langchain-desearch

LangChain Desearch Integration

This project integrates the Desearch API with LangChain tools to enable various search and data-fetching functionalities, such as web searches, Twitter data retrieval, and AI-powered searches.

Features

• Grouped Tools:
  • Search Tools: General-purpose search tools for AI, web, and Twitter searches.
  • Twitter Tools: Tools specifically for Twitter-related operations.

Installation

Install the package using pip:

pip install langchain-desearch

Usage

Grouped Tools

Search Tools

The search_tools group contains tools for general-purpose searches:

• DesearchTool: Perform AI searches, web link searches, and Twitter post searches.
• BasicWebSearchTool: Conduct basic web searches.
• BasicTwitterSearchTool: Perform advanced Twitter searches with filters.

Twitter Tools

The twitter_tools group contains tools specifically for Twitter-related operations:

• BasicTwitterSearchTool: Perform a basic Twitter search using Desearch.
• FetchTweetsByUrlsTool: Retrieve tweets from specific URLs.
• FetchTweetsByIdTool: Fetch tweets using their unique IDs.
• FetchLatestTweetsTool: Get the latest tweets from a specific user.
• FetchTweetsAndRepliesByUserTool: Retrieve tweets and replies from a user.
• FetchRepliesByPostTool: Fetch replies to a specific Twitter post.
• FetchRetweetsByPostTool: Retrieve retweets of a specific post.
• FetchTwitterUserTool: Get detailed information about a Twitter user.

Examples

Using Tools

from langchain_desearch.tools import DesearchTool, BasicWebSearchTool, BasicTwitterSearchTool
from dotenv import load_dotenv
load_dotenv()

# Example 1: Using DesearchTool
tool = DesearchTool()
result = tool._run(
    prompt="Bittensor",
    tool=['web'],
    model="NOVA",
    date_filter="PAST_24_HOURS",
    streaming=False
)
print(result)

# Example 2: Using BasicWebSearchTool
tool = BasicWebSearchTool()
result = tool._run(
    query="Latest news on AI",
    num=5,
    start=1
)
print(result)

# Example 3: Using BasicTwitterSearchTool
tool = BasicTwitterSearchTool()
result = tool._run(
    query="AI trends",
    sort="Top",
    count=5
)
print(result)

Using RAG (Retrieval-Augmented Generation)

from langchain_core.prompts import PromptTemplate, ChatPromptTemplate
from langchain_core.runnables import RunnableLambda, RunnablePassthrough, RunnableParallel
from langchain_core.output_parsers import StrOutputParser
from langchain_desearch.tools import DesearchTool
from langchain_deepseek import ChatDeepSeek

# Setup Desearch Tool
desearch_tool = DesearchTool()

# Template to wrap Desearch output
document_prompt = PromptTemplate.from_template("""
<source>
    <result>{result}</result>
</source>
""")

# Retrieval chain using DesearchTool
def get_desearch_context(prompt: str) -> str:
    return desearch_tool._run(prompt=prompt, tool="desearch_web", model="NOVA")

retrieval_chain = RunnableLambda(lambda query: {
    "result": get_desearch_context(query)
}) | document_prompt | (lambda docs: docs.text)

# Prompt for RAG generation
generation_prompt = ChatPromptTemplate.from_messages([
    ("system", "You are an expert research assistant. You use xml-formatted context to research people's questions."),
    ("human", """
Please answer the following query based on the provided context. Please cite your sources at the end of your response.:

Query: {query}
---
<context>
{context}
</context>
""")
])

# Use DeepSeek for LLM
llm = ChatDeepSeek(
    model="deepseek-chat",
    temperature=0.7,
    max_tokens=None,
    timeout=None,
    max_retries=2
)

output_parser = StrOutputParser()

# Final chain
chain = RunnableParallel({
    "query": RunnablePassthrough(),
    "context": retrieval_chain,
}) | generation_prompt | llm | output_parser

# Run it!
query = "Recent trends in AI safety research"
result = chain.invoke(query)
print(result)

Using the LangChain Agent

from langchain_desearch.agent import create_search_agent
from langchain_deepseek import ChatDeepSeek

# Create a DeepSeek LLM instance
llm = ChatDeepSeek(
    model="deepseek-chat",
    temperature=0.7,
    max_tokens=None,
    timeout=None,
    max_retries=2
)

# Initialize the search agent
search_agent = create_search_agent(llm=llm)

# Use the agent to perform a task
state = {
    "input_message": "What's the latest news on AI?",
}
response = search_agent.invoke(state)
print(f"Agent Response: {response['output']}")

Running Tests

Dummy Tests

Run the dummy tests to verify the tools' functionality with mocked data:

pytest tests/test_tools.py

Real API Tests

Run the real tests to verify the tools' functionality with the Desearch API:

pytest tests/test_tools_real.py

Note: Ensure you have a valid DESEARCH_API_KEY in your .env file before running real tests.

Project Structure

langchain_desearch/
├── langchain_desearch/
│   ├── __init__.py
│   ├── tools.py
│   ├── search_tools.py
│   ├── agent.py
├── examples/
│   ├── tools.py
│   ├── RAG.py
│   ├── agent.py
├── tests/
│   ├── test_tools.py
│   ├── test_tools_real.py
├── .env
├── README.md
├── setup.py
├── requirements.txt

Contributing

• Fork the repository.
• Create a new branch for your feature or bug fix:

  git checkout -b feature-name
  

• Commit your changes:

  git commit -m "Add feature-name"
  

• Push to your branch:

  git push origin feature-name
  

• Create a pull request.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Contact

For questions or support, please contact [[e-mail]].