app.py

from huggingface_hub import InferenceClient
import gradio as gr
import os
import re
import requests
import http.client
import typing
import urllib.request
import vertexai
from vertexai.generative_models import GenerativeModel, Image

with open(".config/application_default_credentials.json", 'w') as file:
    file.write(str(os.getenv('credentials')))

vertexai.init(project=os.getenv('project_id'))
model = GenerativeModel("gemini-1.0-pro-vision")
client = InferenceClient("google/gemma-7b-it")

def extract_image_urls(text):
    url_regex = r"(https?:\/\/.*\.(?:png|jpg|jpeg|gif|webp|svg))"
    image_urls = re.findall(url_regex, text, flags=re.IGNORECASE)
    valid_image_url = ""
    for url in image_urls:
        try:
            response = requests.head(url)  # Use HEAD request for efficiency
            if response.status_code in range(200, 300) and 'image' in response.headers.get('content-type', ''):
                valid_image_url = url
        except requests.exceptions.RequestException:
            pass  # Ignore inaccessible URLs
    return valid_image_url

def load_image_from_url(image_url: str) -> Image:
    with urllib.request.urlopen(image_url) as response:
        response = typing.cast(http.client.HTTPResponse, response)
        image_bytes = response.read()
    return Image.from_bytes(image_bytes)

def search(url):
    image = load_image_from_url(url)
    response = model.generate_content([image,"Describe what is shown in this image."])
    return response.text

def format_prompt(message, history):
  prompt = ""
  for user_prompt, bot_response in history:
    prompt += f"<start_of_turn>user\n{user_prompt}<end_of_turn>\n"
    prompt += f"<start_of_turn>model\n{bot_response}<end_of_turn>\n"
  prompt += f"<start_of_turn>user\n{message}<end_of_turn>\n<start_of_turn>model\n"
  return prompt

def generate(
    prompt, history, system_prompt, temperature=0.9, max_new_tokens=256, top_p=0.95, repetition_penalty=1.0,
):
    temperature = float(temperature)
    if temperature < 1e-2:
        temperature = 1e-2
    top_p = float(top_p)

    generate_kwargs = dict(
        temperature=temperature,
        max_new_tokens=max_new_tokens,
        top_p=top_p,
        repetition_penalty=repetition_penalty,
        do_sample=True,
        seed=42,
    )

    image = extract_image_urls(prompt)
    if image:
        image_description = "Image Description: " + search(image)
        prompt = prompt.replace(image, image_description)
        print(prompt)
    formatted_prompt = format_prompt(f"{system_prompt}, {prompt}", history)
    stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
    output = ""

    for response in stream:
        output += response.token.text
        yield output
    return output


additional_inputs=[
    gr.Textbox(
        label="System Prompt",
        max_lines=1,
        interactive=True,
    ),
    gr.Slider(
        label="Temperature",
        value=0.9,
        minimum=0.0,
        maximum=1.0,
        step=0.05,
        interactive=True,
        info="Higher values produce more diverse outputs",
    ),
    gr.Slider(
        label="Max new tokens",
        value=256,
        minimum=0,
        maximum=1048,
        step=64,
        interactive=True,
        info="The maximum numbers of new tokens",
    ),
    gr.Slider(
        label="Top-p (nucleus sampling)",
        value=0.90,
        minimum=0.0,
        maximum=1,
        step=0.05,
        interactive=True,
        info="Higher values sample more low-probability tokens",
    ),
    gr.Slider(
        label="Repetition penalty",
        value=1.2,
        minimum=1.0,
        maximum=2.0,
        step=0.05,
        interactive=True,
        info="Penalize repeated tokens",
    )
]

examples=[["What are they doing here https://upload.wikimedia.org/wikipedia/commons/3/38/Two_dancers.jpg ?", None, None, None, None, None]]

gr.ChatInterface(
    fn=generate,
    chatbot=gr.Chatbot(show_label=True, show_share_button=True, show_copy_button=True, likeable=True, layout="bubble", bubble_full_width=False),
    additional_inputs=additional_inputs,
    title="Hey Gemini",
    description="Gemini Sprint submission by Rishiraj Acharya. Uses Google's Gemini 1.0 Pro Vision multimodal model from Vertex AI with Google's Gemma 7B Instruct model from Hugging Face.",
    theme="Soft",
    examples=examples,
    concurrency_limit=20,
).launch(show_api=False)