Create rag_agent.py

rag_agent.py

@@ -0,0 +1,86 @@
+import os
+import json
+import pandas as pd
+import faiss
+import numpy as np
+from phi.agent import Agent
+from phi.model.openai import OpenAIChat
+from phi.embedder.openai import OpenAIEmbedder
+from phi.vectordb.lancedb import LanceDb, SearchType
+from phi.knowledge.json import JSONKnowledgeBase
+
+# Load environment variables (API keys, etc.)
+from dotenv import load_dotenv
+load_dotenv()
+
+#####################################################################################
+#                               BNI RAG Agent (Auto-RAG)                            #
+#####################################################################################
+
+# Define paths
+BNI_CSV_FILE = "bni_pearl_chapter.csv"
+BNI_JSON_FILE = "bni_pearl_chapter.json"
+
+# 1️⃣ **Convert CSV to JSON for structured retrieval**
+def convert_csv_to_json():
+    df = pd.read_csv(BNI_CSV_FILE)
+    json_data = df.to_dict(orient="records")
+    
+    with open(BNI_JSON_FILE, "w") as file:
+        json.dump(json_data, file, indent=4)
+
+    print(f"✅ Converted {BNI_CSV_FILE} to JSON format.")
+
+# Run the conversion **only if JSON does not exist**
+if not os.path.exists(BNI_JSON_FILE):
+    convert_csv_to_json()
+
+# 2️⃣ **Create a Vector Database for RAG**
+vector_db = LanceDb(
+    table_name="bni_pearl",
+    uri="tmp/lancedb",
+    search_type=SearchType.vector,
+    embedder=OpenAIEmbedder(model="text-embedding-3-small"),
+)
+
+# 3️⃣ **Create Knowledge Base from JSON**
+knowledge_base = JSONKnowledgeBase(
+    file_path=BNI_JSON_FILE,  
+    vector_db=vector_db,  # Use vector DB for faster search
+)
+
+# **Load Knowledge Base** (Run this once)
+knowledge_base.load()
+
+# 4️⃣ **Define the RAG Agent**
+rag_agent = Agent(
+    model=OpenAIChat(id="gpt-4o"),
+    knowledge=knowledge_base,  # Attach the Knowledge Base
+    show_tool_calls=True,
+    markdown=True,
+)
+
+# 5️⃣ **Define Function to Recommend BNI Members**
+def recommend_bni_connections(company_data: str) -> str:
+    """
+    Retrieves BNI Pearl Chapter members who can help based on the user's company details.
+    
+    Args:
+        company_data (str): The user's company information.
+
+    Returns:
+        str: A list of relevant BNI members and how they can help.
+    """
+    query = f"""
+    Given the following business details:
+    - **Company Information:** {company_data}
+
+    Search the BNI Pearl Chapter database and recommend **3-5 members** who can 
+    help this business **through referrals, networking, or industry connections.** 
+    Provide:
+    1. **Member Name**
+    2. **Company & Industry**
+    3. **How They Can Help**
+    """
+    response = rag_agent.run(query)
+    return response.content  # Return recommended BNI connections