app.py.old2

import gradio as gr
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer, BitsAndBytesConfig
from threading import Thread
import os
from selenium import webdriver
from selenium.webdriver.common.by import By
from selenium.webdriver.common.keys import Keys
from selenium.webdriver.chrome.options import Options
import time

# --- MODEL AYARLARI ---
MODEL_ID = "deepseek-ai/DeepSeek-V3"  # veya "deepseek-ai/DeepSeek-V3" dene

# 4-bit Quantization
quantization_config = BitsAndBytesConfig(
    load_in_4bit=True,
    bnb_4bit_compute_dtype=torch.float16,
    bnb_4bit_quant_type="nf4",
    bnb_4bit_use_double_quant=True,
)

print("🔄 Tokenizer yükleniyor...")
tokenizer = AutoTokenizer.from_pretrained(MODEL_ID, trust_remote_code=True)
if tokenizer.pad_token is None:
    tokenizer.pad_token = tokenizer.eos_token

print("🔄 Model yükleniyor (bu biraz sürebilir)...")
model = AutoModelForCausalLM.from_pretrained(
    MODEL_ID,
    quantization_config=quantization_config,
    device_map="auto",
    trust_remote_code=True
)
print("✅ Model başarıyla yüklendi!")

def search_google_selenium(query):
    """Selenium ile Google araması"""
    try:
        chrome_options = Options()
        chrome_options.add_argument("--headless")
        chrome_options.add_argument("--no-sandbox")
        chrome_options.add_argument("--disable-dev-shm-usage")
        chrome_options.add_argument("--disable-blink-features=AutomationControlled")
        chrome_options.add_argument("user-agent=Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36")
        
        driver = webdriver.Chrome(options=chrome_options)
        driver.get("https://www.google.com")
        time.sleep(1)
        
        search_box = driver.find_element(By.NAME, "q")
        search_box.send_keys(query)
        search_box.send_keys(Keys.RETURN)
        time.sleep(2)
        
        results = []
        result_blocks = driver.find_elements(By.CSS_SELECTOR, "div.g")
        
        for block in result_blocks[:3]:
            try:
                title = block.find_element(By.CSS_SELECTOR, "h3").text
                link = block.find_element(By.CSS_SELECTOR, "a").get_attribute("href")
                desc = block.find_element(By.CSS_SELECTOR, "div.VwiC3b").text
                results.append(f"Başlık: {title}\nLink: {link}\nÖzet: {desc}\n")
            except:
                continue
        
        driver.quit()
        return "\n".join(results) if results else "Sonuç bulunamadı."
    except Exception as e:
        return f"Arama hatası: {str(e)}"

def respond(message, history, internet_search):
    system_prompt = "NoiseAI/Noise adlı, 32B parametreli, MagnuX tarafından eğitilmiş bir yapay zekasın. Türkçe konuş ve en iyi kod pratiklerini uygula."
    
    full_prompt = f"<|im_start|>system\n{system_prompt}<|im_end|>\n"
    
    for user_msg, assistant_msg in history:
        if user_msg:
            full_prompt += f"<|im_start|>user\n{user_msg}<|im_end|>\n"
        if assistant_msg:
            full_prompt += f"<|im_start|>assistant\n{assistant_msg}<|im_end|>\n"

    if internet_search:
        search_results = search_google_selenium(message)
        message = f"Kullanıcı Sorusu: {message}\n\nİnternet Arama Sonuçları:\n{search_results}\n\nYukarıdaki bilgileri kullanarak cevap ver."

    full_prompt += f"<|im_start|>user\n{message}<|im_end|>\n<|im_start|>assistant\n"

    inputs = tokenizer([full_prompt], return_tensors="pt").to(model.device)
    streamer = TextIteratorStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)

    generation_kwargs = dict(
        inputs,
        streamer=streamer,
        max_new_tokens=2048,
        temperature=0.6,
        top_p=0.9,
        do_sample=True,
    )

    thread = Thread(target=model.generate, kwargs=generation_kwargs)
    thread.start()

    partial_text = ""
    for new_text in streamer:
        partial_text += new_text
        yield partial_text

# Gradio Arayüzü
demo = gr.ChatInterface(
    respond,
    title="NoiseAI (Local Engine) - 32B",
    description="Bu Space, modeli kendi donanımında (4-bit Quantized) çalıştırır.",
    additional_inputs=[
        gr.Checkbox(label="🌐 İnternette Ara", value=False)
    ]
)

if __name__ == "__main__":
    demo.launch(server_name="0.0.0.0", server_port=7860)