far-search-tool 0.1.0

LangChain tool for semantic search over Federal Acquisition Regulations (FAR)

FAR Search Tool for LangChain

A LangChain tool for semantic search over Federal Acquisition Regulations (FAR).

Enable your AI agents to search and understand U.S. government contracting regulations, procurement rules, and compliance requirements.

Installation

pip install far-search-tool

Quick Start

Basic Usage (Free Tier)

from far_search import FARSearchTool

# Initialize the tool
tool = FARSearchTool()

# Search FAR regulations
result = tool.invoke({
    "query": "small business set aside requirements",
    "top_k": 5
})

print(result)

With RapidAPI Key (Higher Limits)

For production use with higher rate limits, get an API key from RapidAPI.

from far_search import FARSearchTool

tool = FARSearchTool(rapidapi_key="your-rapidapi-key")

result = tool.invoke({
    "query": "cybersecurity requirements for contractors"
})

Use with LangChain Agents

from langchain.agents import initialize_agent, AgentType
from langchain_openai import ChatOpenAI
from far_search import FARSearchTool

# Initialize LLM and tool
llm = ChatOpenAI(model="gpt-4")
tools = [FARSearchTool()]

# Create agent
agent = initialize_agent(
    tools=tools,
    llm=llm,
    agent=AgentType.OPENAI_FUNCTIONS,
    verbose=True
)

# Ask about regulations
response = agent.run(
    "What are the FAR requirements for small business subcontracting plans?"
)
print(response)

Use with LangGraph

from langgraph.prebuilt import create_react_agent
from langchain_openai import ChatOpenAI
from far_search import FARSearchTool

# Create agent with FAR search capability
agent = create_react_agent(
    ChatOpenAI(model="gpt-4"),
    tools=[FARSearchTool()]
)

# Query the agent
result = agent.invoke({
    "messages": [("user", "What cybersecurity clauses should I include in a DoD contract?")]
})

Features

• Semantic Search: Find relevant regulations using natural language queries
• Pre-vectorized Data: 617 FAR clauses with pre-computed embeddings for fast search
• LLM-Optimized Output: Results formatted for easy consumption by language models
• Retry Logic: Built-in handling for transient failures
• Dual API Support: Free tier or RapidAPI for production use

API Reference

FARSearchTool

FARSearchTool(
    rapidapi_key: str = None,      # Optional RapidAPI key for paid tier
    base_url: str = None,          # Override API URL (for self-hosted)
    timeout: int = 30,             # Request timeout in seconds
    max_retries: int = 2           # Retry attempts on failure
)

Input Schema

Parameter	Type	Default	Description
query	str	required	Natural language search query
top_k	int	5	Number of results (1-20)

Output

Returns a formatted string containing matching FAR clauses with:

• Clause ID and title
• Relevance score
• Source reference
• URL to official documentation
• Clause text (truncated if long)

Example Queries

• "Small business set aside requirements"
• "Cybersecurity contract clauses"
• "Payment terms for government contracts"
• "Contractor ethics and conduct rules"
• "Cost accounting standards"
• "Intellectual property rights in contracts"
• "Subcontracting plan requirements"
• "Contract termination procedures"

Pricing

Tier	Rate Limit	Price
Free	100 requests/day	$0
Pro	5,000 requests/month	$29/month
Ultra	150,000 requests/month	$199/month

Get your API key at RapidAPI.

Error Handling

from far_search import FARSearchTool, FARAPIError, FARRateLimitError

tool = FARSearchTool()

try:
    result = tool.invoke({"query": "my query"})
except FARRateLimitError:
    print("Rate limit exceeded. Upgrade your plan or wait.")
except FARAPIError as e:
    print(f"API error: {e}")

Requirements

• Python 3.9+
• langchain >= 0.1.0
• requests >= 2.28.0
• pydantic >= 2.0.0

License

MIT License - see LICENSE for details.

Links

• API Documentation
• GitHub Repository
• FAR Official Website

Contributing

Contributions welcome! Please open an issue or submit a pull request.