app.py

# =============================================================================
# DON'T STEAL THE FREE CODE OF DEVS! Use it for free an do not touch credits!
# If you steal this code, in the future you will pay for apps like this!
# A bit of respect goes a long way – all rights reserved under German law.
# Copyright Volkan Kücükbudak https://github.com/volkansah
# Repo URL: https://github.com/AiCodeCraft
# =============================================================================
import streamlit as st
import os
import json
import datetime
import openai
from datetime import timedelta
import logging
from datasets import load_dataset, Dataset, concatenate_datasets

# ------------------ Logging konfigurieren ------------------
logging.basicConfig(
    level=logging.INFO,
    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
)
logger = logging.getLogger(__name__)
logger.info("Starte App mit HF-Dataset Memory...")

# ------------------ Hugging Face Token laden ------------------
HF_TOKEN_MEMORY = os.getenv('HF_TOKEN_MEMORY', '').strip()
if HF_TOKEN_MEMORY:
    logger.info("Hugging Face Token gefunden.")
else:
    logger.warning("Kein Hugging Face Token gefunden. Falls benötigt, bitte setzen!")

# ------------------ Einstellungen für das Memory-Dataset ------------------
DATASET_REPO = "AiCodeCarft/customer_memory"

def load_memory_dataset():
    try:
        ds = load_dataset(DATASET_REPO, split="train")
        logger.info("Dataset erfolgreich vom HF Hub geladen.")
    except Exception as e:
        logger.info("Kein Dataset gefunden. Erstelle ein neues Dataset...")
        data = {"user_id": [], "query": [], "response": []}
        ds = Dataset.from_dict(data)
        ds.push_to_hub(DATASET_REPO)
        logger.info("Neues Dataset erfolgreich erstellt und gepusht.")
    return ds

def add_to_memory(user_id, query, response):
    ds = load_memory_dataset()
    new_entry = Dataset.from_dict({
        "user_id": [user_id],
        "query": [query],
        "response": [response]
    })
    updated_ds = concatenate_datasets([ds, new_entry])
    updated_ds.push_to_hub(DATASET_REPO)
    logger.info("Memory-Dataset erfolgreich aktualisiert.")

def get_memory(user_id):
    ds = load_memory_dataset()
    filtered_ds = ds.filter(lambda x: x["user_id"] == user_id)
    logger.info(f"Memory für User {user_id} abgerufen. {len(filtered_ds)} Einträge gefunden.")
    return filtered_ds

# ------------------ Streamlit App UI ------------------
st.title("AI Customer Support Agent with Memory 🛒")
st.caption("Chat with a customer support assistant who remembers your past interactions.")

# OpenAI API Key Eingabe oben in der Haupt-UI
openai_api_key = st.text_input("Enter OpenAI API Key", type="password")

if not openai_api_key:
    st.warning("⚠️ Please enter your OpenAI API key to continue.")
    st.stop()

openai.api_key = openai_api_key  # Direktes Setzen des API-Keys

# ------------------ Klasse: CustomerSupportAIAgent ------------------
class CustomerSupportAIAgent:
    def __init__(self):
        self.client = openai
        self.app_id = "customer-support"

    def handle_query(self, query, user_id=None):
        try:
            memories = get_memory(user_id)
            context = "Relevant past information:\n"
            if len(memories) > 0:
                for entry in memories:
                    context += f"- Query: {entry['query']}\n  Response: {entry['response']}\n"
            
            full_prompt = f"{context}\nCustomer: {query}\nSupport Agent:"
            
            # API-Key wird direkt übergeben
            answer = self.client.ChatCompletion.create(
                model="gpt-3.5-turbo",
                messages=[
                    {"role": "system", "content": "You are a customer support AI for TechGadgets.com."},
                    {"role": "user", "content": full_prompt}
                ]
            ).choices[0].message.content
            
            add_to_memory(user_id, query, answer)
            return answer
        except Exception as e:
            logger.error(f"Fehler bei handle_query: {e}")
            return "Sorry, I encountered an error. Please try again later."

# ------------------ Initialisierung ------------------
support_agent = CustomerSupportAIAgent()

# ------------------ Sidebar-Komponenten ------------------
with st.sidebar:
    st.title("Customer ID")
    customer_id = st.text_input("Enter your Customer ID", key="customer_id")
    
    if 'customer_id' in st.session_state and st.session_state.customer_id:
        if st.button("Generate Synthetic Data"):
            # ... (deine bestehende Synthetic Data Logik)

# ------------------ Chat-History Management ------------------
if "messages" not in st.session_state:
    st.session_state.messages = []

# ------------------ Chat-Eingabe ------------------
if prompt := st.chat_input("How can I assist you today?"):
    if not customer_id:
        st.error("❌ Please enter a customer ID first")
        st.stop()
    
    st.session_state.messages.append({"role": "user", "content": prompt})
    
    with st.spinner("Generating response..."):
        response = support_agent.handle_query(prompt, customer_id)
    
    st.session_state.messages.append({"role": "assistant", "content": response})

# ------------------ Nachrichten anzeigen ------------------
for message in st.session_state.messages:
    with st.chat_message(message["role"]):
        st.markdown(message["content"])