""" This module integrates real-time object detection into live YouTube streams using the YOLO model and provides an interactive user interface through Gradio. It allows users to search for live YouTube streams and apply object detection to these streams in real time. Main Features: - Search for live YouTube streams using specific queries. - Retrieve live stream URLs using the `pytube` library. - Perform real-time object detection on live streams using the YOLO model. - Display the live stream and object detection results through a Gradio interface. Dependencies: - cv2 (OpenCV): Used for image processing tasks. - Gradio: Provides the interactive web-based user interface. - `pytube`: Used for retrieving live stream URLs from YouTube. - innertube: Used for interacting with YouTube's internal API. - numpy: Utilized for numerical operations on image data. - PIL (Pillow): A Python Imaging Library for opening, manipulating, and saving images. - ultralytics YOLO: The YOLO model implementation for object detection. Usage: Run this file to launch the Gradio interface, which allows users to input search queries for YouTube live streams, select a stream, and perform object detection on the selected live stream. """ import logging import sys from enum import Enum from typing import Any, Dict, List, Optional, Tuple import cv2 import gradio as gr import innertube import numpy as np from PIL import Image from ultralytics import YOLO from pytube import YouTube logging.basicConfig(stream=sys.stderr, level=logging.DEBUG) class SearchFilter(Enum): LIVE = ("EgJAAQ%3D%3D", "Live") VIDEO = ("EgIQAQ%3D%3D", "Video") def __init__(self, code, human_readable): self.code = code self.human_readable = human_readable def __str__(self): return self.human_readable class SearchService: @staticmethod def search(query: Optional[str], filter: SearchFilter = SearchFilter.VIDEO): response = SearchService._search(query, filter) results = SearchService.parse(response) return results @staticmethod def parse(data: Dict[str, Any]) -> List[Dict[str, str]]: results = [] try: contents = data["contents"]["twoColumnSearchResultsRenderer"]["primaryContents"]["sectionListRenderer"]["contents"] for content in contents: items = content.get("itemSectionRenderer", {}).get("contents", []) for item in items: if "videoRenderer" in item: renderer = item["videoRenderer"] video_id = renderer.get("videoId", "") thumbnails = renderer.get("thumbnail", {}).get("thumbnails", []) thumbnail_url = thumbnails[-1]["url"] if thumbnails else "" title_runs = renderer.get("title", {}).get("runs", []) title = "".join(run.get("text", "") for run in title_runs) results.append( { "video_id": video_id, "thumbnail_url": thumbnail_url, "title": title, } ) except Exception as e: logging.error(f"Error parsing search results: {e}") return results @staticmethod def _search(query: Optional[str] = None, filter: SearchFilter = SearchFilter.VIDEO) -> Dict[str, Any]: client = innertube.InnerTube(client_name="WEB", client_version="2.20230920.00.00") response = client.search(query=query, params=filter.code if filter else None) return response @staticmethod def get_youtube_url(video_id: str) -> str: return f"https://www.youtube.com/watch?v={video_id}" @staticmethod def get_stream(youtube_url: str) -> Optional[str]: """Retrieves the live stream URL for a given YouTube video URL using pytube. :param youtube_url: The URL of the YouTube video. :type youtube_url: str :return: The HLS manifest URL if available, otherwise None. :rtype: Optional[str] """ try: yt = YouTube(youtube_url) _ = yt.streams is_live = yt.player_response.get('videoDetails', {}).get('isLiveContent', False) if is_live: streaming_data = yt.player_response.get('streamingData', {}) hls_manifest_url = streaming_data.get('hlsManifestUrl') if hls_manifest_url: logging.debug(f"Found HLS manifest URL for live stream: {hls_manifest_url}") return hls_manifest_url else: logging.warning(f"HLS manifest URL not found for live stream: {youtube_url}") return None else: logging.warning(f"Video is not a live stream: {youtube_url}") return None except Exception as e: logging.warning(f"An error occurred while getting stream: {e}") return None INITIAL_STREAMS = SearchService.search("world live cams", SearchFilter.LIVE) class LiveYouTubeObjectDetector: def __init__(self): logging.getLogger().setLevel(logging.DEBUG) self.model = YOLO("yolo11n.pt") self.streams = INITIAL_STREAMS # Gradio UI initial_gallery_items = [(stream["thumbnail_url"], stream["title"]) for stream in self.streams] self.gallery = gr.Gallery(label="Live YouTube Videos", value=initial_gallery_items, show_label=True, columns=[4], rows=[5], object_fit="contain", height="auto", allow_preview=False) self.search_input = gr.Textbox(label="Search Live YouTube Videos") self.stream_input = gr.Textbox(label="URL of Live YouTube Video") self.annotated_image = gr.AnnotatedImage(show_label=False) self.search_button = gr.Button("Search", size="lg") self.submit_button = gr.Button("Detect Objects", variant="primary", size="lg") self.page_title = gr.HTML("

Object Detection in Live YouTube Streams

") def detect_objects(self, url: str) -> Tuple[Image.Image, List[Tuple[Tuple[int, int, int, int], str]]]: stream_url = SearchService.get_stream(url) if not stream_url: logging.error(f"Unable to find a stream for: {url}") return self.create_black_image() frame = self.get_frame(stream_url) if frame is None: logging.error(f"Unable to capture frame for: {url}") return self.create_black_image() return self.annotate(frame) def get_frame(self, stream_url: str) -> Optional[np.ndarray]: if not stream_url: return None try: cap = cv2.VideoCapture(stream_url) ret, frame = cap.read() cap.release() if ret and frame is not None: return cv2.resize(frame, (1920, 1080)) else: logging.warning("Unable to process the live stream with cv2.VideoCapture.") return None except Exception as e: logging.warning(f"An error occurred while capturing the frame: {e}") return None def annotate(self, frame: np.ndarray) -> Tuple[Image.Image, List[Tuple[Tuple[int, int, int, int], str]]]: frame_rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB) predictions = self.model.predict(frame_rgb) annotations = [] result = predictions[0] boxes = result.boxes for box in boxes: x1, y1, x2, y2 = box.xyxy[0].tolist() class_id = int(box.cls[0]) class_name = self.model.names[class_id] bbox_coords = (int(x1), int(y1), int(x2), int(y2)) annotations.append((bbox_coords, class_name)) return Image.fromarray(frame_rgb), annotations @staticmethod def create_black_image() -> Tuple[Image.Image, List]: black_image = np.zeros((1080, 1920, 3), dtype=np.uint8) pil_black_image = Image.fromarray(black_image) return pil_black_image, [] @staticmethod def get_live_streams(query=""): return SearchService.search(query if query else "world live cams", SearchFilter.LIVE) def render(self): with gr.Blocks(title="Object Detection in Live YouTube Streams", css="footer {visibility: hidden}", analytics_enabled=False) as app: self.page_title.render() with gr.Column(): with gr.Group(): with gr.Row(): self.stream_input.render() self.submit_button.render() self.annotated_image.render() with gr.Group(): with gr.Row(): self.search_input.render() self.search_button.render() with gr.Row(): self.gallery.render() @self.gallery.select(inputs=None, outputs=[self.annotated_image, self.stream_input], scroll_to_output=True) def detect_objects_from_gallery_item(evt: gr.SelectData): if evt.index is not None and evt.index < len(self.streams): selected_stream = self.streams[evt.index] stream_url = SearchService.get_youtube_url(selected_stream["video_id"]) self.stream_input.value = stream_url annotated_image_result = self.detect_objects(stream_url) return annotated_image_result, stream_url return self.create_black_image(), "" @self.search_button.click(inputs=[self.search_input], outputs=[self.gallery]) def search_live_streams(query): self.streams = self.get_live_streams(query) gallery_items = [(stream["thumbnail_url"], stream["title"]) for stream in self.streams] return gallery_items @self.submit_button.click(inputs=[self.stream_input], outputs=[self.annotated_image]) def detect_objects_from_url(url): return self.detect_objects(url) app.queue().launch(show_api=False, debug=True) if __name__ == "__main__": LiveYouTubeObjectDetector().render()