Spaces:
Sleeping
Sleeping
File size: 2,760 Bytes
c4f0624 07bd23e c4f0624 07bd23e c4f0624 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 |
from openai import OpenAI
from dotenv import load_dotenv
import os
load_dotenv()
class LLMHandler:
def __init__(self, model_name="gpt-4o-mini"):
"""
Initializes the LLMHandler with the specified OpenAI model.
"""
self.openai_api_key = os.getenv("OPENAI_API_KEY")
if not self.openai_api_key:
raise ValueError("OPENAI_API_KEY environment variable not set.")
# Initialize OpenAI client
self.client = OpenAI(api_key=self.openai_api_key)
self.model_name = model_name
def generate_response(self, user_prompt, data):
"""
Generate a concise response using the LLM based on user prompt and data.
:param user_prompt: Prompt provided by the user.
:param data: Dictionary containing the instance information.
:return: Generated response text.
"""
# Refined prompt to handle encoding and formatting
prompt = (
f"You are a professional AI model tasked with writing personalized invite texts "
f"that are concise (less than 40 words), brochure-suitable, and tailored as per the category in the given sample.\n\n"
f"Consider the user prompt: {user_prompt}\n\n"
f"Details of the individual:\n"
f"- Name: {data['Name']}\n"
f"- Job Title: {data['Job Title']}\n"
f"- Organisation: {data['Organisation']}\n"
f"- Area of Interest: {data['Area of Interest']}\n"
f"- Category: {data['Category']}\n\n"
f"The response **MUST**:\n"
f"- Start with 'Hello {data['Name']}'.\n"
f"- Be concise, professional, and STRICTLY DO NOT generate invalid characters or encoding errors (e.g. 'SoraVR’s').\n"
f"- Use standard English punctuation, such as single quotes (e.g., 'can't', 'it's').\n"
f"- STRICTLY Give only one response for the Category the sample belongs to.\n"
f"- Do NOT include preambles or unnecessary text.\n\n"
f"Return the final response cleanly, without any extraneous symbols or characters."
)
# Query the OpenAI client and return the response
completion = self.client.chat.completions.create(
model=self.model_name,
messages=[
{"role": "system", "content": "You are a professional assistant."},
{"role": "user", "content": prompt},
]
)
# Extract and clean the generated response
response = completion.choices[0].message.content.strip()
# Optional: Post-process to clean invalid characters
#response_cleaned = response.encode('utf-8').decode('utf-8', errors='ignore')
return response
|