required |
|---|---|---|
| Language Model | other | Yes |
| Memory | other | No |
| Input | str | Yes |
| Output | Type |
|---|---|
| Text | Message |
Fake Embeddings
Generate fake embeddings, useful for initial testing and connecting components.
| Input | Type | Required |
|---|---|---|
| Dimensions | int | No |
| Output | Type |
|---|---|
| Embedding Model | Embeddings |
HTML Link Extractor
Extract hyperlinks from HTML content.
| Input | Type | Required |
|---|---|---|
| Input | other | Yes |
| Drop URL fragments | bool | No |
| Kind of edge | str | No |
| Output | Type |
|---|---|
| JSON | JSON |
JsonAgent
Construct a json agent from an LLM and tools.
| Input | Type | Required |
|---|---|---|
| Language Model | other | Yes |
| File Path | file | Yes |
| Input | str | No |
Advanced options: Agent Description [Deprecated], Handle Parse Errors, Max Iterations, Verbose
| Output | Type |
|---|---|
| Response | Message |
| Agent | AgentExecutor |
LLMCheckerChain
Chain for question-answering with self-verification.
| Input | Type | Required |
|---|---|---|
| Language Model | other | Yes |
| Input | str | Yes |
| Output | Type |
|---|---|
| Text | Message |
LLMMathChain
Chain that interprets a prompt and executes python code to do math.
| Input | Type | Required |
|---|---|---|
| Language Model | other | Yes |
| Input | str | Yes |
| Output | Type |
|---|---|
| Message | Message |
Language Recursive Text Splitter
Split text into chunks of a specified length based on language.
| Input | Type | Required |
|---|---|---|
| Input | other | Yes |
| Chunk Overlap | int | No |
| Chunk Size | int | No |
| Code Language | str | No |
| Output | Type |
|---|---|
| JSON | JSON |
Natural Language Text Splitter
Split text based on natural language boundaries, optimized for a specified language.
| Input | Type | Required |
|---|---|---|
| Input | other | Yes |
| Chunk Overlap | int | No |
| Chunk Size | int | No |
| Language | str | No |
| Separator | str | No |
| Output | Type |
|---|---|
| JSON | JSON |
Natural Language to SQL
Generate SQL from natural language.
| Input | Type | Required |
|---|---|---|
| SQLDatabase | other | Yes |
| Language Model | other | Yes |
| Input | str | Yes |
| Prompt | str | No |
| Top K | int | No |
| Output | Type |
|---|---|
| Message | Message |
OpenAI Tools Agent
Agent that uses tools via openai-tools.
| Input | Type | Required |
|---|---|---|
| Tools | other | No |
| watsonx API Endpoint | str | No |
| Input | str | No |
| Language Model | model | Yes |
| watsonx Project ID | str | No |
| System Prompt | str | No |
| Prompt | str | No |
Advanced options: Chat History, Agent Description [Deprecated], API Key, Handle Parse Errors, Max Iterations, Verbose
| Output | Type |
|---|---|
| Response | Message |
| Agent | AgentExecutor |
OpenAPI Agent
Agent to interact with OpenAPI API.
| Input | Type | Required |
|---|---|---|
| File Path | file | Yes |
| Allow Dangerous Requests | bool | Yes |
| watsonx API Endpoint | str | No |
| Input | str | No |
| Language Model | model | Yes |
| watsonx Project ID | str | No |
Advanced options: Agent Description [Deprecated], API Key, Handle Parse Errors, Max Iterations, Verbose
| Output | Type |
|---|---|
| Response | Message |
| Agent | AgentExecutor |
Prompt Hub
Prompt Component that uses LangChain Hub prompts
| Input | Type | Required |
|---|---|---|
| LangChain API Key | str | Yes |
| LangChain Hub Prompt | str | Yes |
| Output | Type |
|---|---|
| Build Prompt | Message |
Recursive Character Text Splitter
Split text trying to keep all related text together.
| Input | Type | Required |
|---|---|---|
| Input | other | Yes |
| Chunk Overlap | int | No |
| Chunk Size | int | No |
| Separators | str | No |
| Output | Type |
|---|---|
| JSON | JSON |
Retrieval QA
Chain for question-answering querying sources from a retriever.
| Input | Type | Required |
|---|---|---|
| Language Model | other | Yes |
| Memory | other | No |
| Retriever | other | Yes |
| Input | str | Yes |
| Return Source Documents | bool | No |
Advanced options: Chain Type
| Output | Type |
|---|---|
| Text | Message |
Runnable Executor
Execute a runnable. It will try to guess the input and output keys.
| Input | Type | Required |
|---|---|---|
| Agent Executor | other | Yes |
| Input | str | Yes |
| Stream | bool | No |
Advanced options: Input Key, Output Key
| Output | Type |
|---|---|
| Message | Message |
SQLAgent
Construct an SQL agent from an LLM and tools.
| Input | Type | Required |
|---|---|---|
| watsonx API Endpoint | str | No |
| Database URI | str | Yes |
| Input | str | No |
| Language Model | model | Yes |
| watsonx Project ID | str | No |
Advanced options: Extra Tools, Agent Description [Deprecated], API Key, Handle Parse Errors, Max Iterations, Verbose
| Output | Type |
|---|---|
| Response | Message |
| Agent | AgentExecutor |
SQLDatabase
SQL Database
| Input | Type | Required |
|---|---|---|
| URI | str | Yes |
| Output | Type |
|---|---|
| SQLDatabase | SQLDatabase |
Self Query Retriever
Retriever that uses a vector store and an LLM to generate the vector store queries.
| Input | Type | Required |
|---|---|---|
| Metadata Field Info | other | No |
| LLM | other | No |
| Query | other | No |
| Vector Store | other | No |
| Document Content Description | str | No |
| Output | Type |
|---|---|
| Retrieved Documents | JSON |
Semantic Text Splitter
Split text into semantically meaningful chunks using semantic similarity.
| Input | Type | Required |
|---|---|---|
| Data Inputs | other | Yes |
| Embeddings | other | Yes |
| Breakpoint Threshold Amount | float | No |
| Breakpoint Threshold Type | str | No |
| Number of Chunks | int | No |
Advanced options: Buffer Size, Sentence Split Regex
| Output | Type |
|---|---|
| Chunks | JSON |
Spider Web Crawler & Scraper
Spider API for web crawling and scraping.
| Input | Type | Required |
|---|---|---|
| Mode | str | Yes |
| Additional Parameters | dict | No |
| Spider API Key | str | Yes |
| URL | str | Yes |
Advanced options: Blacklist, Depth, Limit, Metadata, Use Readability, Request Timeout, Whitelist
| Output | Type |
|---|---|
| Markdown | JSON |
Tool Calling Agent
An agent designed to utilize various tools seamlessly within workflows.
| Input | Type | Required |
|---|---|---|
| Tools | other | No |
| watsonx API Endpoint | str | No |
| Input | str | No |
| Language Model | model | Yes |
| watsonx Project ID | str | No |
| System Prompt | str | No |
Advanced options: Chat Memory, Agent Description [Deprecated], API Key, Handle Parse Errors, Max Iterations, Verbose
| Output | Type |
|---|---|
| Response | Message |
| Agent | AgentExecutor |
VectorStoreInfo
Information about a VectorStore
| Input | Type | Required |
|---|---|---|
| Vector Store | other | Yes |
| Description | str | Yes |
| Name | str | Yes |
| Output | Type |
|---|---|
| Vector Store Info | VectorStoreInfo |
VectorStoreRouterAgent
Construct an agent from a Vector Store Router.
| Input | Type | Required |
|---|---|---|
| Language Model | other | Yes |
| Vector Stores | other | Yes |
| Input | str | No |
Advanced options: Agent Description [Deprecated], Handle Parse Errors, Max Iterations, Verbose
| Output | Type |
|---|---|
| Response | Message |
| Agent | AgentExecutor |
XML Agent
Agent that uses tools formatting instructions as xml to the Language Model.
| Input | Type | Required |
|---|---|---|
| Tools | other | No |
| watsonx API Endpoint | str | No |
| Input | str | No |
| Language Model | model | Yes |
| watsonx Project ID | str | No |
| System Prompt | str | No |
| Prompt | str | No |
Advanced options: Chat History, Agent Description [Deprecated], API Key, Handle Parse Errors, Max Iterations, Verbose
| Output | Type |
|---|---|
| Response | Message |
| Agent | AgentExecutor |