utils.py

### `utils.py`

from typing import List, Tuple, Dict
import re
import gradio as gr
import io, base64
import numpy as np
from PIL import Image

# History/message helpers
History = List[Tuple[str, str]]
Messages = List[Dict[str, str]]

def history_to_messages(history: History, system: str) -> Messages:
    messages = [{'role': 'system', 'content': system}]
    for h in history:
        user_content = h[0]
        if isinstance(user_content, list):
            text_content = ""
            for item in user_content:
                if isinstance(item, dict) and item.get("type") == "text":
                    text_content += item.get("text", "")
            user_content = text_content if text_content else str(user_content)
        messages.append({'role': 'user', 'content': user_content})
        messages.append({'role': 'assistant', 'content': h[1]})
    return messages

def messages_to_history(messages: Messages) -> Tuple[str, History]:
    assert messages[0]['role'] == 'system'
    history = []
    for q, r in zip(messages[1::2], messages[2::2]):
        user_content = q['content']
        if isinstance(user_content, list):
            text_content = ""
            for item in user_content:
                if isinstance(item, dict) and item.get("type") == "text":
                    text_content += item.get("text", "")
            user_content = text_content if text_content else str(user_content)
        history.append([user_content, r['content']])
    return history

def history_render(history: History):
    return gr.update(visible=True), history

def clear_history():
    return [], [], None, ""

def update_image_input_visibility(model):
    is_ernie_vl = model.get("id") == "baidu/ERNIE-4.5-VL-424B-A47B-Base-PT"
    is_glm_vl = model.get("id") == "THUDM/GLM-4.1V-9B-Thinking"
    return gr.update(visible=is_ernie_vl or is_glm_vl)

def process_image_for_model(image):
    if image is None:
        return None
    if isinstance(image, np.ndarray):
        image = Image.fromarray(image)
    buffer = io.BytesIO()
    image.save(buffer, format='PNG')
    img_str = base64.b64encode(buffer.getvalue()).decode()
    return f"data:image/png;base64,{img_str}"

def create_multimodal_message(text, image=None):
    if image is None:
        return {"role": "user", "content": text}
    content = [
        {"type": "text", "text": text},
        {"type": "image_url", "image_url": {"url": process_image_for_model(image)}}
    ]
    return {"role": "user", "content": content}

# Code-block parsing
def remove_code_block(text):
    patterns = [
        r'```(?:html|HTML)\n([\s\S]+?)\n```',
        r'```\n([\s\S]+?)\n```',
        r'```([\s\S]+?)```'
    ]
    for pattern in patterns:
        match = re.search(pattern, text, re.DOTALL)
        if match:
            extracted = match.group(1).strip()
            lines = extracted.split('\n', 1)
            if lines[0].strip().lower() in ['python', 'html', 'css', 'javascript']:
                return lines[1] if len(lines) > 1 else ''
            return extracted
    return text.strip()

def parse_transformers_js_output(text):
    files = {'index.html': '', 'index.js': '', 'style.css': ''}
    files['index.html'] = re.search(r'```html\s*\n([\s\S]+?)\n```', text, re.IGNORECASE).group(1).strip() if re.search(r'```html\s*\n([\s\S]+?)\n```', text, re.IGNORECASE) else ''
    files['index.js'] = re.search(r'```javascript\s*\n([\s\S]+?)\n```', text, re.IGNORECASE).group(1).strip() if re.search(r'```javascript\s*\n([\s\S]+?)\n```', text, re.IGNORECASE) else ''
    files['style.css'] = re.search(r'```css\s*\n([\s\S]+?)\n```', text, re.IGNORECASE).group(1).strip() if re.search(r'```css\s*\n([\s\S]+?)\n```', text, re.IGNORECASE) else ''
    return files

def format_transformers_js_output(files):
    return f"=== index.html ===\n{files['index.html']}\n=== index.js ===\n{files['index.js']}\n=== style.css ===\n{files['style.css']}"