Spaces:
Runtime error
Runtime error
jonathanagustin
commited on
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,2 +1,283 @@
|
|
| 1 |
-
import
|
| 2 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import logging
|
| 2 |
+
import os
|
| 3 |
+
import sys
|
| 4 |
+
import zipfile
|
| 5 |
+
from enum import Enum
|
| 6 |
+
from typing import Any, Dict, List, Optional
|
| 7 |
+
|
| 8 |
+
os.system("python3 -m pip uninstall -y typing-extensions")
|
| 9 |
+
os.system("python3 -m pip install -U typing-extensions")
|
| 10 |
+
os.system(
|
| 11 |
+
"python3 -m pip install -q --progress-bar off streamlink gradio tiktoken ultralytics pillow innertube"
|
| 12 |
+
)
|
| 13 |
+
import cv2
|
| 14 |
+
import gradio as gr
|
| 15 |
+
import innertube
|
| 16 |
+
import numpy as np
|
| 17 |
+
import streamlink
|
| 18 |
+
from PIL import Image, ImageDraw, ImageFont
|
| 19 |
+
from ultralytics import YOLO
|
| 20 |
+
|
| 21 |
+
logging.basicConfig(stream=sys.stderr, level=logging.DEBUG)
|
| 22 |
+
|
| 23 |
+
model = YOLO("yolov8x.pt")
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
class SearchFilter(Enum):
|
| 27 |
+
LIVE = ("EgJAAQ%3D%3D", "Live")
|
| 28 |
+
VIDEO = ("EgIQAQ%3D%3D", "Video")
|
| 29 |
+
|
| 30 |
+
def __init__(self, code, human_readable):
|
| 31 |
+
self.code = code
|
| 32 |
+
self.human_readable = human_readable
|
| 33 |
+
|
| 34 |
+
def __str__(self):
|
| 35 |
+
return self.human_readable
|
| 36 |
+
|
| 37 |
+
|
| 38 |
+
class SearchService:
|
| 39 |
+
@staticmethod
|
| 40 |
+
def search(
|
| 41 |
+
query: Optional[str], filter: SearchFilter = SearchFilter.VIDEO
|
| 42 |
+
) -> (List[Dict[str, Any]], Optional[str]):
|
| 43 |
+
client = innertube.InnerTube("WEB", "2.20230920.00.00")
|
| 44 |
+
response = SearchService._search(query, filter)
|
| 45 |
+
results = SearchService.parse(response)
|
| 46 |
+
return results
|
| 47 |
+
|
| 48 |
+
@staticmethod
|
| 49 |
+
def parse(data: Dict[str, Any]) -> List[Dict[str, Any]]:
|
| 50 |
+
results = []
|
| 51 |
+
items = []
|
| 52 |
+
|
| 53 |
+
contents = (
|
| 54 |
+
data.get("contents", {})
|
| 55 |
+
.get("twoColumnSearchResultsRenderer", {})
|
| 56 |
+
.get("primaryContents", {})
|
| 57 |
+
.get("sectionListRenderer", {})
|
| 58 |
+
.get("contents", [])
|
| 59 |
+
)
|
| 60 |
+
if contents:
|
| 61 |
+
items = contents[0].get("itemSectionRenderer", {}).get("contents", [])
|
| 62 |
+
|
| 63 |
+
for item in items:
|
| 64 |
+
if "videoRenderer" in item:
|
| 65 |
+
renderer = item["videoRenderer"]
|
| 66 |
+
video_id = renderer.get("videoId", "")
|
| 67 |
+
thumbnail_urls = [
|
| 68 |
+
thumb.get("url", "")
|
| 69 |
+
for thumb in renderer.get("thumbnail", {}).get("thumbnails", [])
|
| 70 |
+
]
|
| 71 |
+
title_text = "".join(
|
| 72 |
+
[
|
| 73 |
+
run.get("text", "")
|
| 74 |
+
for run in renderer.get("title", {}).get("runs", [])
|
| 75 |
+
]
|
| 76 |
+
)
|
| 77 |
+
|
| 78 |
+
result = {
|
| 79 |
+
"video_id": video_id,
|
| 80 |
+
"thumbnail_urls": thumbnail_urls,
|
| 81 |
+
"title": title_text,
|
| 82 |
+
}
|
| 83 |
+
results.append(result)
|
| 84 |
+
|
| 85 |
+
return results
|
| 86 |
+
|
| 87 |
+
@staticmethod
|
| 88 |
+
def _search(
|
| 89 |
+
query: Optional[str] = None, filter: SearchFilter = SearchFilter.VIDEO
|
| 90 |
+
) -> Dict[str, Any]:
|
| 91 |
+
client = innertube.InnerTube("WEB", "2.20230920.00.00")
|
| 92 |
+
response = client.search(query=query, params=filter.code if filter else None)
|
| 93 |
+
return response
|
| 94 |
+
|
| 95 |
+
@staticmethod
|
| 96 |
+
def get_youtube_url(video_id: str) -> str:
|
| 97 |
+
return f"https://www.youtube.com/watch?v={video_id}"
|
| 98 |
+
|
| 99 |
+
@staticmethod
|
| 100 |
+
def get_stream_url(youtube_url):
|
| 101 |
+
try:
|
| 102 |
+
session = streamlink.Streamlink()
|
| 103 |
+
streams = session.streams(youtube_url)
|
| 104 |
+
if streams:
|
| 105 |
+
best_stream = streams.get("best")
|
| 106 |
+
return best_stream.url if best_stream else None
|
| 107 |
+
else:
|
| 108 |
+
logging.warning("No streams found for this URL")
|
| 109 |
+
return None
|
| 110 |
+
except Exception as e:
|
| 111 |
+
logging.warning(f"An error occurred: {e}")
|
| 112 |
+
return None
|
| 113 |
+
|
| 114 |
+
|
| 115 |
+
class LiveStreamAnnotator:
|
| 116 |
+
def __init__(self):
|
| 117 |
+
logging.getLogger().setLevel(logging.DEBUG)
|
| 118 |
+
self.model = YOLO("yolov8x.pt")
|
| 119 |
+
self.font_path = self.download_font(
|
| 120 |
+
"https://www.fontsquirrel.com/fonts/download/open-sans",
|
| 121 |
+
"/content/open-sans.zip",
|
| 122 |
+
)
|
| 123 |
+
self.current_page_token = None
|
| 124 |
+
self.streams = self.fetch_live_streams("world live cams")
|
| 125 |
+
# Gradio UI Elements
|
| 126 |
+
initial_gallery_items = [
|
| 127 |
+
(stream["thumbnail_url"], stream["title"]) for stream in self.streams
|
| 128 |
+
]
|
| 129 |
+
self.gallery = gr.Gallery(
|
| 130 |
+
label="Live YouTube Videos",
|
| 131 |
+
value=initial_gallery_items,
|
| 132 |
+
show_label=False,
|
| 133 |
+
columns=[3],
|
| 134 |
+
rows=[10],
|
| 135 |
+
object_fit="contain",
|
| 136 |
+
height="auto",
|
| 137 |
+
)
|
| 138 |
+
self.search_input = gr.Textbox(label="Search Live YouTube Videos")
|
| 139 |
+
self.stream_input = gr.Textbox(label="URL of Live YouTube Video")
|
| 140 |
+
self.output_image = gr.AnnotatedImage(show_label=False)
|
| 141 |
+
self.search_button = gr.Button("Search")
|
| 142 |
+
self.submit_button = gr.Button("Detect Objects", variant="primary", size="lg")
|
| 143 |
+
self.prev_page_button = gr.Button("Previous Page", interactive=False)
|
| 144 |
+
self.next_page_button = gr.Button("Next Page", interactive=False)
|
| 145 |
+
|
| 146 |
+
@staticmethod
|
| 147 |
+
def download_font(url, save_path):
|
| 148 |
+
os.system(f"wget {url} -O {save_path}")
|
| 149 |
+
with zipfile.ZipFile(save_path, "r") as zip_ref:
|
| 150 |
+
font_dir = "/usr/share/fonts/open-sans"
|
| 151 |
+
zip_ref.extractall(font_dir)
|
| 152 |
+
return os.path.join(font_dir, "OpenSans-Regular.ttf")
|
| 153 |
+
|
| 154 |
+
def capture_frame(self, url):
|
| 155 |
+
stream_url = SearchService.get_stream_url(url)
|
| 156 |
+
if not stream_url:
|
| 157 |
+
return self.create_error_image("No stream found"), []
|
| 158 |
+
frame = self.get_frame(stream_url)
|
| 159 |
+
if frame is None:
|
| 160 |
+
return self.create_error_image("Failed to capture frame"), []
|
| 161 |
+
return self.process_frame(frame)
|
| 162 |
+
|
| 163 |
+
def get_frame(self, stream_url):
|
| 164 |
+
if not stream_url:
|
| 165 |
+
return None
|
| 166 |
+
try:
|
| 167 |
+
cap = cv2.VideoCapture(stream_url)
|
| 168 |
+
ret, frame = cap.read()
|
| 169 |
+
cap.release()
|
| 170 |
+
if ret:
|
| 171 |
+
return cv2.resize(frame, (1920, 1080))
|
| 172 |
+
else:
|
| 173 |
+
logging.warning(
|
| 174 |
+
"Unable to process the HLS stream with cv2.VideoCapture."
|
| 175 |
+
)
|
| 176 |
+
return None
|
| 177 |
+
except Exception as e:
|
| 178 |
+
logging.warning(f"An error occurred while capturing the frame: {e}")
|
| 179 |
+
return None
|
| 180 |
+
|
| 181 |
+
def process_frame(self, frame):
|
| 182 |
+
frame_rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
|
| 183 |
+
results = self.model(frame_rgb)
|
| 184 |
+
annotations = self.get_annotations(results)
|
| 185 |
+
return Image.fromarray(frame_rgb), annotations
|
| 186 |
+
|
| 187 |
+
def get_annotations(self, results):
|
| 188 |
+
annotations = []
|
| 189 |
+
for result in results:
|
| 190 |
+
for box in result.boxes:
|
| 191 |
+
class_id = int(box.cls[0])
|
| 192 |
+
class_name = result.names[class_id]
|
| 193 |
+
bbox = tuple(map(int, box.xyxy[0]))
|
| 194 |
+
annotations.append((bbox, class_name))
|
| 195 |
+
return annotations
|
| 196 |
+
|
| 197 |
+
@staticmethod
|
| 198 |
+
def create_error_image(message):
|
| 199 |
+
error_image = np.zeros((1920, 1080, 3), dtype=np.uint8)
|
| 200 |
+
pil_image = Image.fromarray(error_image)
|
| 201 |
+
draw = ImageDraw.Draw(pil_image)
|
| 202 |
+
font = ImageFont.truetype("/usr/share/fonts/open-sans/OpenSans-Regular.ttf", 24)
|
| 203 |
+
text_size = draw.textsize(message, font=font)
|
| 204 |
+
position = ((1920 - text_size[0]) // 2, (1080 - text_size[1]) // 2)
|
| 205 |
+
draw.text(position, message, (0, 0, 255), font=font)
|
| 206 |
+
return cv2.cvtColor(np.array(pil_image), cv2.COLOR_RGB2BGR)
|
| 207 |
+
|
| 208 |
+
def fetch_live_streams(self, query=""):
|
| 209 |
+
streams = []
|
| 210 |
+
results = SearchService.search(
|
| 211 |
+
query if query else "world live cams", SearchFilter.LIVE
|
| 212 |
+
)
|
| 213 |
+
for result in results:
|
| 214 |
+
if "video_id" in result and "thumbnail_urls" in result:
|
| 215 |
+
streams.append(
|
| 216 |
+
{
|
| 217 |
+
"thumbnail_url": result["thumbnail_urls"][0]
|
| 218 |
+
if result["thumbnail_urls"]
|
| 219 |
+
else "",
|
| 220 |
+
"title": result["title"],
|
| 221 |
+
"video_id": result["video_id"],
|
| 222 |
+
"label": result["video_id"],
|
| 223 |
+
}
|
| 224 |
+
)
|
| 225 |
+
return streams
|
| 226 |
+
|
| 227 |
+
def render(self):
|
| 228 |
+
with gr.Blocks(
|
| 229 |
+
title="Object Detection in Live YouTube Streams",
|
| 230 |
+
css="footer {visibility: hidden}",
|
| 231 |
+
) as app:
|
| 232 |
+
gr.HTML(
|
| 233 |
+
"<center><h1><b>Object Detection in Live YouTube Streams</b></h1></center>"
|
| 234 |
+
)
|
| 235 |
+
with gr.Column():
|
| 236 |
+
self.stream_input.render()
|
| 237 |
+
with gr.Group():
|
| 238 |
+
self.output_image.render()
|
| 239 |
+
self.submit_button.render()
|
| 240 |
+
with gr.Group():
|
| 241 |
+
with gr.Row():
|
| 242 |
+
self.search_input.render()
|
| 243 |
+
self.search_button.render()
|
| 244 |
+
with gr.Row():
|
| 245 |
+
self.gallery.render()
|
| 246 |
+
|
| 247 |
+
@self.gallery.select(
|
| 248 |
+
inputs=None, outputs=[self.output_image, self.stream_input]
|
| 249 |
+
)
|
| 250 |
+
def on_gallery_select(evt: gr.SelectData):
|
| 251 |
+
selected_index = evt.index
|
| 252 |
+
if selected_index is not None and selected_index < len(self.streams):
|
| 253 |
+
selected_stream = self.streams[selected_index]
|
| 254 |
+
stream_url = SearchService.get_youtube_url(
|
| 255 |
+
selected_stream["video_id"]
|
| 256 |
+
)
|
| 257 |
+
frame_output = self.capture_frame(stream_url)
|
| 258 |
+
return frame_output, stream_url
|
| 259 |
+
return None, ""
|
| 260 |
+
|
| 261 |
+
@self.search_button.click(
|
| 262 |
+
inputs=[self.search_input], outputs=[self.gallery]
|
| 263 |
+
)
|
| 264 |
+
def on_search_click(query):
|
| 265 |
+
self.streams = self.fetch_live_streams(query)
|
| 266 |
+
gallery_items = [
|
| 267 |
+
(stream["thumbnail_url"], stream["title"])
|
| 268 |
+
for stream in self.streams
|
| 269 |
+
]
|
| 270 |
+
return gallery_items
|
| 271 |
+
|
| 272 |
+
@self.submit_button.click(
|
| 273 |
+
inputs=[self.stream_input], outputs=[self.output_image]
|
| 274 |
+
)
|
| 275 |
+
def annotate_stream(url):
|
| 276 |
+
return self.capture_frame(url)
|
| 277 |
+
|
| 278 |
+
app.queue().launch(
|
| 279 |
+
share=True, debug=True, quiet=True, show_api=False, height=800
|
| 280 |
+
)
|
| 281 |
+
|
| 282 |
+
|
| 283 |
+
LiveStreamAnnotator().render()
|