modules/app.py

from fastapi import FastAPI, HTTPException, Request, Depends
from fastapi.staticfiles import StaticFiles
from pytubefix import YouTube
from pytubefix.exceptions import PytubeFixError
from moviepy.editor import VideoFileClip, concatenate_videoclips, CompositeVideoClip, AudioFileClip, ImageClip
from pydantic import BaseModel, HttpUrl
from typing import List
import requests
import tempfile
import time
import uuid
import aiohttp
import base64
import io
import os
import random
import traceback
import string
import json

app = FastAPI()

whisper_origin = os.getenv("WHISPER_ORIGIN")
whisper_base_url = os.getenv("WHISPER_BASE_URL")
img2location_name = os.getenv("IMG2LOCATION_NAME")
img2location_origin = os.getenv("IMG2LOCATION_ORIGIN")
img2location_base_url = os.getenv("IMG2LOCATION_BASE_URL")
pixart_sigma_base_url = os.getenv("PIXART_SIGMA_BASE_URL")
allowed_user_agent = os.getenv("ALLOWED_USER_AGENT")


def generate_hash(length=12):
    # Characters that can appear in the hash
    characters = string.ascii_lowercase + string.digits
    # Generate a random string of the specified length
    hash_string = ''.join(random.choice(characters) for _ in range(length))
    return hash_string

@app.get("/")
async def read_root():
    return {"message": "Saqib's API"}

# Create a directory to store MP3 files if it doesn't exist
AUDIO_DIR = "audio_files"
os.makedirs(AUDIO_DIR, exist_ok=True)

# Create a directory for storing output files
OUTPUT_DIR = "output"
os.makedirs(OUTPUT_DIR, exist_ok=True)

# Mount the audio directory
app.mount("/audio", StaticFiles(directory=AUDIO_DIR), name="audio")

# Mount the output directory
app.mount("/output", StaticFiles(directory=OUTPUT_DIR), name="output")

class AudioImageInput(BaseModel):
    image_url: HttpUrl
    audio_url: HttpUrl

class VideosInput(BaseModel):
    video_urls: List[HttpUrl]

def download_file(url: str, suffix: str):
    response = requests.get(url)
    if response.status_code != 200:
        raise HTTPException(status_code=400, detail=f"Failed to download file from {url}")
    with tempfile.NamedTemporaryFile(delete=False, suffix=suffix) as temp_file:
        temp_file.write(response.content)
        return temp_file.name

@app.post("/add_audio_to_image")
async def add_audio_to_image(input_data: AudioImageInput):
    # Download image and audio files
    temp_image_path = download_file(str(input_data.image_url), ".png")
    temp_audio_path = download_file(str(input_data.audio_url), ".mp3")

    # Create a video from the image
    image_clip = ImageClip(temp_image_path).set_duration(5)  # 5 seconds duration
    image_clip = image_clip.set_fps(24)  # Set fps to 24

    # Load the audio
    audio_clip = AudioFileClip(temp_audio_path)

    # Set the audio of the video clip
    video_with_audio = image_clip.set_audio(audio_clip)

    # Generate a unique filename
    output_filename = f"{uuid.uuid4()}.mp4"
    output_path = os.path.join(OUTPUT_DIR, output_filename)

    # Write the result to a file
    video_with_audio.write_videofile(output_path, codec="libx264", audio_codec="aac")

    # Clean up temporary files
    os.unlink(temp_image_path)
    os.unlink(temp_audio_path)

    # Return the URL path to the output file
    return f"/output/{output_filename}"

@app.post("/concatenate_videos")
async def concatenate_videos(input_data: VideosInput):
    temp_video_paths = []

    # Download videos to temporary files
    for video_url in input_data.video_urls:
        temp_video_paths.append(download_file(str(video_url), ".mp4"))

    # Load video clips
    video_clips = []
    for path in temp_video_paths:
        clip = VideoFileClip(path)
        if not clip.fps:
            clip = clip.set_fps(24)  # Set fps to 24 if not defined
        video_clips.append(clip)

    # Concatenate video clips
    final_clip = concatenate_videoclips(video_clips)

    # Generate a unique filename
    output_filename = f"{uuid.uuid4()}.mp4"
    output_path = os.path.join(OUTPUT_DIR, output_filename)

    # Write the result to a file
    final_clip.write_videofile(output_path, codec="libx264", audio_codec="aac")

    # Clean up temporary files
    for path in temp_video_paths:
        os.unlink(path)

    # Return the URL path to the output file
    return f"/output/{output_filename}"

@app.get("/get_audio")
async def get_audio(url: str):    
    if not url:
        raise HTTPException(status_code=400, detail="URL is required")

    try:
        print(f"Attempting to process YouTube URL: {url}")
        yt = YouTube(url)
        video = yt.streams.filter(only_audio=True).first()
        
        if not video:
            print(f"No audio stream found for URL: {url}")
            raise HTTPException(status_code=404, detail="No audio stream found for this video")
        
        # Generate a unique filename
        unique_filename = f"{uuid.uuid4().hex}.mp3"
        out_file = os.path.join(AUDIO_DIR, unique_filename)

        print((f"Downloading audio to: {out_file}"))
        
        # Download the audio
        video.download(output_path=AUDIO_DIR, filename=unique_filename)
        
        file_stats = os.stat(out_file)
        print(f'Size of audio file in Bytes: {file_stats.st_size}')
        
        if file_stats.st_size <= 30000000:
            # Construct the URL for the MP3 file
            mp3_url = f"/audio/{unique_filename}"
            return mp3_url
        else:
            os.remove(out_file)
            raise HTTPException(status_code=413, detail="Audio file is too large. Limited to about 1.5 hours.")
    except PytubeFixError as e:
        print(f"PytubeError occurred: {str(e)}")
        raise HTTPException(status_code=400, detail=f"Error processing YouTube video: {str(e)}")
    except HTTPException as he:
        # Re-raise HTTP exceptions
        raise he
    except Exception as e:
        print(f"Unexpected error occurred: {str(e)}")
        print(traceback.format_exc())
        raise HTTPException(status_code=500, detail=f"An unexpected error occurred: {str(e)}")

@app.post("/whisper")
async def whisper(request: Request):
    data = await request.json()  # Extracting JSON data from request
    if "audio_url" not in data:
        raise HTTPException(status_code=400, detail="audio_url not found in request")
    url = data["audio_url"]

    headers = {
        'Accept': 'application/json, text/plain, */*',
        'Accept-Language': 'en-US,en;q=0.9',
        'Cache-Control': 'no-cache',
        'Connection': 'keep-alive',
        'Content-Type': 'application/json',
        'DNT': '1',
        'Origin': whisper_origin,
        'Pragma': 'no-cache',
        'Referer': f'{whisper_origin}/',
        'Sec-Fetch-Dest': 'empty',
        'Sec-Fetch-Mode': 'cors',
        'Sec-Fetch-Site': 'same-site',
        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36 Edg/124.0.0.0',
        'sec-ch-ua': '"Chromium";v="124", "Microsoft Edge";v="124", "Not-A.Brand";v="99"',
        'sec-ch-ua-mobile': '?0',
        'sec-ch-ua-platform': '"Windows"',
    }

    # Async HTTP request to get the audio file
    async with aiohttp.ClientSession() as session:
        async with session.get(url) as resp:
            if resp.status != 200:
                return f"Failed to download audio: {resp.status}"
            audio_data = await resp.read()

            # Encode the audio data to base64
            audio_base64 = base64.b64encode(audio_data).decode("utf-8")

            json_data = '{"audio": "' + audio_base64 + '"}'

            # Post request to the API
            async with session.post(f'{whisper_base_url}/v1/inference/openai/whisper-large', headers=headers, data=json_data) as post_resp:
                if post_resp.status != 200:
                    return f"API request failed: {post_resp.status}"
                return await post_resp.json()

@app.post("/img2location")
async def img2location(request: Request):
    request_json = await request.json()
    image_url = request_json.get("image_url", None)
    
    if not image_url:
        raise HTTPException(status_code=400, detail="image_url not found in request")

    headers = {
        'accept': '*/*',
        'accept-language': 'en-US,en;q=0.9',
        'cache-control': 'no-cache',
        'dnt': '1',
        'origin': img2location_origin,
        'pragma': 'no-cache',
        'priority': 'u=1, i',
        'referer': f'{img2location_origin}/',
        'sec-ch-ua': '"Chromium";v="124", "Microsoft Edge";v="124", "Not-A.Brand";v="99"',
        'sec-ch-ua-mobile': '?0',
        'sec-ch-ua-platform': '"Windows"',
        'sec-fetch-dest': 'empty',
        'sec-fetch-mode': 'cors',
        'sec-fetch-site': 'cross-site',
        'user-agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36 Edg/124.0.0.0',
    }

    async with aiohttp.ClientSession() as session:
        # Fetch the image from the URL
        async with session.get(image_url) as img_response:
            if img_response.status != 200:
                return f"Failed to fetch image: HTTP {img_response.status}"
            image_data = await img_response.read()

        # Using BytesIO to handle the byte content
        data = aiohttp.FormData()
        data.add_field('image', io.BytesIO(image_data), filename="image.png", content_type='image/png')

        # Sending the POST request
        async with session.post(img2location_base_url, headers=headers, data=data) as response:
            if response.status != 200:
                return f"Failed to upload image: HTTP {response.status}"
            json_response = await response.json()
            if json_response["message"]["latitude"] and json_response["message"]["longitude"]:
                json_response["message"]["latitude"] = str(json_response["message"]["latitude"])
                json_response["message"]["longitude"] = str(json_response["message"]["longitude"])
            return json_response

@app.post("/pixart-sigma")
async def pixart_sigma(request: Request):
    request_json = await request.json()
    prompt = request_json.get("prompt", None)
    negative_prompt = request_json.get("negative_prompt", "")
    style = request_json.get("style", "(No style)")
    use_negative_prompt = request_json.get("use_negative_prompt", True)
    num_imgs = request_json.get("num_imgs", 1)
    seed = request_json.get("seed", 0)
    width = request_json.get("width", 1024)
    height = request_json.get("height", 1024)
    schedule = request_json.get("schedule", "DPM-Solver")
    dpms_guidance_scale = request_json.get("dpms_guidance_scale", 4.5)
    sas_guidance_scale = request_json.get("sas_guidance_scale", 3)
    dpms_inference_steps = request_json.get("dpms_inference_steps", 14)
    sas_inference_steps = request_json.get("sas_inference_steps", 25)
    randomize_seed = request_json.get("randomize_seed", True)

    hash = generate_hash()

    headers = {
        'accept': '*/*'
    }

    params = {
        '__theme': 'light',
    }

    json_data = {
        'data': [
            prompt,
            negative_prompt,
            style,
            use_negative_prompt,
            num_imgs,
            seed,
            width,
            height,
            schedule,
            dpms_guidance_scale,
            sas_guidance_scale,
            dpms_inference_steps,
            sas_inference_steps,
            randomize_seed,
        ],
        'event_data': None,
        'fn_index': 3,
        'trigger_id': 7,
        'session_hash': hash,
    }

    async with aiohttp.ClientSession() as session:
        async with session.post(f'{pixart_sigma_base_url}/queue/join', params=params, headers=headers, json=json_data, ssl=False) as response:
            print(response.status)

        params = {
            'session_hash': hash,
        }

        async with session.get(f'{pixart_sigma_base_url}/queue/data', params=params, headers=headers, ssl=False) as response:
            async for line in response.content:
                try:
                    if line:
                        line = line.decode('utf-8')
                        line = line.replace('data: ', '')
                        line_json = json.loads(line)
                        if line_json["msg"] == "process_completed":
                            image_url = line_json["output"]["data"][0][0]["image"]["url"]
                            return {"image_url": image_url}
                except:
                    pass

# if __name__ == "__main__":
#     import uvicorn
#     uvicorn.run(app, host="0.0.0.0", port=8000)