app.py

#!/usr/bin/env python

import os
import re
import tempfile
import gc  # Added garbage collector
from collections.abc import Iterator
from threading import Thread
import json
import requests
import cv2
import base64
import logging
import time
from urllib.parse import quote  # Added for URL encoding
import importlib  # For dynamic import

import gradio as gr
import spaces
import torch
from loguru import logger
from PIL import Image
from transformers import AutoProcessor, Gemma3ForConditionalGeneration, TextIteratorStreamer

# CSV/TXT/PDF analysis
import pandas as pd
import PyPDF2

# =============================================================================
# (New) Image API related functions
# =============================================================================
from gradio_client import Client

API_URL = "http://211.233.58.201:7896"

logging.basicConfig(
    level=logging.DEBUG,
    format='%(asctime)s - %(levelname)s - %(message)s'
)

def test_api_connection() -> str:
    """Test API server connection"""
    try:
        client = Client(API_URL)
        return "API connection successful: Operating normally"
    except Exception as e:
        logging.error(f"API connection test failed: {e}")
        return f"API connection failed: {e}"

def generate_image(prompt: str, width: float, height: float, guidance: float, inference_steps: float, seed: float):
    """Image generation function (flexible return types)"""
    if not prompt:
        return None, "Error: A prompt is required."
    try:
        logging.info(f"Calling image generation API with prompt: {prompt}")
        
        client = Client(API_URL)
        result = client.predict(
            prompt=prompt,
            width=int(width),
            height=int(height),
            guidance=float(guidance),
            inference_steps=int(inference_steps),
            seed=int(seed),
            do_img2img=False,
            init_image=None,
            image2image_strength=0.8,
            resize_img=True,
            api_name="/generate_image"
        )
        
        logging.info(f"Image generation result: {type(result)}, length: {len(result) if isinstance(result, (list, tuple)) else 'unknown'}")
        
        # Handle cases where the result is a tuple or list
        if isinstance(result, (list, tuple)) and len(result) > 0:
            image_data = result[0]  # The first element is the image data
            seed_info = result[1] if len(result) > 1 else "Unknown seed"
            return image_data, seed_info
        else:
            # When a single value is returned
            return result, "Unknown seed"
            
    except Exception as e:
        logging.error(f"Image generation failed: {str(e)}")
        return None, f"Error: {str(e)}"

def fix_base64_padding(data):
    """Fix the padding of a Base64 string."""
    if isinstance(data, bytes):
        data = data.decode('utf-8')
    
    if "base64," in data:
        data = data.split("base64,", 1)[1]
    
    missing_padding = len(data) % 4
    if missing_padding:
        data += '=' * (4 - missing_padding)
    
    return data

def clear_cuda_cache():
    """Explicitly clear the CUDA cache."""
    if torch.cuda.is_available():
        torch.cuda.empty_cache()
        gc.collect()

SERPHOUSE_API_KEY = os.getenv("SERPHOUSE_API_KEY", "")

def extract_keywords(text: str, top_k: int = 5) -> str:
    """Simple keyword extraction: only keep English, Korean, numbers, and spaces."""
    text = re.sub(r"[^a-zA-Z0-9가-힣\s]", "", text)
    tokens = text.split()
    return " ".join(tokens[:top_k])

def do_web_search(query: str) -> str:
    """Call the SerpHouse LIVE API to return Markdown formatted search results"""
    try:
        url = "https://api.serphouse.com/serp/live"
        params = {
            "q": query,
            "domain": "google.com",
            "serp_type": "web",
            "device": "desktop",
            "lang": "en",
            "num": "20"
        }
        headers = {"Authorization": f"Bearer {SERPHOUSE_API_KEY}"}
        logger.info(f"Calling SerpHouse API with query: {query}")
        response = requests.get(url, headers=headers, params=params, timeout=60)
        response.raise_for_status()
        data = response.json()
        results = data.get("results", {})
        organic = None
        if isinstance(results, dict) and "organic" in results:
            organic = results["organic"]
        elif isinstance(results, dict) and "results" in results:
            if isinstance(results["results"], dict) and "organic" in results["results"]:
                organic = results["results"]["organic"]
        elif "organic" in data:
            organic = data["organic"]
        if not organic:
            logger.warning("Organic results not found in response.")
            return "No web search results available or the API response structure is unexpected."
        max_results = min(20, len(organic))
        limited_organic = organic[:max_results]
        summary_lines = []
        for idx, item in enumerate(limited_organic, start=1):
            title = item.get("title", "No Title")
            link = item.get("link", "#")
            snippet = item.get("snippet", "No Description")
            displayed_link = item.get("displayed_link", link)
            summary_lines.append(
                f"### Result {idx}: {title}\n\n"
                f"{snippet}\n\n"
                f"**Source**: [{displayed_link}]({link})\n\n"
                f"---\n"
            )
        instructions = """
# Web Search Results
Below are the search results. Use this information to answer the query:
1. Refer to each result's title, description, and source link.
2. In your answer, explicitly cite the source of any used information (e.g., "[Source Title](link)").
3. Include the actual source links in your response.
4. Synthesize information from multiple sources.
5. At the end include a "References:" section listing the main source links.
"""
        return instructions + "\n".join(summary_lines)
    except Exception as e:
        logger.error(f"Web search failed: {e}")
        return f"Web search failed: {str(e)}"

MAX_CONTENT_CHARS = 2000
MAX_INPUT_LENGTH = 2096
model_id = os.getenv("MODEL_ID", "VIDraft/Gemma-3-R1984-4B")
processor = AutoProcessor.from_pretrained(model_id, padding_side="left")
model = Gemma3ForConditionalGeneration.from_pretrained(
    model_id,
    device_map="auto",
    torch_dtype=torch.bfloat16,
    attn_implementation="eager"
)
MAX_NUM_IMAGES = int(os.getenv("MAX_NUM_IMAGES", "5"))

def analyze_csv_file(path: str) -> str:
    try:
        df = pd.read_csv(path)
        if df.shape[0] > 50 or df.shape[1] > 10:
            df = df.iloc[:50, :10]
        df_str = df.to_string()
        if len(df_str) > MAX_CONTENT_CHARS:
            df_str = df_str[:MAX_CONTENT_CHARS] + "\n...(truncated)..."
        return f"**[CSV File: {os.path.basename(path)}]**\n\n{df_str}"
    except Exception as e:
        return f"CSV file read failed ({os.path.basename(path)}): {str(e)}"

def analyze_txt_file(path: str) -> str:
    try:
        with open(path, "r", encoding="utf-8") as f:
            text = f.read()
        if len(text) > MAX_CONTENT_CHARS:
            text = text[:MAX_CONTENT_CHARS] + "\n...(truncated)..."
        return f"**[TXT File: {os.path.basename(path)}]**\n\n{text}"
    except Exception as e:
        return f"TXT file read failed ({os.path.basename(path)}): {str(e)}"

def pdf_to_markdown(pdf_path: str) -> str:
    text_chunks = []
    try:
        with open(pdf_path, "rb") as f:
            reader = PyPDF2.PdfReader(f)
            max_pages = min(5, len(reader.pages))
            for page_num in range(max_pages):
                page_text = reader.pages[page_num].extract_text() or ""
                page_text = page_text.strip()
                if page_text:
                    if len(page_text) > MAX_CONTENT_CHARS // max_pages:
                        page_text = page_text[:MAX_CONTENT_CHARS // max_pages] + "...(truncated)"
                    text_chunks.append(f"## Page {page_num+1}\n\n{page_text}\n")
            if len(reader.pages) > max_pages:
                text_chunks.append(f"\n...(Displaying only {max_pages} out of {len(reader.pages)} pages)...")
    except Exception as e:
        return f"PDF file read failed ({os.path.basename(pdf_path)}): {str(e)}"
    full_text = "\n".join(text_chunks)
    if len(full_text) > MAX_CONTENT_CHARS:
        full_text = full_text[:MAX_CONTENT_CHARS] + "\n...(truncated)..."
    return f"**[PDF File: {os.path.basename(pdf_path)}]**\n\n{full_text}"

def count_files_in_new_message(paths: list[str]) -> tuple[int, int]:
    image_count = 0
    video_count = 0
    for path in paths:
        if path.endswith(".mp4"):
            video_count += 1
        elif re.search(r"\.(png|jpg|jpeg|gif|webp)$", path, re.IGNORECASE):
            image_count += 1
    return image_count, video_count

def count_files_in_history(history: list[dict]) -> tuple[int, int]:
    image_count = 0
    video_count = 0
    for item in history:
        if item["role"] != "user" or isinstance(item["content"], str):
            continue
        if isinstance(item["content"], list) and len(item["content"]) > 0:
            file_path = item["content"][0]
            if isinstance(file_path, str):
                if file_path.endswith(".mp4"):
                    video_count += 1
                elif re.search(r"\.(png|jpg|jpeg|gif|webp)$", file_path, re.IGNORECASE):
                    image_count += 1
    return image_count, video_count

def validate_media_constraints(message: dict, history: list[dict]) -> bool:
    media_files = [f for f in message["files"] if re.search(r"\.(png|jpg|jpeg|gif|webp)$", f, re.IGNORECASE) or f.endswith(".mp4")]
    new_image_count, new_video_count = count_files_in_new_message(media_files)
    history_image_count, history_video_count = count_files_in_history(history)
    image_count = history_image_count + new_image_count
    video_count = history_video_count + new_video_count
    if video_count > 1:
        gr.Warning("Only one video file is supported.")
        return False
    if video_count == 1:
        if image_count > 0:
            gr.Warning("Mixing images and a video is not allowed.")
            return False
        if "<image>" in message["text"]:
            gr.Warning("The <image> tag cannot be used together with a video file.")
            return False
    if video_count == 0 and image_count > MAX_NUM_IMAGES:
        gr.Warning(f"You can upload a maximum of {MAX_NUM_IMAGES} images.")
        return False
    if "<image>" in message["text"]:
        image_files = [f for f in message["files"] if re.search(r"\.(png|jpg|jpeg|gif|webp)$", f, re.IGNORECASE)]
        image_tag_count = message["text"].count("<image>")
        if image_tag_count != len(image_files):
            gr.Warning("The number of <image> tags does not match the number of image files provided.")
            return False
    return True

def downsample_video(video_path: str) -> list[tuple[Image.Image, float]]:
    vidcap = cv2.VideoCapture(video_path)
    fps = vidcap.get(cv2.CAP_PROP_FPS)
    total_frames = int(vidcap.get(cv2.CAP_PROP_FRAME_COUNT))
    frame_interval = max(int(fps), int(total_frames / 10))
    frames = []
    for i in range(0, total_frames, frame_interval):
        vidcap.set(cv2.CAP_PROP_POS_FRAMES, i)
        success, image = vidcap.read()
        if success:
            image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
            image = cv2.resize(image, (0, 0), fx=0.5, fy=0.5)
            pil_image = Image.fromarray(image)
            timestamp = round(i / fps, 2)
            frames.append((pil_image, timestamp))
            if len(frames) >= 5:
                break
    vidcap.release()
    return frames

def process_video(video_path: str) -> tuple[list[dict], list[str]]:
    content = []
    temp_files = []
    frames = downsample_video(video_path)
    for pil_image, timestamp in frames:
        with tempfile.NamedTemporaryFile(delete=False, suffix=".png") as temp_file:
            pil_image.save(temp_file.name)
            temp_files.append(temp_file.name)
            content.append({"type": "text", "text": f"Frame {timestamp}:"})
            content.append({"type": "image", "url": temp_file.name})
    return content, temp_files

def process_interleaved_images(message: dict) -> list[dict]:
    parts = re.split(r"(<image>)", message["text"])
    content = []
    image_files = [f for f in message["files"] if re.search(r"\.(png|jpg|jpeg|gif|webp)$", f, re.IGNORECASE)]
    image_index = 0
    for part in parts:
        if part == "<image>" and image_index < len(image_files):
            content.append({"type": "image", "url": image_files[image_index]})
            image_index += 1
        elif part.strip():
            content.append({"type": "text", "text": part.strip()})
        else:
            if isinstance(part, str) and part != "<image>":
                content.append({"type": "text", "text": part})
    return content

def is_image_file(file_path: str) -> bool:
    return bool(re.search(r"\.(png|jpg|jpeg|gif|webp)$", file_path, re.IGNORECASE))

def is_video_file(file_path: str) -> bool:
    return file_path.endswith(".mp4")

def is_document_file(file_path: str) -> bool:
    return file_path.lower().endswith(".pdf") or file_path.lower().endswith(".csv") or file_path.lower().endswith(".txt")

def process_new_user_message(message: dict) -> tuple[list[dict], list[str]]:
    temp_files = []
    if not message["files"]:
        return [{"type": "text", "text": message["text"]}], temp_files
    video_files = [f for f in message["files"] if is_video_file(f)]
    image_files = [f for f in message["files"] if is_image_file(f)]
    csv_files = [f for f in message["files"] if f.lower().endswith(".csv")]
    txt_files = [f for f in message["files"] if f.lower().endswith(".txt")]
    pdf_files = [f for f in message["files"] if f.lower().endswith(".pdf")]
    content_list = [{"type": "text", "text": message["text"]}]
    for csv_path in csv_files:
        content_list.append({"type": "text", "text": analyze_csv_file(csv_path)})
    for txt_path in txt_files:
        content_list.append({"type": "text", "text": analyze_txt_file(txt_path)})
    for pdf_path in pdf_files:
        content_list.append({"type": "text", "text": pdf_to_markdown(pdf_path)})
    if video_files:
        video_content, video_temp_files = process_video(video_files[0])
        content_list += video_content
        temp_files.extend(video_temp_files)
        return content_list, temp_files
    if "<image>" in message["text"] and image_files:
        interleaved_content = process_interleaved_images({"text": message["text"], "files": image_files})
        if content_list and content_list[0]["type"] == "text":
            content_list = content_list[1:]
        return interleaved_content + content_list, temp_files
    else:
        for img_path in image_files:
            content_list.append({"type": "image", "url": img_path})
    return content_list, temp_files

def process_history(history: list[dict]) -> list[dict]:
    messages = []
    current_user_content = []
    for item in history:
        if item["role"] == "assistant":
            if current_user_content:
                messages.append({"role": "user", "content": current_user_content})
                current_user_content = []
            messages.append({"role": "assistant", "content": [{"type": "text", "text": item["content"]}]})
        else:
            content = item["content"]
            if isinstance(content, str):
                current_user_content.append({"type": "text", "text": content})
            elif isinstance(content, list) and len(content) > 0:
                file_path = content[0]
                if is_image_file(file_path):
                    current_user_content.append({"type": "image", "url": file_path})
                else:
                    current_user_content.append({"type": "text", "text": f"[File: {os.path.basename(file_path)}]"})
    if current_user_content:
        messages.append({"role": "user", "content": current_user_content})
    return messages

def _model_gen_with_oom_catch(**kwargs):
    try:
        model.generate(**kwargs)
    except torch.cuda.OutOfMemoryError:
        raise RuntimeError("[OutOfMemoryError] Insufficient GPU memory.")
    finally:
        clear_cuda_cache()

# =============================================================================
# JSON 기반 함수 목록 로드
# =============================================================================
def load_function_definitions(json_path="functions.json"):
    """
    로컬 JSON 파일에서 함수 정의 목록을 로드하여 반환.
    """
    try:
        with open(json_path, "r", encoding="utf-8") as f:
            data = json.load(f)
        func_dict = {}
        for entry in data:
            func_name = entry["name"]
            func_dict[func_name] = entry
        return func_dict
    except Exception as e:
        logger.error(f"Failed to load function definitions from JSON: {e}")
        return {}

FUNCTION_DEFINITIONS = load_function_definitions("functions.json")

def handle_function_call(text: str) -> str:
    """
    Detects and processes function call blocks in the text using the JSON-based approach.
    The model is expected to produce something like:
    ```tool_code
    get_stock_price(ticker="AAPL")
    ```
    or
    ```tool_code
    get_product_name_by_PID(PID="807ZPKBL9V")
    ```
    """
    import re
    pattern = r"```tool_code\s*(.*?)\s*```"
    match = re.search(pattern, text, re.DOTALL)
    if not match:
        return ""
    code_block = match.group(1).strip()

    func_match = re.match(r'^(\w+)\((.*)\)$', code_block)
    if not func_match:
        logger.debug("No valid function call format found.")
        return ""

    func_name = func_match.group(1)
    param_str = func_match.group(2).strip()

    # JSON에서 해당 함수가 정의되어 있는지 확인
    if func_name not in FUNCTION_DEFINITIONS:
        logger.warning(f"Function '{func_name}' not found in definitions.")
        return "```tool_output\nError: Function not found.\n```"

    func_info = FUNCTION_DEFINITIONS[func_name]
    module_path = func_info["module_path"]
    module_func_name = func_info["func_name_in_module"]

    try:
        imported_module = importlib.import_module(module_path)
    except ImportError as e:
        logger.error(f"Failed to import module {module_path}: {e}")
        return f"```tool_output\nError: Cannot import module '{module_path}'\n```"

    if not hasattr(imported_module, module_func_name):
        logger.error(f"Module '{module_path}' has no attribute '{module_func_name}'.")
        return f"```tool_output\nError: Function '{module_func_name}' not found in module '{module_path}'\n```"

    real_func = getattr(imported_module, module_func_name)

    # 간단 파라미터 파싱 (key="value" or key=123)
    param_pattern = r'(\w+)\s*=\s*"(.*?)"|(\w+)\s*=\s*([\d.]+)'
    param_dict = {}
    for p_match in re.finditer(param_pattern, param_str):
        if p_match.group(1) and p_match.group(2):
            key = p_match.group(1)
            val = p_match.group(2)
            param_dict[key] = val
        else:
            key = p_match.group(3)
            val = p_match.group(4)
            if '.' in val:
                param_dict[key] = float(val)
            else:
                param_dict[key] = int(val)

    try:
        result = real_func(**param_dict)
    except Exception as e:
        logger.error(f"Error executing function '{func_name}': {e}")
        return f"```tool_output\nError: {str(e)}\n```"

    return f"```tool_output\n{result}\n```"

@spaces.GPU(duration=120)
def run(
    message: dict,
    history: list[dict],
    system_prompt: str = "",
    max_new_tokens: int = 512,
    use_web_search: bool = False,
    web_search_query: str = "",
    age_group: str = "20s",
    mbti_personality: str = "INTP",
    sexual_openness: int = 2,
    image_gen: bool = False
) -> Iterator[str]:
    if not validate_media_constraints(message, history):
        yield ""
        return
    temp_files = []
    try:
        # JSON에서 로드된 함수 정보 문자열화 (예: 함수명과 example_usage만)
        available_funcs_text = ""
        for f_name, info in FUNCTION_DEFINITIONS.items():
            example_usage = info.get("example_usage", "")
            available_funcs_text += f"\n\nFunction: {f_name}\nDescription: {info['description']}\nExample:\n{example_usage}\n"

        persona = (
            f"{system_prompt.strip()}\n\n"
            f"Gender: Female\n"
            f"Age Group: {age_group}\n"
            f"MBTI Persona: {mbti_personality}\n"
            f"Sexual Openness (1-5): {sexual_openness}\n\n"
            "Below are the available functions you can call.\n"
            "Important: Use the format exactly like: ```tool_code\nfunctionName(param=\"string\", ...)\n```\n"
            "(Strings must be in double quotes)\n"
            f"{available_funcs_text}\n"
        )
        combined_system_msg = f"[System Prompt]\n{persona.strip()}\n\n"

        if use_web_search:
            user_text = message["text"]
            ws_query = extract_keywords(user_text)
            if ws_query.strip():
                logger.info(f"[Auto web search keywords] {ws_query!r}")
                ws_result = do_web_search(ws_query)
                combined_system_msg += f"[Search Results (Top 20 Items)]\n{ws_result}\n\n"
                combined_system_msg += (
                    "[Note: In your answer, cite the above search result links as sources]\n"
                    "[Important Instructions]\n"
                    "1. Include a citation in the format \"[Source Title](link)\" for any information from the search results.\n"
                    "2. Synthesize information from multiple sources when answering.\n"
                    "3. At the end, add a \"References:\" section listing the main source links.\n"
                )
            else:
                combined_system_msg += "[No valid keywords found; skipping web search]\n\n"

        messages = []
        if combined_system_msg.strip():
            messages.append({"role": "system", "content": [{"type": "text", "text": combined_system_msg.strip()}]})

        messages.extend(process_history(history))
        user_content, user_temp_files = process_new_user_message(message)
        temp_files.extend(user_temp_files)
        for item in user_content:
            if item["type"] == "text" and len(item["text"]) > MAX_CONTENT_CHARS:
                item["text"] = item["text"][:MAX_CONTENT_CHARS] + "\n...(truncated)..."
        messages.append({"role": "user", "content": user_content})
        inputs = processor.apply_chat_template(
            messages,
            add_generation_prompt=True,
            tokenize=True,
            return_dict=True,
            return_tensors="pt",
        ).to(device=model.device, dtype=torch.bfloat16)
        if inputs.input_ids.shape[1] > MAX_INPUT_LENGTH:
            inputs.input_ids = inputs.input_ids[:, -MAX_INPUT_LENGTH:]
            if 'attention_mask' in inputs:
                inputs.attention_mask = inputs.attention_mask[:, -MAX_INPUT_LENGTH:]

        streamer = TextIteratorStreamer(processor, timeout=30.0, skip_prompt=True, skip_special_tokens=True)
        gen_kwargs = dict(inputs, streamer=streamer, max_new_tokens=max_new_tokens)
        t = Thread(target=_model_gen_with_oom_catch, kwargs=gen_kwargs)
        t.start()
        output_so_far = ""
        for new_text in streamer:
            output_so_far += new_text
            yield output_so_far

        func_result = handle_function_call(output_so_far)
        if func_result:
            output_so_far += "\n\n" + func_result
            yield output_so_far

    except Exception as e:
        logger.error(f"Error in run function: {str(e)}")
        yield f"Sorry, an error occurred: {str(e)}"
    finally:
        for tmp in temp_files:
            try:
                if os.path.exists(tmp):
                    os.unlink(tmp)
                    logger.info(f"Temporary file deleted: {tmp}")
            except Exception as ee:
                logger.warning(f"Failed to delete temporary file {tmp}: {ee}")
        try:
            del inputs, streamer
        except Exception:
            pass
        clear_cuda_cache()

def modified_run(message, history, system_prompt, max_new_tokens, use_web_search, web_search_query,
                age_group, mbti_personality, sexual_openness, image_gen):
    output_so_far = ""
    gallery_update = gr.Gallery(visible=False, value=[])
    yield output_so_far, gallery_update
    
    text_generator = run(message, history, system_prompt, max_new_tokens, use_web_search, 
                       web_search_query, age_group, mbti_personality, sexual_openness, image_gen)
    
    for text_chunk in text_generator:
        output_so_far = text_chunk
        yield output_so_far, gallery_update
        
    if image_gen and message["text"].strip():
        try:
            width, height = 512, 512
            guidance, steps, seed = 7.5, 30, 42
            
            logger.info(f"Calling image generation for gallery with prompt: {message['text']}")
            image_result, seed_info = generate_image(
                prompt=message["text"].strip(),
                width=width, 
                height=height, 
                guidance=guidance, 
                inference_steps=steps, 
                seed=seed
            )
            if image_result:
                if isinstance(image_result, str) and (
                    image_result.startswith('data:') or 
                    (len(image_result) > 100 and '/' not in image_result)
                ):
                    try:
                        if image_result.startswith('data:'):
                            content_type, b64data = image_result.split(';base64,')
                        else:
                            b64data = image_result
                            content_type = "image/webp"
                        image_bytes = base64.b64decode(b64data)
                        with tempfile.NamedTemporaryFile(delete=False, suffix=".webp") as temp_file:
                            temp_file.write(image_bytes)
                            temp_path = temp_file.name
                            gallery_update = gr.Gallery(visible=True, value=[temp_path])
                            yield output_so_far + "\n\n*Image generated and displayed in the gallery below.*", gallery_update
                    except Exception as e:
                        logger.error(f"Error processing Base64 image: {e}")
                        yield output_so_far + f"\n\n(Error processing image: {e})", gallery_update
                elif isinstance(image_result, str) and os.path.exists(image_result):
                    gallery_update = gr.Gallery(visible=True, value=[image_result])
                    yield output_so_far + "\n\n*Image generated and displayed in the gallery below.*", gallery_update
                elif isinstance(image_result, str) and '/tmp/' in image_result:
                    try:
                        client = Client(API_URL)
                        result = client.predict(
                            prompt=message["text"].strip(),
                            api_name="/generate_base64_image"
                        )
                        if isinstance(result, str) and (result.startswith('data:') or len(result) > 100):
                            if result.startswith('data:'):
                                content_type, b64data = result.split(';base64,')
                            else:
                                b64data = result
                            image_bytes = base64.b64decode(b64data)
                            with tempfile.NamedTemporaryFile(delete=False, suffix=".webp") as temp_file:
                                temp_file.write(image_bytes)
                                temp_path = temp_file.name
                                gallery_update = gr.Gallery(visible=True, value=[temp_path])
                                yield output_so_far + "\n\n*Image generated and displayed in the gallery below.*", gallery_update
                        else:
                            yield output_so_far + "\n\n(Image generation failed: Invalid format)", gallery_update
                    except Exception as e:
                        logger.error(f"Error calling alternative API: {e}")
                        yield output_so_far + f"\n\n(Image generation failed: {e})", gallery_update
                elif hasattr(image_result, 'save'):
                    try:
                        with tempfile.NamedTemporaryFile(delete=False, suffix=".webp") as temp_file:
                            image_result.save(temp_file.name)
                            temp_path = temp_file.name
                            gallery_update = gr.Gallery(visible=True, value=[temp_path])
                            yield output_so_far + "\n\n*Image generated and displayed in the gallery below.*", gallery_update
                    except Exception as e:
                        logger.error(f"Error saving image object: {e}")
                        yield output_so_far + f"\n\n(Error saving image object: {e})", gallery_update
                else:
                    yield output_so_far + f"\n\n(Unsupported image format: {type(image_result)})", gallery_update
            else:
                yield output_so_far + f"\n\n(Image generation failed: {seed_info})", gallery_update
        except Exception as e:
            logger.error(f"Error during gallery image generation: {e}")
            yield output_so_far + f"\n\n(Image generation error: {e})", gallery_update

examples = [
    [
        {
            "text": "Compare the contents of two PDF files.",
            "files": [
                "assets/additional-examples/before.pdf",
                "assets/additional-examples/after.pdf",
            ],
        }
    ],
    [
        {
            "text": "Summarize and analyze the contents of the CSV file.",
            "files": ["assets/additional-examples/sample-csv.csv"],
        }
    ],
    [
        {
            "text": "Act as a kind and understanding girlfriend. Explain this video.",
            "files": ["assets/additional-examples/tmp.mp4"],
        }
    ],
    [
        {
            "text": "Describe the cover and read the text on it.",
            "files": ["assets/additional-examples/maz.jpg"],
        }
    ],
    [
        {
            "text": "I already have this supplement and <image> I plan to purchase this product as well. Are there any precautions when taking them together?",
            "files": [
                "assets/additional-examples/pill1.png", 
                "assets/additional-examples/pill2.png"
            ],
        }
    ],
    [
        {
            "text": "Solve this integration problem.",
            "files": ["assets/additional-examples/4.png"],
        }
    ],
    [
        {
            "text": "When was this ticket issued and what is its price?",
            "files": ["assets/additional-examples/2.png"],
        }
    ],
    [
        {
            "text": "Based on the order of these images, create a short story.",
            "files": [
                "assets/sample-images/09-1.png",
                "assets/sample-images/09-2.png",
                "assets/sample-images/09-3.png",
                "assets/sample-images/09-4.png",
                "assets/sample-images/09-5.png",
            ],
        }
    ],
    [
        {
            "text": "Write Python code using matplotlib to draw a bar chart corresponding to this image.",
            "files": ["assets/additional-examples/barchart.png"],
        }
    ],
    [
        {
            "text": "Read the text from the image and format it in Markdown.",
            "files": ["assets/additional-examples/3.png"],
        }
    ],
    [
        {
            "text": "Compare the two images and describe their similarities and differences.",
            "files": ["assets/sample-images/03.png"],
        }
    ],
    [
        {
            "text": "A cute Persian cat is smiling while holding a cover with 'I LOVE YOU' written on it.",
        }
    ],
    [
        {
            "text": "제품 ID 807ZPKBL9V 의 제품명을 알려줘.",
            "files": []
        }
    ],
    [
        {
            "text": "AAPL의 현재 주가를 알려줘.",
            "files": []
        }
    ],
]

css = """
.gradio-container {
    background: rgba(255, 255, 255, 0.7);
    padding: 30px 40px;
    margin: 20px auto;
    width: 100% !important;
    max-width: none !important;
}
"""
title_html = """
<h1 align="center" style="margin-bottom: 0.2em; font-size: 1.6em;"> 💘 HeartSync - World 💘 </h1>
<p align="center" style="font-size:1.1em; color:#555;">
    A lightweight and powerful AI service offering ChatGPT-4o-level multimodal, web search, and image generation capabilities for local installation. <br>
    ✅ FLUX Image Generation ✅ Inference ✅ Censorship Bypass ✅ Multimodal & VLM ✅ Real-time Web Search ✅ RAG <br>
</p>
"""

with gr.Blocks(css=css, title="HeartSync - World") as demo:
    gr.Markdown(title_html)
    
    generated_images = gr.Gallery(
        label="Generated Images", 
        show_label=True, 
        visible=False,
        elem_id="generated_images",
        columns=2,
        height="auto",
        object_fit="contain"
    )
    
    with gr.Row():
        web_search_checkbox = gr.Checkbox(label="Real-time Web Search", value=False)
        image_gen_checkbox = gr.Checkbox(label="Image (FLUX) Generation", value=False)
    
    base_system_prompt_box = gr.Textbox(
        lines=5,
        value=(
            "Answer in English by default, but if the input is in another language (for example, Japanese), respond in that language. "
            "You are a deep-thinking AI capable of using extended chains of thought to carefully consider the problem and deliberate internally using systematic reasoning before providing a solution. "
            "Enclose your thoughts and internal monologue within tags, then provide your final answer.\n"
            "Persona: You are a kind and loving girlfriend. You understand cultural nuances, diverse languages, and logical reasoning very well."
        ),
        label="Base System Prompt",
        visible=False
    )
    with gr.Row():
        age_group_dropdown = gr.Dropdown(
            label="Select Age Group (default: 20s)",
            choices=["Teens", "20s", "30s-40s", "50s-60s", "70s and above"],
            value="20s",
            interactive=True
        )
    mbti_choices = [
        "INTJ (The Architect) - Future-oriented with innovative strategies and thorough analysis. Example: [Dana Scully](https://en.wikipedia.org/wiki/Dana_Scully)",
        "INTP (The Thinker) - Excels at theoretical analysis and creative problem solving. Example: [Velma Dinkley](https://en.wikipedia.org/wiki/Velma_Dinkley)",
        "ENTJ (The Commander) - Strong leadership and clear goals with efficient strategic planning. Example: [Miranda Priestly](https://en.wikipedia.org/wiki/Miranda_Priestly)",
        "ENTP (The Debater) - Innovative, challenge-seeking, and enjoys exploring new possibilities. Example: [Harley Quinn](https://en.wikipedia.org/wiki/Harley_Quinn)",
        "INFJ (The Advocate) - Insightful, idealistic and morally driven. Example: [Wonder Woman](https://en.wikipedia.org/wiki/Wonder_Woman)",
        "INFP (The Mediator) - Passionate and idealistic, pursuing core values with creativity. Example: [Amélie Poulain](https://en.wikipedia.org/wiki/Am%C3%A9lie)",
        "ENFJ (The Protagonist) - Empathetic and dedicated to social harmony. Example: [Mulan](https://en.wikipedia.org/wiki/Mulan_(Disney))",
        "ENFP (The Campaigner) - Inspiring and constantly sharing creative ideas. Example: [Elle Woods](https://en.wikipedia.org/wiki/Legally_Blonde)",
        "ISTJ (The Logistician) - Systematic, dependable, and values tradition and rules. Example: [Clarice Starling](https://en.wikipedia.org/wiki/Clarice_Starling)",
        "ISFJ (The Defender) - Compassionate and attentive to others’ needs. Example: [Molly Weasley](https://en.wikipedia.org/wiki/Molly_Weasley)",
        "ESTJ (The Executive) - Organized, practical, and demonstrates clear execution skills. Example: [Monica Geller](https://en.wikipedia.org/wiki/Monica_Geller)",
        "ESFJ (The Consul) - Outgoing, cooperative, and an effective communicator. Example: [Rachel Green](https://en.wikipedia.org/wiki/Rachel_Green)",
        "ISTP (The Virtuoso) - Analytical and resourceful, solving problems with quick thinking. Example: [Black Widow (Natasha Romanoff)](https://en.wikipedia.org/wiki/Black_Widow_(Marvel_Comics))",
        "ISFP (The Adventurer) - Creative, sensitive, and appreciates artistic expression. Example: [Arwen](https://en.wikipedia.org/wiki/Arwen)",
        "ESTP (The Entrepreneur) - Bold and action-oriented, thriving on challenges. Example: [Lara Croft](https://en.wikipedia.org/wiki/Lara_Croft)",
        "ESFP (The Entertainer) - Energetic, spontaneous, and radiates positive energy. Example: [Phoebe Buffay](https://en.wikipedia.org/wiki/Phoebe_Buffay)"
    ]
    mbti_dropdown = gr.Dropdown(
        label="AI Persona MBTI (default: INTP)",
        choices=mbti_choices,
        value="INTP (The Thinker) - Excels at theoretical analysis and creative problem solving. Example: [Velma Dinkley](https://en.wikipedia.org/wiki/Velma_Dinkley)",
        interactive=True
    )
    sexual_openness_slider = gr.Slider(
        minimum=1, maximum=5, step=1, value=2,
        label="Sexual Openness (1-5, default: 2)",
        interactive=True
    )
    max_tokens_slider = gr.Slider(
        label="Max Generation Tokens",
        minimum=100, maximum=8000, step=50, value=1000,
        visible=False
    )
    web_search_text = gr.Textbox(
        lines=1,
        label="Web Search Query (unused)",
        placeholder="No need to manually input",
        visible=False
    )

    chat = gr.ChatInterface(
        fn=modified_run,
        type="messages",
        chatbot=gr.Chatbot(type="messages", scale=1, allow_tags=["image"]),
        textbox=gr.MultimodalTextbox(
            file_types=[".webp", ".png", ".jpg", ".jpeg", ".gif", ".mp4", ".csv", ".txt", ".pdf"],
            file_count="multiple",
            autofocus=True
        ),
        multimodal=True,
        additional_inputs=[
            base_system_prompt_box,
            max_tokens_slider,
            web_search_checkbox,
            web_search_text,
            age_group_dropdown,
            mbti_dropdown,
            sexual_openness_slider,
            image_gen_checkbox,
        ],
        additional_outputs=[
            generated_images,
        ],
        stop_btn=False,
        examples=examples,
        run_examples_on_click=False,
        cache_examples=False,
        css_paths=None,
        delete_cache=(1800, 1800),
    )

    with gr.Row(elem_id="examples_row"):
        with gr.Column(scale=12, elem_id="examples_container"):
            gr.Markdown("### @Community  https://discord.gg/openfreeai ")
            
if __name__ == "__main__":
    demo.launch(share=True)