from agents import Agent, Tool, Runner
import numpy as np
from pinecone import Pinecone

# Initialize vector database
pc = Pinecone(api_key="YOUR_API_KEY")
index = pc.Index("mood-mnky-knowledge")

# Define vector search tool
def search_knowledge_base(query: str, limit: int = 5) -> list:
    """Search the knowledge base for relevant information."""
    # Generate embedding for the query
    embedding = generate_embedding(query)
    
    # Search in vector database
    results = index.query(
        vector=embedding,
        top_k=limit,
        include_metadata=True
    )
    
    return [item["metadata"] for item in results["matches"]]

# Create tool for the agent
knowledge_tool = Tool(
    name="search_knowledge_base",
    description="Search for information in the MOOD MNKY knowledge base",
    function=search_knowledge_base
)

# Create agent with the tool
agent = Agent(
    name="MOOD MNKY",
    instructions="You are MOOD MNKY, the brand assistant. Help customers with product information.",
    tools=[knowledge_tool]
)

# Run the agent
runner = Runner()
result = await runner.run(agent, "Tell me about your lavender-based products")

from agents import Agent, Runner
from agents.mcp.server import MCPServerStdio, MCPServerStdioParams

# Set up MCP server for knowledge base
async with MCPServerStdio(
    params=MCPServerStdioParams(
        command="node",
        args=["./mcp-servers/knowledge-server.js"]
    ),
    name="knowledge_mcp",
    cache_tools_list=True
) as knowledge_server:
    
    # Create agent with MCP server
    agent = Agent(
        name="MOOD MNKY",
        instructions="You are MOOD MNKY, the brand assistant. Help customers with product information.",
        mcp_servers=[knowledge_server]
    )
    
    # Run the agent
    runner = Runner()
    result = await runner.run(agent, "What ingredients are in your sleep mist?")

from agents import Agent, Runner

# Query knowledge base for agent instructions
instructions = get_agent_instructions("mood_mnky_001")

# Create agent with knowledge-infused instructions
agent = Agent(
    name="MOOD MNKY",
    instructions=instructions,
)

runner = Runner()
result = await runner.run(agent, "Tell me about your brand values")

from agents import Agent, Tool, Runner
from knowledge_base import KnowledgeBase

# Initialize knowledge base
kb = KnowledgeBase()

# Define knowledge lookup tool
def lookup_knowledge(topic: str) -> str:
    """Look up information on a specific topic."""
    return kb.get_knowledge_by_topic(topic)

# Create tool for the agent
knowledge_tool = Tool(
    name="lookup_knowledge",
    description="Look up information about MOOD MNKY products, policies, or brand",
    function=lookup_knowledge
)

# Create agent with the tool
agent = Agent(
    name="MOOD MNKY",
    instructions="You are MOOD MNKY, the brand assistant.",
    tools=[knowledge_tool]
)

from agents import Agent, Runner
from knowledge_base import KnowledgeBase

# Initialize knowledge base
kb = KnowledgeBase()

async def run_with_dynamic_context(query: str) -> dict:
    # Retrieve relevant knowledge based on the query
    relevant_knowledge = kb.retrieve_relevant(query)
    
    # Create agent with dynamic context
    agent = Agent(
        name="MOOD MNKY",
        instructions=f"""You are MOOD MNKY, the brand assistant. 
        Use the following information to answer questions:
        
        {relevant_knowledge}
        
        Always base your answers on the provided information."""
    )
    
    # Run the agent
    runner = Runner()
    return await runner.run(agent, query)

result = await run_with_dynamic_context("What's your return policy?")