app.py

import gradio as gr
from transformers import AutoModelForCausalLM, AutoTokenizer
from PIL import Image
import re
import copy
import secrets
from pathlib import Path
import os
os.system("pip install git+https://github.com/openai/whisper.git")
import whisper


model_whisper = whisper.load_model("small")

# Constants
BOX_TAG_PATTERN = r"<box>([\s\S]*?)</box>"
PUNCTUATION = "!\"#$%&'()*+,-./:;<=>?@[\\]^_`{|}~"

# Initialize model and tokenizer
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen-VL-Chat-Int4", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen-VL-Chat-Int4", device_map="auto", trust_remote_code=True).eval()

def format_text(text):
    """Format text for rendering in the chat UI."""
    lines = text.split("\n")
    lines = [line for line in lines if line != ""]
    count = 0
    for i, line in enumerate(lines):
        if "```" in line:
            count += 1
            items = line.split("`")
            if count % 2 == 1:
                lines[i] = f'<pre><code class="language-{items[-1]}">'
            else:
                lines[i] = f"<br></code></pre>"
        else:
            if i > 0:
                if count % 2 == 1:
                    line = line.replace("`", r"\`")
                    line = line.replace("<", "&lt;")
                    line = line.replace(">", "&gt;")
                    line = line.replace(" ", "&nbsp;")
                    line = line.replace("*", "&ast;")
                    line = line.replace("_", "&lowbar;")
                    line = line.replace("-", "&#45;")
                    line = line.replace(".", "&#46;")
                    line = line.replace("!", "&#33;")
                    line = line.replace("(", "&#40;")
                    line = line.replace(")", "&#41;")
                    line = line.replace("$", "&#36;")
                lines[i] = "<br>" + line
    text = "".join(lines)
    return text

def transcribe_audio(audio):
    audio = whisper.load_audio(audio)
    audio = whisper.pad_or_trim(audio)
    mel = whisper.log_mel_spectrogram(audio).to(model_whisper.device)
    _, probs = model_whisper.detect_language(mel)
    options = whisper.DecodingOptions(fp16 = False)
    result = whisper.decode(model_whisper, mel, options)
    return result.text


def get_chat_response(chatbot, task_history):
    global model, tokenizer 
    chat_query = chatbot[-1][0]
    query = task_history[-1][0]
    history_cp = copy.deepcopy(task_history)
    full_response = ""

    history_filter = []
    pic_idx = 1
    pre = ""
    for i, (q, a) in enumerate(history_cp):
        if isinstance(q, (tuple, list)):
            q = f'Picture {pic_idx}: <img>{q[0]}</img>'
            pre += q + '\n'
            pic_idx += 1
        else:
            pre += q
            history_filter.append((pre, a))
            pre = ""
    history, message = history_filter[:-1], history_filter[-1][0]
    response, history = model.chat(tokenizer, message, history=history)
    image = tokenizer.draw_bbox_on_latest_picture(response, history)
    if image is not None:
        temp_dir = secrets.token_hex(20)
        temp_dir = Path("/tmp") / temp_dir
        temp_dir.mkdir(exist_ok=True, parents=True)
        name = f"tmp{secrets.token_hex(5)}.jpg"
        filename = temp_dir / name
        image.save(str(filename))
        chatbot[-1] = (format_text(chat_query), (str(filename),))  # Hier verwenden wir format_text statt _parse_text
        chat_response = response.replace("<ref>", "")
        chat_response = chat_response.replace(r"</ref>", "")
        chat_response = re.sub(BOX_TAG_PATTERN, "", chat_response)
        if chat_response != "":
            chatbot.append((None, chat_response))
    else:
        chatbot[-1] = (format_text(chat_query), response)
    full_response = format_text(response)
    task_history[-1] = (query, full_response)
    return chatbot


def handle_text_input(history, task_history, text):
    """Handle text input from the user."""
    task_text = text
    if len(text) >= 2 and text[-1] in PUNCTUATION and text[-2] not in PUNCTUATION:
        task_text = text[:-1]
    history = history + [(format_text(text), None)]
    task_history = task_history + [(task_text, None)]
    return history, task_history, ""

def handle_file_upload(history, task_history, file):
    """Handle file upload from the user."""
    history = history + [((file.name,), None)]
    task_history = task_history + [((file.name,), None)]
    return history, task_history

def clear_input():
    """Clear the user input."""
    return gr.update(value="")

def clear_history(task_history):
    """Clear the chat history."""
    task_history.clear()
    return []

def handle_regeneration(chatbot, task_history):
    """Handle the regeneration of the last response."""
    print("Regenerate clicked")
    print("Before:", task_history, chatbot)
    if not task_history:
        return chatbot
    item = task_history[-1]
    if item[1] is None:
        return chatbot
    task_history[-1] = (item[0], None)
    chatbot_item = chatbot.pop(-1)
    if chatbot_item[0] is None:
        chatbot[-1] = (chatbot[-1][0], None)
    else:
        chatbot.append((chatbot_item[0], None))
    print("After:", task_history, chatbot)
    return get_chat_response(chatbot, task_history)


with gr.Blocks(theme='gradio/soft') as demo:
    audio = gr.Audio(
        label="Input Audio",
        show_label=False,
        source="microphone",
        type="filepath"
    )
    gr.Markdown("# Qwen-VL Multimodal-Vision-Insight")
    gr.Markdown(
        "## Developed by Keyvan Hardani (Keyvven on [Twitter](https://twitter.com/Keyvven))\n"
        "Special thanks to [@Artificialguybr](https://twitter.com/artificialguybr) for the inspiration from his code.\n"
        "### Qwen-VL: A Multimodal Large Vision Language Model by Alibaba Cloud\n"
    )
    chatbot = gr.Chatbot(label='Qwen-VL-Chat', elem_classes="control-height", height=520)
    query = gr.Textbox(lines=2, label='Input')
    task_history = gr.State([])

    with gr.Row():
        with gr.Column(width=4):
            upload_btn = gr.UploadButton("📁 Upload", file_types=["image"], elem_classes="control-width")
        with gr.Column(width=2):
            submit_btn = gr.Button("🚀 Submit", elem_classes="control-width")
        with gr.Column(width=2):
            regen_btn = gr.Button("🤔️ Regenerate", elem_classes="control-width")
        with gr.Column(width=2):
            clear_btn = gr.Button("🧹 Clear History", elem_classes="control-width")
    
    gr.Markdown("### Key Features:\n- **Strong Performance**: Surpasses existing LVLMs on multiple English benchmarks including Zero-shot Captioning and VQA.\n- **Multi-lingual Support**: Supports English, Chinese, and multi-lingual conversation.\n- **High Resolution**: Utilizes 448*448 resolution for fine-grained recognition and understanding.")
    submit_btn.click(handle_text_input, [chatbot, task_history, query], [chatbot, task_history]).then(
        get_chat_response, [chatbot, task_history], [chatbot], show_progress=True
    )

    submit_btn.click(clear_input, [], [query])
    clear_btn.click(clear_history, [task_history], [chatbot], show_progress=True)
    regen_btn.click(handle_regeneration, [chatbot, task_history], [chatbot], show_progress=True)
    upload_btn.upload(handle_file_upload, [chatbot, task_history, upload_btn], [chatbot, task_history], show_progress=True)
    audio.on_change(transcribe_audio, inputs=[audio], outputs=[query])


demo.launch()