import os import requests from openai import AzureOpenAI from dotenv import load_dotenv load_dotenv() # Config AZURE_SEARCH_ENDPOINT = os.getenv("AZURE_SEARCH_ENDPOINT") AZURE_SEARCH_KEY = os.getenv("AZURE_SEARCH_KEY") AZURE_SEARCH_INDEX_NAME = "it-ticket-solutions-index" HF_TOKEN = os.getenv("HF_TOKEN") API_BASE_URL = os.getenv("API_BASE_URL") MODEL_NAME = os.getenv("MODEL_NAME") # Use the specific deployment name for embeddings from your .env EMBEDDING_MODEL = os.getenv("AZURE_OPENAI_DEPLOYMENT") AZURE_OPENAI_API_VERSION = "2024-02-15-preview" # Clients openai_client = AzureOpenAI( api_key=HF_TOKEN, api_version=AZURE_OPENAI_API_VERSION, azure_endpoint=API_BASE_URL ) def embed_text(text: str): # FIX: Changed model from MODEL_NAME to EMBEDDING_MODEL response = openai_client.embeddings.create( input=[text], model=EMBEDDING_MODEL ) return response.data[0].embedding def search_similar_solution(query: str, category: str) -> str: embedding = embed_text(query) url = f"{AZURE_SEARCH_ENDPOINT}/indexes/{AZURE_SEARCH_INDEX_NAME}/docs/search?api-version=2023-07-01-Preview" headers = { "Content-Type": "application/json", "api-key": AZURE_SEARCH_KEY } payload = { "search": "", "vectors": [ { "value": embedding, "fields": "embedding", "k": 3 } ], "select": "category,problem,solution", "filter": f"category eq '{category}'" } response = requests.post(url, headers=headers, json=payload) if response.status_code != 200: return f"Error while searching: {response.text}" results = response.json().get("value", []) if not results: return "No matching solutions found." response_text = "" for idx, doc in enumerate(results, 1): response_text += ( f"\nResult {idx}:\n" f"Category: {doc.get('category')}\n" f"Problem: {doc.get('problem')}\n" f"Solution: {doc.get('solution')}\n" ) return response_text if __name__ == "__main__": query = "My Outlook crashes every time I open it" rag_output = search_similar_solution(query, category="Software Bug") print(rag_output)