app.py

import gradio as gr
import requests
import time
import json
from contextlib import closing
from websocket import create_connection
from deep_translator import GoogleTranslator
from langdetect import detect
import os
from PIL import Image
import io
import base64
import os
import random
import tempfile


def animate_img(encoded_string):
    r = requests.post("https://stable-video-diffusion.com/api/upload", files={"file": open(encoded_string, 'rb')})
    print(r.text)
    hash_ = r.json()['hash']
    time.sleep(10)
    while True:
        r2 = requests.get(f"https://stable-video-diffusion.com/result?hash={hash_}")
        source_string = r2.text
        if "Generation has been in progress for" in source_string:
            time.sleep(10)
            continue
        if "Generation has been in progress for" not in source_string:
            pattern = r'https://storage.stable-video-diffusion.com/([a-f0-9]{32})\.mp4'
            matches = re.findall(pattern, source_string)
            sd_video = []
            for match in matches:
                sd_video.append(f"https://storage.stable-video-diffusion.com/{match}.mp4")
            print(sd_video[0])
            return sd_video[0]


def flip_text(prompt, motion):
    try:
        language = detect(prompt)
        if language == 'ru':
            prompt = GoogleTranslator(source='ru', target='en').translate(prompt)
            print(prompt)
    except:
        prompt = 'video'

    url_video_g = os.getenv("url_video_g")
    url_video_c = os.getenv("url_video_c")

    if motion == "Приближение →←":
        motion = 'zoom in'
    if motion == "Отдаление ←→":
        motion = 'zoom out'
    if motion == "Вверх ↑":
        motion = 'up'
    if motion == "Вниз ↓":
        motion = 'down'
    if motion == "Влево ←":
        motion = 'left'
    if motion == "Вправо →":
        motion = 'right'
    if motion == "По часовой стрелке ⟳":
        motion = 'rotate cw'
    if motion == "Против часовой стрелки ⟲":
        motion = 'rotate ccw'
    
    data = {"prompt": f"{prompt}","image": "null", "denoise": 0.75,"motion": motion}
    r = requests.post(f"{url_video_g}", json=data)
    while True:
        data2 = {"task_id": f"{r.json()['task_id']}"}
        r2 = requests.post(f"{url_video_c}", json=data2)
        time.sleep(3)
        try:
            if r2.json()['status'] == "QUEUED":
                continue
            if r2.json()['status'] == "PROCESSING":
                continue
        except:
            try:
                n_im2 = f"{time.time()}"
                with tempfile.NamedTemporaryFile(prefix=f'aaafff{n_im2}', suffix='.mp4', delete=False) as file:
                    for chunk in r2.iter_content(chunk_size=1024):
                        if chunk:
                            file.write(chunk)
                    return file.name
            except Exception as e:
                print(e)
                break




def flip_text2(encoded_string, prompt, motion):

    url_video_g = os.getenv("url_video_g")
    url_video_c = os.getenv("url_video_c")
    
    try:
        language = detect(prompt)
        if language == 'ru':
            prompt = GoogleTranslator(source='ru', target='en').translate(prompt)
            print(prompt)
    except:
        pass

    if motion == "Приближение →←":
        motion = 'zoom in'
    if motion == "Отдаление ←→":
        motion = 'zoom out'
    if motion == "Вверх ↑":
        motion = 'up'
    if motion == "Вниз ↓":
        motion = 'down'
    if motion == "Влево ←":
        motion = 'left'
    if motion == "Вправо →":
        motion = 'right'
    if motion == "По часовой стрелке ⟳":
        motion = 'rotate cw'
    if motion == "Против часовой стрелки ⟲":
        motion = 'rotate ccw'
 
    with open(encoded_string, "rb") as image_file:
        encoded_string2 = base64.b64encode(image_file.read())
        encoded_string2 = str(encoded_string2).replace("b'", '')

    data = {"prompt": f"{prompt}","image": f"{encoded_string2}","denoise":0.75,"motion": motion}
    r = requests.post(f"{url_video_g}", json=data)
    while True:
        data2 = {"task_id": f"{r.json()['task_id']}"}
        r2 = requests.post(f"{url_video_c}", json=data2)
        time.sleep(3)
        try:
            if r2.json()['status'] == "QUEUED":
                continue
            if r2.json()['status'] == "PROCESSING":
                continue
        except:
            try:
                n_im2 = f"{time.time()}"
                with tempfile.NamedTemporaryFile(prefix=f'aaafff{n_im2}', suffix='.mp4', delete=False) as file:
                    for chunk in r2.iter_content(chunk_size=1024):
                        if chunk:
                            file.write(chunk)
                    return file.name
            except Exception as e:
                print(e)
                break

                


css = """
#generate {
    width: 100%;
    background: #e253dd !important;
    border: none;
    border-radius: 50px;
    outline: none !important;
    color: white;
}
#generate:hover {
    background: #de6bda !important;
    outline: none !important;
    color: #fff;
    }
footer {visibility: hidden !important;}
"""

with gr.Blocks(css=css) as demo:

    with gr.Tab("Сгенерировать видео"):
        with gr.Column():
            prompt = gr.Textbox(placeholder="Введите описание видео...", show_label=True, label='Описание:', lines=3)
            motion1 = gr.Dropdown(value="Приближение →←", interactive=True, show_label=True, label="Движение камеры:", choices=[
                    "Приближение →←", "Отдаление ←→", "Вверх ↑", "Вниз ↓", "Влево ←", "Вправо →", "По часовой стрелке ⟳", "Против часовой стрелки ⟲"])
        with gr.Column():
            text_button = gr.Button("Сгенерировать видео", variant='primary', elem_id="generate")
        with gr.Column():
            video_output = gr.Video(show_label=True, label='Результат:', type="file")
            text_button.click(flip_text, inputs=[prompt, motion1], outputs=video_output)
            
    with gr.Tab("Анимировать изображение"):
        with gr.Column():
            prompt2 = gr.Image(show_label=True, interactive=True, type='filepath', label='Исходное изображение:')
            prompt12 = gr.Textbox(placeholder="Введите описание видео...", show_label=True, label='Описание видео (опционально):', lines=3)
            motion2 = gr.Dropdown(value="Приближение →←", interactive=True, show_label=True, label="Движение камеры:", choices=[
                    "Приближение →←", "Отдаление ←→", "Вверх ↑", "Вниз ↓", "Влево ←", "Вправо →", "По часовой стрелке ⟳", "Против часовой стрелки ⟲"])
        with gr.Column():
            text_button2 = gr.Button("Анимировать изображение", variant='primary', elem_id="generate")
        with gr.Column():
            video_output2 = gr.Video(show_label=True, label='Результат:', type="file")
            text_button2.click(animate_img, inputs=[prompt2], outputs=video_output2)
    
demo.queue(concurrency_count=12)
demo.launch()