mariano/media-analyzer
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

integrate webcam support with real-time detection

Browse Source
This commit is contained in:
buenosairesam
2025-08-25 04:01:24 -03:00
parent c7822c1ec4
commit ad2b1a8a8f
38 changed files with 1252 additions and 458 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
.gitignore vendored
View File

@@ -7,3 +7,4 @@ postgres_data
media/*
!media/.gitkeep
media-analyzer-src.tar.gz
media-analyzer/media

59
GEMINI.md Normal file
View File

@@ -0,0 +1,59 @@
# Media Analyzer
## Project Overview
This project is a real-time video streaming and AI analysis platform. It ingests RTMP video streams, processes them with computer vision AI models, and provides live analysis results through a responsive web dashboard.
The architecture is based on microservices and includes:
* **Backend**: A Django application that handles video stream management, AI processing, and WebSocket communication for real-time updates.
* **Frontend**: An Angular single-page application that provides a user interface for stream viewing and analysis visualization.
* **AI Processing**: A Python-based analysis engine that uses various adapters for different types of analysis, such as object detection, logo detection, and motion analysis. The engine can be configured to run locally, on a remote LAN worker, or in the cloud.
* **Streaming**: An NGINX server with the RTMP module ingests video streams and converts them to HLS for web playback.
* **Infrastructure**: The entire platform is containerized using Docker and can be deployed with Docker Compose for development or Kubernetes for production.
## Building and Running
### Docker Compose (Development)
To run the application in a development environment, use Docker Compose:
```bash
# Start all services
docker compose up
# Run migrations (in a separate terminal)
docker compose --profile tools up migrate
```
The application will be accessible at the following URLs:
* **Frontend**: `http://localhost:4200`
* **Backend API**: `http://localhost:8000`
* **RTMP Stream**: `rtmp://localhost:1935/live`
* **HLS Stream**: `http://localhost:8081/hls`
### Kubernetes (Production-like)
To deploy the application to a Kubernetes cluster, you can use the provided scripts:
```bash
# Build and push images to a local registry
./k8s/build-for-ctlptl.sh
# Deploy to Kubernetes
kubectl apply -k k8s/overlays/development
# Check deployment status
kubectl get pods -n media-analyzer
# Access the application via port forwarding
kubectl port-forward service/frontend -n media-analyzer 4200:80
```
## Development Conventions
* **Backend**: The backend is a Django application. Follow Django best practices for development.
* **Frontend**: The frontend is an Angular application. Follow Angular best practices for development.
* **AI Processing**: The AI processing engine is designed to be extensible. To add a new analysis capability, create a new adapter and integrate it with the `AnalysisEngine`.
* **Testing**: The project includes a `test_unified_ai.py` file, which suggests that there is a testing framework in place. Run existing tests and add new ones when making changes.

88
def/improvements/angular_compose Normal file
View File

@@ -0,0 +1,88 @@
You have a few options to speed up frontend development while keeping the backend services
running:
Option 1: Hybrid Development (Recommended)
Run frontend locally, backend in Docker:
# Terminal 1: Start backend services only
docker compose up postgres redis backend celery-logo celery-default file-watcher nginx-rtmp
# Terminal 2: Run Angular dev server locally
cd frontend && ng serve --host 0.0.0.0 --port 4200
Pros: Hot reload, fast rebuilds, best developer experience
Cons: Need to manage CORS and proxy config
Option 2: Docker Bind Mount + Manual Rebuild
Keep frontend container but rebuild manually:
# Build frontend assets locally
cd frontend && ng build --watch
# Or for production build
cd frontend && ng build --configuration production
# Assets go to dist/ and can be volume mounted to nginx
Option 3: Frontend Container with Volume Mount
Modify docker-compose.yml for dev:
frontend:
build:
context: ./frontend
target: development
volumes:
- ./frontend/src:/app/src # Mount source
- ./frontend/dist:/usr/share/nginx/html # Mount build output
command: sh -c "ng build --watch & nginx -g 'daemon off;'"
Option 4: Live Reload in Container
Add this to your frontend Dockerfile development stage:
FROM node:18 AS development
WORKDIR /app
COPY package*.json ./
RUN npm ci
COPY . .
# Start both ng build --watch and nginx
CMD ["sh", "-c", "ng build --watch --output-path /usr/share/nginx/html & nginx -g 'daemon off;'"]
My Recommendation: Option 1
For your use case, I'd run:
# Keep backend running
docker compose up -d postgres redis backend celery-logo celery-default file-watcher nginx-rtmp
# Run frontend locally with proxy
cd frontend && ng serve --proxy-config proxy.conf.json
Create frontend/proxy.conf.json:
{
"/api/*": {
"target": "http://localhost:8000",
"secure": false,
"logLevel": "debug"
},
"/streaming/*": {
"target": "http://localhost:8000",
"secure": false
},
"/ws/*": {
"target": "http://localhost:8000",
"secure": false,
"ws": true
}
}
This gives you:
- ✅ Instant hot reload
- ✅ Fast TypeScript compilation
- ✅ Backend services running
- ✅ No container rebuilds
Would you like me to set up the proxy config for the hybrid approach?

17
media-analyzer/backend/ai_processing/analysis_engine.py
View File

@@ -15,6 +15,8 @@ logger = logging.getLogger(__name__)
class AnalysisEngine:
"""Main analysis engine that orchestrates capability-specific adapters with execution strategies"""
_strategy_logged = False
def __init__(self):
self.object_detector = None
self.logo_detector = None
@@ -66,7 +68,9 @@ class AnalysisEngine:
logger.warning(f"Unknown strategy type {strategy_type}, falling back to local")
self.execution_strategy = strategy_configs['local']()
logger.info(f"Configured execution strategy: {strategy_type}")
if not AnalysisEngine._strategy_logged:
logger.info(f"Configured execution strategy: {strategy_type}")
AnalysisEngine._strategy_logged = True
except Exception as e:
logger.error(f"Failed to configure execution strategy: {e}")
@@ -76,8 +80,17 @@ class AnalysisEngine:
def extract_frame_from_segment(self, segment_path, timestamp=None):
"""Extract frame from video segment"""
try:
import os
logger.debug(f"Attempting to extract frame from: {segment_path}")
if not os.path.exists(segment_path):
logger.error(f"Segment file does not exist: {segment_path}")
return None
cap = cv2.VideoCapture(segment_path)
if not cap.isOpened():
logger.error(f"OpenCV failed to open: {segment_path}")
return None
# For TS segments, seeking is problematic, just read first frame
@@ -88,6 +101,8 @@ class AnalysisEngine:
if ret:
frame_rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
return Image.fromarray(frame_rgb)
else:
logger.error(f"Failed to read frame from {segment_path}")
return None
except Exception as e:

23
media-analyzer/backend/ai_processing/migrations/0003_rename_stream_id_to_stream_key.py Normal file
View File

@@ -0,0 +1,23 @@
# Generated migration for stream_id to stream_key rename
from django.db import migrations
class Migration(migrations.Migration):
dependencies = [
('ai_processing', '0002_make_provider_nullable'),
]
operations = [
migrations.RenameField(
model_name='videoanalysis',
old_name='stream_id',
new_name='stream_key',
),
migrations.RenameField(
model_name='processingqueue',
old_name='stream_id',
new_name='stream_key',
),
]

35
media-analyzer/backend/ai_processing/migrations/0004_remove_processingqueue_ai_processi_stream__60b035_idx_and_more.py Normal file
View File

@@ -0,0 +1,35 @@
# Generated by Django 5.0.6 on 2025-08-22 00:19
import ai_processing.models
from django.db import migrations, models
class Migration(migrations.Migration):
dependencies = [
('ai_processing', '0003_rename_stream_id_to_stream_key'),
]
operations = [
migrations.RemoveIndex(
model_name='processingqueue',
name='ai_processi_stream__60b035_idx',
),
migrations.RemoveIndex(
model_name='videoanalysis',
name='ai_processi_stream__b961da_idx',
),
migrations.AlterField(
model_name='videoanalysis',
name='confidence_threshold',
field=models.FloatField(default=ai_processing.models.get_default_confidence_threshold),
),
migrations.AddIndex(
model_name='processingqueue',
index=models.Index(fields=['stream_key'], name='ai_processi_stream__aecb17_idx'),
),
migrations.AddIndex(
model_name='videoanalysis',
index=models.Index(fields=['stream_key', 'timestamp'], name='ai_processi_stream__d99710_idx'),
),
]

21
media-analyzer/backend/ai_processing/models.py
View File

@@ -2,6 +2,15 @@ from django.db import models
import uuid
def get_default_confidence_threshold():
"""Get default confidence threshold from settings"""
try:
from django.conf import settings
return settings.LOGO_DETECTION_CONFIG['confidence_threshold']
except:
return 0.6 # Fallback if settings not available
class AnalysisProvider(models.Model):
id = models.UUIDField(primary_key=True, default=uuid.uuid4, editable=False)
name = models.CharField(max_length=100, unique=True)
@@ -28,20 +37,20 @@ class Brand(models.Model):
class VideoAnalysis(models.Model):
id = models.UUIDField(primary_key=True, default=uuid.uuid4, editable=False)
stream_id = models.CharField(max_length=100)
stream_key = models.CharField(max_length=100) # Use stream_key instead of stream_id
segment_path = models.CharField(max_length=500)
timestamp = models.DateTimeField(auto_now_add=True)
processing_time = models.FloatField(null=True)
provider = models.ForeignKey(AnalysisProvider, on_delete=models.CASCADE, null=True, blank=True)
analysis_type = models.CharField(max_length=50)
confidence_threshold = models.FloatField(default=0.5)
confidence_threshold = models.FloatField(default=get_default_confidence_threshold)
frame_timestamp = models.FloatField()
external_request_id = models.CharField(max_length=200, null=True)
def to_dict(self):
return {
'id': str(self.id),
'stream_id': self.stream_id,
'stream_key': self.stream_key,
'timestamp': self.timestamp.isoformat(),
'processing_time': self.processing_time,
'analysis_type': self.analysis_type,
@@ -53,7 +62,7 @@ class VideoAnalysis(models.Model):
class Meta:
indexes = [
models.Index(fields=['stream_id', 'timestamp']),
models.Index(fields=['stream_key', 'timestamp']),
models.Index(fields=['analysis_type']),
]
@@ -128,7 +137,7 @@ class Brand(models.Model):
class ProcessingQueue(models.Model):
id = models.UUIDField(primary_key=True, default=uuid.uuid4, editable=False)
stream_id = models.CharField(max_length=100)
stream_key = models.CharField(max_length=100) # Use stream_key instead of stream_id
segment_path = models.CharField(max_length=500)
priority = models.IntegerField(default=0)
status = models.CharField(max_length=20, choices=[
@@ -146,5 +155,5 @@ class ProcessingQueue(models.Model):
class Meta:
indexes = [
models.Index(fields=['status', 'priority']),
models.Index(fields=['stream_id']),
models.Index(fields=['stream_key']),
]

8
media-analyzer/backend/ai_processing/processors/video_analyzer.py
View File

@@ -50,12 +50,12 @@ class VideoAnalyzer:
except Exception as e:
logger.error(f"Error setting up providers: {e}")
def queue_segment_analysis(self, stream_id, segment_path):
def queue_segment_analysis(self, stream_key, segment_path):
"""Queue video segment for analysis"""
try:
# Check if already queued
existing = ProcessingQueue.objects.filter(
stream_id=stream_id,
stream_key=stream_key,
segment_path=segment_path,
status__in=['pending', 'processing']
).exists()
@@ -66,14 +66,14 @@ class VideoAnalyzer:
# Create queue item
queue_item = ProcessingQueue.objects.create(
stream_id=stream_id,
stream_key=stream_key,
segment_path=segment_path,
analysis_types=['logo_detection'],
priority=1
)
# Trigger async processing
process_video_segment.delay(stream_id, segment_path)
process_video_segment.delay(stream_key, segment_path)
logger.info(f"Queued segment for analysis: {segment_path}")
return True

47
media-analyzer/backend/ai_processing/tasks.py
View File

@@ -11,13 +11,13 @@ channel_layer = get_channel_layer()
@shared_task(bind=True, queue='logo_detection')
def analyze_logo_detection(self, stream_id, segment_path):
def analyze_logo_detection(self, stream_key, segment_path):
"""Dedicated task for logo detection analysis"""
queue_item = None
try:
# Update queue status
queue_item = ProcessingQueue.objects.filter(
stream_id=stream_id,
stream_key=stream_key,
segment_path=segment_path,
status='pending'
).first()
@@ -50,20 +50,22 @@ def analyze_logo_detection(self, stream_id, segment_path):
queue_item.save()
return {"error": "Failed to extract frame"}
# Analyze for logos only
analysis_results = engine.analyze_frame(frame, ['logo_detection'], confidence_threshold=0.3)
# Analyze for logos only - use configured threshold
from django.conf import settings
confidence = settings.LOGO_DETECTION_CONFIG['confidence_threshold']
analysis_results = engine.analyze_frame(frame, ['logo_detection'], confidence_threshold=confidence)
# Store results
provider_info = config_manager.get_provider_by_type(logo_config['provider_type'])
provider = AnalysisProvider.objects.get(id=provider_info['id'])
analysis = VideoAnalysis.objects.create(
stream_id=stream_id,
stream_key=stream_key,
segment_path=segment_path,
provider=provider,
analysis_type='logo_detection',
frame_timestamp=0.0,
confidence_threshold=0.3
confidence_threshold=confidence
)
detections = []
@@ -83,8 +85,10 @@ def analyze_logo_detection(self, stream_id, segment_path):
# Send results via WebSocket if detections found
if detections:
websocket_group = f"stream_{stream_key}"
logger.info(f"Sending websocket update to group: {websocket_group}")
async_to_sync(channel_layer.group_send)(
f"stream_{stream_id}",
websocket_group,
{
"type": "analysis_update",
"analysis": analysis.to_dict()
@@ -96,11 +100,12 @@ def analyze_logo_detection(self, stream_id, segment_path):
queue_item.status = 'completed'
queue_item.save()
if detections:
logger.info(f"Logo detection: {len(detections)} detections found")
else:
logger.debug(f"Logo detection: no detections found")
return {"detections": len(detections), "analysis_id": str(analysis.id)}
result = {
"detections": len(detections),
"analysis_id": str(analysis.id),
"brands": [d['label'] for d in detections] if detections else []
}
return result
except Exception as e:
logger.error(f"Logo detection failed for {segment_path}: {e}")
@@ -112,13 +117,13 @@ def analyze_logo_detection(self, stream_id, segment_path):
@shared_task(bind=True, queue='visual_analysis')
def analyze_visual_properties(self, stream_id, segment_path):
def analyze_visual_properties(self, stream_key, segment_path):
"""Dedicated task for visual property analysis"""
queue_item = None
try:
# Update queue status
queue_item = ProcessingQueue.objects.filter(
stream_id=stream_id,
stream_key=stream_key,
segment_path=segment_path,
status='pending'
).first()
@@ -145,7 +150,7 @@ def analyze_visual_properties(self, stream_id, segment_path):
# Store results (no provider needed for local visual analysis)
analysis = VideoAnalysis.objects.create(
stream_id=stream_id,
stream_key=stream_key,
segment_path=segment_path,
provider=None, # Local analysis
analysis_type='visual_analysis',
@@ -165,7 +170,7 @@ def analyze_visual_properties(self, stream_id, segment_path):
# Send results via WebSocket
async_to_sync(channel_layer.group_send)(
f"stream_{stream_id}",
f"stream_{stream_key}",
{
"type": "analysis_update",
"analysis": analysis.to_dict()
@@ -190,17 +195,17 @@ def analyze_visual_properties(self, stream_id, segment_path):
@shared_task(bind=True)
def process_video_segment(self, stream_id, segment_path):
def process_video_segment(self, stream_key, segment_path):
"""Main task that dispatches to specialized analysis tasks"""
try:
# Dispatch to specialized queues based on available capabilities
active_capabilities = config_manager.get_active_capabilities()
if 'logo_detection' in active_capabilities:
analyze_logo_detection.delay(stream_id, segment_path)
analyze_logo_detection.delay(stream_key, segment_path)
# Visual analysis disabled for performance - only logo detection
# analyze_visual_properties.delay(stream_id, segment_path)
# analyze_visual_properties.delay(stream_key, segment_path)
return {"dispatched": True, "capabilities": active_capabilities}
@@ -222,7 +227,7 @@ def reload_analysis_config():
@shared_task
def analyze_frame_task(stream_id, segment_path, frame_timestamp=0.0):
def analyze_frame_task(stream_key, segment_path, frame_timestamp=0.0):
"""Analyze a single frame from video segment"""
try:
engine = AnalysisEngine()
@@ -251,7 +256,7 @@ def analyze_frame_task(stream_id, segment_path, frame_timestamp=0.0):
results = engine.analyze_frame(frame, ['logo_detection', 'visual_analysis'])
return {
"stream_id": stream_id,
"stream_key": stream_key,
"results": results,
"frame_timestamp": frame_timestamp
}

2
media-analyzer/backend/media_analyzer/asgi.py
View File

@@ -26,7 +26,7 @@ application = ProtocolTypeRouter({
"http": django_asgi_app,
"websocket": AuthMiddlewareStack(
URLRouter([
path("ws/stream/<str:stream_id>/", StreamAnalysisConsumer.as_asgi()),
path("ws/stream/", StreamAnalysisConsumer.as_asgi()),
])
),
})

2
media-analyzer/backend/media_analyzer/settings/base.py
View File

@@ -244,7 +244,7 @@ else:
# Logo Detection Configuration
LOGO_DETECTION_CONFIG = {
'confidence_threshold': float(os.getenv('LOGO_CONFIDENCE_THRESHOLD', '0.3')),
'confidence_threshold': float(os.getenv('LOGO_CONFIDENCE_THRESHOLD', '0.6')),
'enabled_brands': os.getenv('ENABLED_BRANDS', 'Apple,Google,Nike,Coca-Cola,McDonald,Amazon').split(','),
'use_cloud_vision': USE_CLOUD_VISION,
}

3
media-analyzer/backend/media_analyzer/urls.py
View File

@@ -21,7 +21,8 @@ from django.conf.urls.static import static
urlpatterns = [
path('admin/', admin.site.urls),
path('api/', include('streaming.urls')),
path('api/streaming/', include('streaming.urls')),
path('api/', include('api.urls')),
path('streaming/', include('streaming.urls')),
]

53
media-analyzer/backend/streaming/consumers.py
View File

@@ -11,28 +11,16 @@ class StreamAnalysisConsumer(AsyncWebsocketConsumer):
"""WebSocket consumer for real-time analysis updates"""
async def connect(self):
self.stream_id = self.scope['url_route']['kwargs']['stream_id']
self.room_group_name = f'stream_{self.stream_id}'
# Join stream group
await self.channel_layer.group_add(
self.room_group_name,
self.channel_name
)
# Initialize subscription set for dynamic stream groups
self.subscribed_streams = set()
await self.accept()
logger.info(f"WebSocket connected for stream {self.stream_id}")
# Send recent analysis results
await self.send_recent_analysis()
logger.info("WebSocket connected - ready to subscribe to streams")
async def disconnect(self, close_code):
# Leave stream group
await self.channel_layer.group_discard(
self.room_group_name,
self.channel_name
)
logger.info(f"WebSocket disconnected for stream {self.stream_id}")
# Leave all subscribed stream groups
for stream_key in getattr(self, 'subscribed_streams', []):
await self.channel_layer.group_discard(f"stream_{stream_key}", self.channel_name)
logger.info("WebSocket disconnected")
async def receive(self, text_data):
"""Handle incoming WebSocket messages"""
@@ -45,10 +33,20 @@ class StreamAnalysisConsumer(AsyncWebsocketConsumer):
'type': 'pong',
'timestamp': data.get('timestamp')
}))
elif message_type == 'subscribe':
stream_key = data.get('stream_id') # Frontend still sends 'stream_id' but it's actually stream_key
if stream_key and stream_key not in self.subscribed_streams:
self.subscribed_streams.add(stream_key)
await self.channel_layer.group_add(f"stream_{stream_key}", self.channel_name)
await self.send_recent_analysis(stream_key)
elif message_type == 'unsubscribe':
stream_key = data.get('stream_id') # Frontend still sends 'stream_id' but it's actually stream_key
if stream_key and stream_key in self.subscribed_streams:
self.subscribed_streams.remove(stream_key)
await self.channel_layer.group_discard(f"stream_{stream_key}", self.channel_name)
elif message_type == 'request_analysis':
# Trigger analysis if needed
pass
except json.JSONDecodeError:
logger.error("Invalid JSON received")
@@ -60,21 +58,20 @@ class StreamAnalysisConsumer(AsyncWebsocketConsumer):
}))
@database_sync_to_async
def get_recent_analysis(self):
"""Get recent analysis results for stream"""
def get_recent_analysis(self, stream_key):
"""Get recent analysis results for a given stream"""
try:
analyses = VideoAnalysis.objects.filter(
stream_id=self.stream_id
stream_key=stream_key
).order_by('-timestamp')[:5]
return [analysis.to_dict() for analysis in analyses]
except Exception as e:
logger.error(f"Error getting recent analysis: {e}")
logger.error(f"Error getting recent analysis for {stream_key}: {e}")
return []
async def send_recent_analysis(self):
"""Send recent analysis results to client"""
recent_analyses = await self.get_recent_analysis()
async def send_recent_analysis(self, stream_key):
"""Send recent analysis results to client for the given stream"""
recent_analyses = await self.get_recent_analysis(stream_key)
if recent_analyses:
await self.send(text_data=json.dumps({
'type': 'recent_analysis',

42
media-analyzer/backend/streaming/ffmpeg_handler.py
View File

@@ -1,6 +1,7 @@
import ffmpeg
import logging
import subprocess
import platform
from pathlib import Path
logger = logging.getLogger(__name__)
@@ -32,5 +33,46 @@ class FFmpegHandler:
)
return ffmpeg.run_async(output, pipe_stdout=True, pipe_stderr=True)
def webcam_to_hls(self, device_index: int, output_path: str, width: int = 640, height: int = 480, fps: int = 30) -> subprocess.Popen:
"""Convert webcam stream to HLS (cross-platform)"""
system = platform.system().lower()
if system == 'windows':
# Windows: DirectShow
stream = ffmpeg.input(
f'video="Integrated Camera"',
f='dshow',
video_size=f'{width}x{height}',
framerate=fps
)
elif system == 'darwin': # macOS
# macOS: AVFoundation
stream = ffmpeg.input(
f'{device_index}',
f='avfoundation',
video_size=f'{width}x{height}',
framerate=fps
)
else: # Linux and others
# Linux: Video4Linux2
stream = ffmpeg.input(
f'/dev/video{device_index}',
f='v4l2',
s=f'{width}x{height}',
framerate=fps
)
output = ffmpeg.output(
stream, output_path,
vcodec='libx264',
preset='ultrafast', # Fast encoding for real-time
tune='zerolatency', # Low latency
f='hls',
hls_time=2, # Short segments for responsiveness
hls_list_size=10,
hls_flags='delete_segments'
)
return ffmpeg.run_async(output, pipe_stdout=True, pipe_stderr=True)
# Singleton
ffmpeg_handler = FFmpegHandler()

84
media-analyzer/backend/streaming/file_watcher.py Normal file
View File

@@ -0,0 +1,84 @@
import os
import time
import logging
from pathlib import Path
from django.conf import settings
from ai_processing.processors.video_analyzer import VideoAnalyzer
from .models import VideoStream, StreamStatus
logger = logging.getLogger(__name__)
class HLSFileWatcher:
"""Watch for new HLS segment files and trigger analysis"""
def __init__(self, media_dir=None, poll_interval=1.0):
self.media_dir = Path(media_dir or settings.MEDIA_ROOT)
self.poll_interval = poll_interval
self.processed_files = set()
self.analyzer = VideoAnalyzer()
def get_stream_key_from_filename(self, filename):
"""Extract stream_key from filename: 'stream_key-segment_number.ts' -> 'stream_key'"""
if not filename.endswith('.ts'):
return None
base_name = filename.rsplit('.', 1)[0] # Remove .ts extension
stream_key = base_name.rsplit('-', 1)[0] # Remove last segment: "-123"
return stream_key if stream_key else None
def process_new_segment(self, file_path):
"""Process a new HLS segment file"""
try:
# Determine the active stream from the database
active_stream = VideoStream.objects.filter(status=StreamStatus.ACTIVE).first()
if not active_stream:
logger.warning(f"File watcher: No active stream found, skipping segment {file_path.name}")
return
stream_key = active_stream.stream_key
logger.info(f"File watcher: Processing new segment {file_path.name} (stream: {stream_key})")
# Queue for analysis
self.analyzer.queue_segment_analysis(stream_key, str(file_path))
logger.info(f"File watcher: Queued segment for analysis: {file_path.name}")
except Exception as e:
logger.error(f"File watcher: Error processing {file_path}: {e}")
import traceback
logger.error(f"File watcher: Traceback: {traceback.format_exc()}")
def scan_for_new_files(self):
"""Scan for new .ts files in the media directory"""
try:
if not self.media_dir.exists():
return
current_files = set()
for ts_file in self.media_dir.glob("*.ts"):
if ts_file.is_file():
current_files.add(ts_file)
# Find new files
new_files = current_files - self.processed_files
for new_file in new_files:
self.process_new_segment(new_file)
self.processed_files.add(new_file)
except Exception as e:
logger.error(f"File watcher: Error scanning directory: {e}")
def start_watching(self):
"""Start the file watching loop"""
logger.info(f"File watcher: Starting to watch {self.media_dir}")
# Initial scan to catch existing files
self.scan_for_new_files()
while True:
try:
self.scan_for_new_files()
time.sleep(self.poll_interval)
except KeyboardInterrupt:
logger.info("File watcher: Stopped by user")
break
except Exception as e:
logger.error(f"File watcher: Unexpected error: {e}")
time.sleep(self.poll_interval)

30
media-analyzer/backend/streaming/management/commands/watch_hls_files.py Normal file
View File

@@ -0,0 +1,30 @@
from django.core.management.base import BaseCommand
from streaming.file_watcher import HLSFileWatcher
class Command(BaseCommand):
help = 'Watch for new HLS segment files and trigger analysis'
def add_arguments(self, parser):
parser.add_argument(
'--poll-interval',
type=float,
default=1.0,
help='Polling interval in seconds (default: 1.0)'
)
def handle(self, *args, **options):
poll_interval = options['poll_interval']
self.stdout.write(
self.style.SUCCESS(f'Starting HLS file watcher (poll interval: {poll_interval}s)')
)
watcher = HLSFileWatcher(poll_interval=poll_interval)
try:
watcher.start_watching()
except KeyboardInterrupt:
self.stdout.write(
self.style.SUCCESS('HLS file watcher stopped')
)

157
media-analyzer/backend/streaming/source_adapters.py
View File

@@ -4,6 +4,9 @@ from pathlib import Path
from django.conf import settings
from .models import VideoStream, StreamStatus
from .ffmpeg_handler import ffmpeg_handler
import threading
import os
import signal
logger = logging.getLogger(__name__)
@@ -43,6 +46,13 @@ class RTMPSourceAdapter(VideoSourceAdapter):
try:
self.update_stream_status(StreamStatus.STARTING)
# Check if any other stream is active (only one stream allowed)
active_streams = VideoStream.objects.filter(status=StreamStatus.ACTIVE).exclude(id=self.stream.id)
if active_streams.exists():
logger.warning(f"Cannot start RTMP - another stream is active: {active_streams.first().name}")
self.update_stream_status(StreamStatus.ERROR)
return False
# Files go directly in media directory
media_dir = Path(settings.MEDIA_ROOT)
@@ -53,9 +63,14 @@ class RTMPSourceAdapter(VideoSourceAdapter):
# Start FFmpeg conversion
self.process = ffmpeg_handler.rtmp_to_hls(rtmp_url, playlist_path)
# Persist FFmpeg PID for stop operations
try:
pid_file = media_dir / f'{self.stream.stream_key}.pid'
with pid_file.open('w') as f:
f.write(str(self.process.pid))
except Exception as e:
logger.error(f"RTMPSourceAdapter: Failed to write PID file: {e}")
# HLS URL is now generated dynamically from settings
self.update_stream_status(StreamStatus.ACTIVE)
logger.info(f"Started RTMP processing for stream {self.stream.id}")
return True
@@ -68,15 +83,31 @@ class RTMPSourceAdapter(VideoSourceAdapter):
def stop_processing(self) -> bool:
try:
self.update_stream_status(StreamStatus.STOPPING)
media_dir = Path(settings.MEDIA_ROOT)
pid_file = media_dir / f'{self.stream.stream_key}.pid'
# Attempt to terminate in-memory process
if self.process and self.process.poll() is None:
self.process.terminate()
self.process.wait(timeout=10)
try:
self.process.wait(timeout=10)
except Exception:
pass
# Fallback: terminate by PID file
elif pid_file.exists():
try:
pid = int(pid_file.read_text())
os.kill(pid, signal.SIGTERM)
except Exception as kill_err:
logger.error(f"RTMPSourceAdapter: Failed to kill PID {pid}: {kill_err}")
# Cleanup PID file
if pid_file.exists():
try:
pid_file.unlink()
except Exception as unlink_err:
logger.error(f"RTMPSourceAdapter: Failed to remove PID file: {unlink_err}")
self.update_stream_status(StreamStatus.INACTIVE)
logger.info(f"Stopped RTMP processing for stream {self.stream.id}")
return True
except Exception as e:
logger.error(f"Failed to stop RTMP processing: {e}")
self.update_stream_status(StreamStatus.ERROR)
@@ -125,6 +156,119 @@ class FileSourceAdapter(VideoSourceAdapter):
return str(media_dir / f'{self.stream.stream_key}.m3u8')
class WebcamSourceAdapter(VideoSourceAdapter):
"""Adapter for webcam streams"""
def start_processing(self) -> bool:
try:
logger.info(f"Starting webcam processing for stream {self.stream.id} with key {self.stream.stream_key}")
self.update_stream_status(StreamStatus.STARTING)
# Check if any other stream is active (only one stream allowed)
active_streams = VideoStream.objects.filter(status=StreamStatus.ACTIVE).exclude(id=self.stream.id)
if active_streams.exists():
logger.warning(f"Cannot start webcam - another stream is active: {active_streams.first().name}")
self.update_stream_status(StreamStatus.ERROR)
return False
# Files go directly in media directory
media_dir = Path(settings.MEDIA_ROOT)
playlist_path = str(media_dir / f'{self.stream.stream_key}.m3u8')
logger.info(f"Webcam playlist path: {playlist_path}")
# Default to webcam 0
device_index = 0
# Start FFmpeg conversion
logger.info(f"Starting FFmpeg webcam conversion with device {device_index}")
self.process = ffmpeg_handler.webcam_to_hls(device_index, playlist_path)
# Check if FFmpeg process started successfully
if self.process.poll() is not None:
# Process already exited - get error details
try:
stdout, stderr = self.process.communicate(timeout=2)
error_msg = stderr.decode('utf-8') if stderr else "Unknown FFmpeg error"
logger.error(f"FFmpeg failed to start webcam: {error_msg}")
except Exception as comm_error:
logger.error(f"FFmpeg failed and couldn't read error: {comm_error}")
error_msg = "FFmpeg process failed to start"
self.update_stream_status(StreamStatus.ERROR)
raise Exception(f"Webcam initialization failed: {error_msg}")
logger.info(f"FFmpeg process started successfully with PID: {self.process.pid}")
# Persist FFmpeg PID for stop operations
try:
pid_file = media_dir / f'{self.stream.stream_key}.pid'
with pid_file.open('w') as f:
f.write(str(self.process.pid))
except Exception as e:
logger.error(f"WebcamSourceAdapter: Failed to write PID file: {e}")
self.update_stream_status(StreamStatus.ACTIVE)
logger.info(f"Started webcam processing for stream {self.stream.id}")
# Monitor FFmpeg process and handle unexpected termination
threading.Thread(target=self._monitor_webcam, daemon=True).start()
return True
except Exception as e:
logger.error(f"Failed to start webcam processing: {e}")
logger.exception(f"Full exception details:")
self.update_stream_status(StreamStatus.ERROR)
return False
def stop_processing(self) -> bool:
try:
self.update_stream_status(StreamStatus.STOPPING)
media_dir = Path(settings.MEDIA_ROOT)
pid_file = media_dir / f'{self.stream.stream_key}.pid'
# Attempt to terminate in-memory process
if self.process and self.process.poll() is None:
self.process.terminate()
try:
self.process.wait(timeout=10)
except Exception:
pass
# Fallback: terminate by PID file
elif pid_file.exists():
try:
pid = int(pid_file.read_text())
os.kill(pid, signal.SIGTERM)
except Exception as kill_err:
logger.error(f"WebcamSourceAdapter: Failed to kill PID {pid}: {kill_err}")
# Cleanup PID file
if pid_file.exists():
try:
pid_file.unlink()
except Exception as unlink_err:
logger.error(f"WebcamSourceAdapter: Failed to remove PID file: {unlink_err}")
self.update_stream_status(StreamStatus.INACTIVE)
logger.info(f"Stopped webcam processing for stream {self.stream.id}")
return True
except Exception as e:
logger.error(f"Failed to stop webcam processing: {e}")
self.update_stream_status(StreamStatus.ERROR)
return False
def get_hls_output_path(self) -> str:
media_dir = Path(settings.MEDIA_ROOT)
return str(media_dir / f'{self.stream.stream_key}.m3u8')
def _monitor_webcam(self):
"""Monitor the FFmpeg webcam process and update stream status on exit"""
try:
exit_code = self.process.wait()
if exit_code != 0:
logger.error(f"FFmpeg webcam process terminated unexpectedly with code {exit_code}")
new_status = StreamStatus.ERROR
else:
logger.info(f"FFmpeg webcam process terminated normally with code {exit_code}")
new_status = StreamStatus.INACTIVE
self.update_stream_status(new_status)
except Exception as e:
logger.error(f"Error monitoring FFmpeg webcam process: {e}")
class SourceAdapterFactory:
"""Factory for creating video source adapters"""
@@ -133,6 +277,7 @@ class SourceAdapterFactory:
adapters = {
'rtmp': RTMPSourceAdapter,
'file': FileSourceAdapter,
'webcam': WebcamSourceAdapter,
}
adapter_class = adapters.get(stream.source_type)

9
media-analyzer/backend/streaming/urls.py
View File

@@ -4,8 +4,9 @@ from . import views
urlpatterns = [
path('streams/', views.list_streams, name='list_streams'),
path('streams/create/', views.create_stream, name='create_stream'),
path('streams/<int:stream_id>/start/', views.start_stream, name='start_stream'),
path('streams/<int:stream_id>/stop/', views.stop_stream, name='stop_stream'),
path('streams/<str:stream_id>/analyze/', views.trigger_analysis, name='trigger_analysis'),
path('hls/<str:filename>', views.serve_hls_file, name='serve_hls_file'),
path('streams/webcam/start/', views.start_webcam_stream, name='start_webcam_stream'),
path('streams/<str:stream_key>/start/', views.start_stream, name='start_stream'),
path('streams/<str:stream_key>/stop/', views.stop_stream, name='stop_stream'),
path('streams/<str:stream_key>/analyze/', views.trigger_analysis, name='trigger_analysis'),
path('<str:filename>', views.serve_hls_file, name='serve_hls_file'),
]

152
media-analyzer/backend/streaming/views.py
View File

@@ -17,15 +17,30 @@ logger = logging.getLogger(__name__)
@csrf_exempt
@require_http_methods(["POST"])
def create_stream(request):
"""Create new stream"""
"""Create or update RTMP stream (single stream pattern like webcam)"""
try:
data = json.loads(request.body)
stream = VideoStream.objects.create(
name=data['name'],
source_type=data.get('source_type', 'rtmp'),
processing_mode=data.get('processing_mode', 'live'),
stream_key=str(uuid.uuid4())
)
source_type = data.get('source_type', 'rtmp')
# Look for existing stream of this type first
existing_stream = VideoStream.objects.filter(source_type=source_type).first()
if existing_stream:
# Update existing stream
existing_stream.name = data['name']
existing_stream.processing_mode = data.get('processing_mode', 'live')
existing_stream.save()
stream = existing_stream
logger.info(f"Updated existing {source_type} stream: {stream.id}")
else:
# Create new stream
stream = VideoStream.objects.create(
name=data['name'],
source_type=source_type,
processing_mode=data.get('processing_mode', 'live'),
stream_key=str(uuid.uuid4())
)
logger.info(f"Created new {source_type} stream: {stream.id}")
return JsonResponse({
'id': stream.id,
@@ -35,7 +50,7 @@ def create_stream(request):
'stream_key': stream.stream_key,
'status': stream.status,
'hls_playlist_url': f"{settings.HLS_BASE_URL}{settings.HLS_ENDPOINT_PATH}{stream.stream_key}.m3u8" if stream.status == 'active' else None,
'rtmp_ingest_url': f"rtmp://{request.get_host().split(':')[0]}:{settings.RTMP_PORT}/live/{stream.stream_key}",
'rtmp_ingest_url': f"rtmp://{request.get_host().split(':')[0]}:{settings.RTMP_PORT}/live",
'created_at': stream.created_at.isoformat()
})
@@ -52,9 +67,10 @@ def list_streams(request):
'name': s.name,
'source_type': s.source_type,
'processing_mode': s.processing_mode,
'stream_key': s.stream_key,
'status': s.status,
'hls_playlist_url': f"{settings.HLS_BASE_URL}{settings.HLS_ENDPOINT_PATH}{s.stream_key}.m3u8" if s.status == 'active' else None,
'rtmp_ingest_url': f"rtmp://{request.get_host().split(':')[0]}:{settings.RTMP_PORT}/live/{s.stream_key}",
'rtmp_ingest_url': f"rtmp://{request.get_host().split(':')[0]}:{settings.RTMP_PORT}/live",
'created_at': s.created_at.isoformat()
} for s in streams]
})
@@ -62,9 +78,9 @@ def list_streams(request):
@csrf_exempt
@require_http_methods(["POST"])
def start_stream(request, stream_id):
def start_stream(request, stream_key):
"""Start stream processing"""
stream = get_object_or_404(VideoStream, id=stream_id)
stream = get_object_or_404(VideoStream, stream_key=stream_key)
try:
adapter = SourceAdapterFactory.create_adapter(stream)
@@ -79,15 +95,15 @@ def start_stream(request, stream_id):
return JsonResponse({'error': 'Failed to start stream'}, status=500)
except Exception as e:
logger.error(f"Error starting stream {stream_id}: {e}")
logger.error(f"Error starting stream {stream_key}: {e}")
return JsonResponse({'error': str(e)}, status=500)
@csrf_exempt
@require_http_methods(["POST"])
def stop_stream(request, stream_id):
def stop_stream(request, stream_key):
"""Stop stream processing"""
stream = get_object_or_404(VideoStream, id=stream_id)
stream = get_object_or_404(VideoStream, stream_key=stream_key)
try:
adapter = SourceAdapterFactory.create_adapter(stream)
@@ -99,7 +115,7 @@ def stop_stream(request, stream_id):
return JsonResponse({'error': 'Failed to stop stream'}, status=500)
except Exception as e:
logger.error(f"Error stopping stream {stream_id}: {e}")
logger.error(f"Error stopping stream {stream_key}: {e}")
return JsonResponse({'error': str(e)}, status=500)
@@ -115,21 +131,27 @@ def serve_hls_file(request, filename):
# Trigger analysis for new .ts segments
if filename.endswith('.ts'):
logger.info(f"Processing .ts file request: {filename}")
try:
# Extract stream ID from UUID-based filename: 43606ec7-786c-4f7d-acf3-95981f9e5ebe-415.ts
if '-' in filename:
# Split by dash and take first 5 parts (UUID format)
parts = filename.split('-')
if len(parts) >= 5:
stream_id = '-'.join(parts[:5]) # Reconstruct UUID
# Extract stream_key from filename: "stream_key-segment_number.ts" -> "stream_key"
# Example: "69f79422-5816-4cf0-9f44-0ac1421b8b8e-123.ts" -> "69f79422-5816-4cf0-9f44-0ac1421b8b8e"
base_name = filename.rsplit('.', 1)[0] # Remove .ts extension
stream_key = base_name.rsplit('-', 1)[0] # Remove last segment: "-123"
logger.info(f"Parsed stream_key: {stream_key} from filename: {filename}")
# Queue for analysis
analyzer = VideoAnalyzer()
analyzer.queue_segment_analysis(stream_id, file_path)
logger.info(f"Queued segment for analysis: {filename} (stream: {stream_id})")
if stream_key:
# Queue for analysis
logger.info(f"Attempting to queue analysis for {filename}")
analyzer = VideoAnalyzer()
analyzer.queue_segment_analysis(stream_key, file_path)
logger.info(f"Queued segment for analysis: {filename} (stream: {stream_key})")
else:
logger.warning(f"No stream_key extracted from {filename}")
except Exception as e:
logger.error(f"Error queuing analysis for {filename}: {e}")
import traceback
logger.error(f"Traceback: {traceback.format_exc()}")
# Determine content type
if filename.endswith('.m3u8'):
@@ -148,23 +170,25 @@ def serve_hls_file(request, filename):
@csrf_exempt
@require_http_methods(["POST"])
def trigger_analysis(request, stream_id):
def trigger_analysis(request, stream_key):
"""Manually trigger analysis for testing"""
try:
data = json.loads(request.body) if request.body else {}
segment_path = data.get('segment_path')
if not segment_path:
# Find latest segment
media_dir = os.path.join(settings.BASE_DIR.parent.parent, 'media')
# Find latest segment in media directory
media_dir = settings.MEDIA_ROOT
ts_files = [f for f in os.listdir(media_dir) if f.endswith('.ts')]
if ts_files:
# Sort by filename to get the latest segment
ts_files.sort()
segment_path = os.path.join(media_dir, ts_files[-1])
else:
return JsonResponse({'error': 'No segments found'}, status=404)
analyzer = VideoAnalyzer()
success = analyzer.queue_segment_analysis(stream_id, segment_path)
success = analyzer.queue_segment_analysis(stream_key, segment_path)
if success:
return JsonResponse({'message': 'Analysis triggered', 'segment': segment_path})
@@ -174,3 +198,73 @@ def trigger_analysis(request, stream_id):
except Exception as e:
logger.error(f"Error triggering analysis: {e}")
return JsonResponse({'error': str(e)}, status=500)
@csrf_exempt
@require_http_methods(["POST"])
def start_webcam_stream(request):
"""Start or reuse existing webcam stream"""
try:
# Look for existing webcam stream first
webcam_stream = VideoStream.objects.filter(source_type='webcam').first()
if not webcam_stream:
# Create new webcam stream
webcam_stream = VideoStream.objects.create(
name='Webcam Stream',
source_type='webcam',
processing_mode='live',
stream_key=str(uuid.uuid4())
)
logger.info(f"Created new webcam stream: {webcam_stream.id}")
# Check if another stream is active
active_streams = VideoStream.objects.filter(status=StreamStatus.ACTIVE).exclude(id=webcam_stream.id)
if active_streams.exists():
other = active_streams.first()
return JsonResponse({
'error': f'Another stream is active: {other.name}',
'active_stream_key': other.stream_key,
'active_stream_name': other.name
}, status=409)
# Start the webcam stream if not already active
if webcam_stream.status != StreamStatus.ACTIVE:
adapter = SourceAdapterFactory.create_adapter(webcam_stream)
success = adapter.start_processing()
if not success:
return JsonResponse({'error': 'Failed to start webcam'}, status=500)
# Wait for HLS playlist to be ready before returning
import time
playlist_path = os.path.join(settings.MEDIA_ROOT, f"{webcam_stream.stream_key}.m3u8")
max_wait_time = 10 # seconds
wait_interval = 0.5 # seconds
elapsed_time = 0
logger.info(f"Waiting for HLS playlist to be ready: {playlist_path}")
while elapsed_time < max_wait_time:
if os.path.exists(playlist_path) and os.path.getsize(playlist_path) > 0:
logger.info(f"HLS playlist ready after {elapsed_time:.1f}s")
break
time.sleep(wait_interval)
elapsed_time += wait_interval
if not os.path.exists(playlist_path):
logger.warning(f"HLS playlist not ready after {max_wait_time}s, returning anyway")
return JsonResponse({
'id': webcam_stream.id,
'name': webcam_stream.name,
'source_type': webcam_stream.source_type,
'processing_mode': webcam_stream.processing_mode,
'stream_key': webcam_stream.stream_key,
'status': webcam_stream.status,
'hls_playlist_url': f"{settings.HLS_BASE_URL}{settings.HLS_ENDPOINT_PATH}{webcam_stream.stream_key}.m3u8",
'created_at': webcam_stream.created_at.isoformat()
})
except Exception as e:
logger.error(f"Error starting webcam stream: {e}")
return JsonResponse({'error': str(e)}, status=500)

57
media-analyzer/backend/test_ai.py
View File

@@ -1,57 +0,0 @@
#!/usr/bin/env python
"""Quick test script to verify AI pipeline works"""
import os
import django
import sys
from pathlib import Path
# Add the backend directory to Python path
backend_dir = Path(__file__).parent
sys.path.insert(0, str(backend_dir))
# Configure Django
os.environ.setdefault('DJANGO_SETTINGS_MODULE', 'media_analyzer.settings.development')
django.setup()
from ai_processing.processors.video_analyzer import VideoAnalyzer
from ai_processing.models import AnalysisProvider
from PIL import Image
import numpy as np
def test_ai_pipeline():
print("🧪 Testing AI Pipeline...")
# Check providers
providers = AnalysisProvider.objects.all()
print(f"📊 Found {providers.count()} providers:")
for p in providers:
print(f" - {p.name} ({p.provider_type}) - Active: {p.active}")
# Create test analyzer
analyzer = VideoAnalyzer()
# Create a test image (simple colored rectangle)
print("\n🖼️ Creating test image...")
test_image = Image.new('RGB', (640, 480), color='red')
# Test synchronous analysis
print("🔍 Running synchronous analysis...")
try:
result = analyzer.analyze_frame_sync(test_image)
print(f"✅ Analysis result: {result}")
if 'error' in result:
print(f"❌ Error: {result['error']}")
else:
print(f"✅ Found {len(result.get('logos', []))} logo detections")
for logo in result.get('logos', []):
print(f" - {logo['label']}: {logo['confidence']:.3f}")
except Exception as e:
print(f"❌ Analysis failed: {e}")
import traceback
traceback.print_exc()
if __name__ == "__main__":
test_ai_pipeline()

88
media-analyzer/backend/test_frontend_integration.py
View File

@@ -1,88 +0,0 @@
#!/usr/bin/env python
"""Test frontend integration by creating sample analysis data"""
import os
import django
import sys
from pathlib import Path
# Configure Django
backend_dir = Path(__file__).parent
sys.path.insert(0, str(backend_dir))
os.environ.setdefault('DJANGO_SETTINGS_MODULE', 'media_analyzer.settings.development')
django.setup()
from ai_processing.models import VideoAnalysis, DetectionResult, VisualAnalysis, AnalysisProvider
from channels.layers import get_channel_layer
from asgiref.sync import async_to_sync
def create_sample_analysis():
"""Create sample analysis data for testing frontend"""
print("🎯 Creating sample analysis data...")
# Get CLIP provider
provider = AnalysisProvider.objects.filter(provider_type='local_clip').first()
# Create analysis
analysis = VideoAnalysis.objects.create(
stream_id='test_stream',
segment_path='/fake/path.ts',
provider=provider,
analysis_type='logo_detection',
frame_timestamp=0.0,
confidence_threshold=0.3
)
# Create sample detections
DetectionResult.objects.create(
analysis=analysis,
label='Apple',
confidence=0.85,
bbox_x=0.2,
bbox_y=0.3,
bbox_width=0.3,
bbox_height=0.2,
detection_type='logo'
)
DetectionResult.objects.create(
analysis=analysis,
label='Google',
confidence=0.72,
bbox_x=0.5,
bbox_y=0.1,
bbox_width=0.25,
bbox_height=0.15,
detection_type='logo'
)
# Create visual analysis
VisualAnalysis.objects.create(
analysis=analysis,
dominant_colors=[[255, 0, 0], [0, 255, 0], [0, 0, 255]],
brightness_level=0.7,
contrast_level=0.5,
saturation_level=0.8
)
print(f"✅ Created analysis: {analysis.to_dict()}")
# Try to send via WebSocket
try:
channel_layer = get_channel_layer()
if channel_layer:
async_to_sync(channel_layer.group_send)(
"stream_test_stream",
{
"type": "analysis_update",
"analysis": analysis.to_dict()
}
)
print("📡 Sent WebSocket update")
else:
print("⚠️ No channel layer configured")
except Exception as e:
print(f"❌ WebSocket send failed: {e}")
if __name__ == "__main__":
create_sample_analysis()

92
media-analyzer/backend/test_full_pipeline.py
View File

@@ -1,92 +0,0 @@
#!/usr/bin/env python
"""Test the complete pipeline: AI analysis -> WebSocket -> Frontend"""
import os
import django
import sys
import json
import asyncio
from pathlib import Path
# Add the backend directory to Python path
backend_dir = Path(__file__).parent
sys.path.insert(0, str(backend_dir))
# Configure Django
os.environ.setdefault('DJANGO_SETTINGS_MODULE', 'media_analyzer.settings.development')
django.setup()
from ai_processing.tasks import process_video_segment
from ai_processing.models import VideoAnalysis, AnalysisProvider
from PIL import Image, ImageDraw, ImageFont
import tempfile
def create_test_image_with_apple_logo():
"""Create a test image with 'Apple' text as logo simulation"""
img = Image.new('RGB', (640, 480), color='white')
draw = ImageDraw.Draw(img)
# Draw "Apple" text in the center
try:
font = ImageFont.truetype("/usr/share/fonts/truetype/liberation/LiberationSans-Bold.ttf", 48)
except:
font = ImageFont.load_default()
text = "Apple iPhone"
bbox = draw.textbbox((0, 0), text, font=font)
text_width = bbox[2] - bbox[0]
text_height = bbox[3] - bbox[1]
x = (640 - text_width) // 2
y = (480 - text_height) // 2
draw.text((x, y), text, fill='black', font=font)
return img
def test_full_pipeline():
print("🧪 Testing Complete Pipeline...")
# Create test image
print("🖼️ Creating test image with Apple logo simulation...")
test_image = create_test_image_with_apple_logo()
# Save to temporary file as a fake video segment
with tempfile.NamedTemporaryFile(suffix='.jpg', delete=False) as tmp_file:
test_image.save(tmp_file.name)
fake_segment_path = tmp_file.name
print(f"💾 Saved test image to: {fake_segment_path}")
# Test the processing task directly
print("🚀 Triggering analysis task...")
try:
result = process_video_segment('test_stream', fake_segment_path)
print(f"✅ Task result: {result}")
# Check if analysis was stored
analysis = VideoAnalysis.objects.filter(stream_id='test_stream').last()
if analysis:
print(f"📊 Analysis stored: {analysis.to_dict()}")
detections = analysis.detections.all()
print(f"🎯 Found {detections.count()} detections:")
for detection in detections:
print(f" - {detection.label}: {detection.confidence:.3f}")
else:
print("❌ No analysis found in database")
except Exception as e:
print(f"❌ Task failed: {e}")
import traceback
traceback.print_exc()
finally:
# Cleanup
try:
os.unlink(fake_segment_path)
except:
pass
if __name__ == "__main__":
test_full_pipeline()

36
media-analyzer/docker-compose.yml
View File

@@ -9,6 +9,12 @@ services:
volumes:
- ./backend:/app
- ./media:/app/media
# Linux webcam support (comment out on macOS/Windows if needed)
devices:
- /dev/video0:/dev/video0
group_add:
- video
user: "${UID:-1000}:${GID:-1000}"
environment:
- DEBUG=1
- DB_HOST=postgres
@@ -17,6 +23,8 @@ services:
- DB_PASSWORD=media_pass
- REDIS_HOST=redis
- REDIS_PORT=6379
# RTMP ingestion host reachable from backend container
- RTMP_HOST=nginx-rtmp
- HLS_BASE_URL=http://nginx-rtmp:8081
- HLS_ENDPOINT_PATH=/
- MEDIA_ROOT=/app/media
@@ -104,6 +112,30 @@ services:
condition: service_started
command: ./entrypoint-celery.sh --queues=default,config_management --hostname=default-worker@%h
# File Watcher - Monitor HLS segments
file-watcher:
build:
context: ./backend
target: development
volumes:
- ./backend:/app
- ./media:/app/media
environment:
- DEBUG=1
- DB_HOST=postgres
- DB_NAME=media_analyzer
- DB_USER=media_user
- DB_PASSWORD=media_pass
- REDIS_HOST=redis
- REDIS_PORT=6379
- MEDIA_ROOT=/app/media
depends_on:
postgres:
condition: service_healthy
redis:
condition: service_started
command: python manage.py watch_hls_files --poll-interval=2
# Angular Frontend with NGINX (unified approach)
frontend:
build:
@@ -145,7 +177,9 @@ services:
- "0.0.0.0:1935:1935" # RTMP port - bind to all interfaces
- "0.0.0.0:8081:80" # HTTP port for HLS - bind to all interfaces
volumes:
- ./media:/var/www/html
# Persist HLS segments and playlists to host media directory
- ./media:/var/www/media
# Custom nginx.conf for RTMP/HLS configuration
- ./docker/nginx.conf:/etc/nginx/nginx.conf
depends_on:
- postgres

56
media-analyzer/docker/nginx.conf
View File

@@ -13,7 +13,7 @@ rtmp {
# Turn on HLS
hls on;
hls_path /var/www/html;
hls_path /var/www/media;
hls_fragment 3;
hls_playlist_length 60;
@@ -34,42 +34,26 @@ http {
server {
listen 80;
# HLS playlist files
location ~* \.m3u8$ {
add_header Content-Type application/vnd.apple.mpegurl;
add_header Access-Control-Allow-Origin *;
add_header Cache-Control no-cache;
root /var/www/media;
}
# HLS segment files
location ~* \.ts$ {
add_header Content-Type video/mp2t;
add_header Access-Control-Allow-Origin *;
add_header Cache-Control no-cache;
root /var/www/media;
}
# Fallback for other requests
location / {
# Disable cache
add_header 'Cache-Control' 'no-cache';
# CORS setup
add_header 'Access-Control-Allow-Origin' '*' always;
add_header 'Access-Control-Expose-Headers' 'Content-Length,Content-Range';
add_header 'Access-Control-Allow-Headers' 'Range';
# allow CORS preflight requests
if ($request_method = 'OPTIONS') {
add_header 'Access-Control-Allow-Origin' '*';
add_header 'Access-Control-Max-Age' 1728000;
add_header 'Content-Type' 'text/plain charset=UTF-8';
add_header 'Content-Length' 0;
return 204;
}
# HLS playlist files
location ~* \.m3u8$ {
add_header Content-Type application/vnd.apple.mpegurl;
add_header Cache-Control no-cache;
root /var/www/html/;
}
# HLS segment files
location ~* \.ts$ {
add_header Content-Type video/mp2t;
add_header Cache-Control no-cache;
root /var/www/html/;
}
# Default location
location / {
root /var/www/html/;
}
root /var/www/media;
}
}
}

15
media-analyzer/frontend/nginx.conf
View File

@@ -62,15 +62,16 @@ server {
proxy_set_header X-Forwarded-Proto $scheme;
}
# HLS streaming proxy
# HLS streaming proxy: route to nginx-rtmp HLS server
location /streaming/ {
resolver 127.0.0.11 valid=30s;
set $backend backend:8000;
proxy_pass http://$backend;
# HLS proxy: strip /streaming/ prefix and forward to nginx-rtmp
rewrite ^/streaming/(.*)$ /$1 break;
proxy_pass http://nginx-rtmp;
proxy_http_version 1.1;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header X-Forwarded-Proto $scheme;
add_header Access-Control-Allow-Origin *;
add_header Cache-Control no-cache;
proxy_buffering off;
}

3
media-analyzer/frontend/src/app/app.component.html
View File

@@ -6,7 +6,8 @@
<main class="main-content">
<div class="controls-section">
<app-stream-control
(streamSelected)="onStreamSelected($event)">
(streamSelected)="onStreamSelected($event)"
(streamStopped)="onStreamStopped()">
</app-stream-control>
</div>

20
media-analyzer/frontend/src/app/app.component.scss
View File

@@ -14,26 +14,44 @@
.main-content {
display: grid;
grid-template-columns: 1fr 2fr;
grid-template-columns: 350px 1fr;
grid-template-rows: auto auto;
grid-template-areas:
"controls viewer"
"controls analysis";
gap: 2rem;
max-width: 1400px;
margin: 0 auto;
@media (max-width: 768px) {
grid-template-columns: 1fr;
grid-template-rows: auto auto auto;
grid-template-areas:
"controls"
"viewer"
"analysis";
gap: 1rem;
}
}
.controls-section {
grid-area: controls;
background: #f8f9fa;
padding: 1.5rem;
border-radius: 8px;
}
.viewer-section {
grid-area: viewer;
background: #000;
padding: 1rem;
border-radius: 8px;
}
.analysis-section {
grid-area: analysis;
background: #f8f9fa;
padding: 1rem;
border-radius: 8px;
}
}

72
media-analyzer/frontend/src/app/app.component.ts
View File

@@ -1,12 +1,12 @@
import { Component, OnInit, OnDestroy } from '@angular/core';
import { Component, OnInit, OnDestroy, ViewChild } from '@angular/core';
import { RouterOutlet } from '@angular/router';
import { HttpClientModule } from '@angular/common/http';
import { StreamControlComponent } from './components/stream-control/stream-control.component';
import { StreamViewerComponent } from './components/stream-viewer/stream-viewer.component';
import { AnalysisPanelComponent } from './components/analysis-panel/analysis-panel.component';
import { AnalysisService } from './services/analysis.service';
import { StreamService } from './services/stream.service';
import { DetectionResult, VisualAnalysis, Analysis } from './models/analysis';
import { environment } from '../environments/environment';
@Component({
selector: 'app-root',
@@ -16,6 +16,8 @@ import { environment } from '../environments/environment';
styleUrl: './app.component.scss'
})
export class AppComponent implements OnInit, OnDestroy {
@ViewChild(StreamViewerComponent) streamViewer!: StreamViewerComponent;
title = 'Media Analyzer';
selectedStreamUrl: string = '';
currentStreamId: string = '';
@@ -23,7 +25,10 @@ export class AppComponent implements OnInit, OnDestroy {
currentVisual?: VisualAnalysis;
recentAnalyses: Analysis[] = [];
constructor(private analysisService: AnalysisService) {}
constructor(
private analysisService: AnalysisService,
private streamService: StreamService
) {}
ngOnInit() {
// Subscribe to analysis updates
@@ -47,20 +52,63 @@ export class AppComponent implements OnInit, OnDestroy {
onStreamSelected(streamUrl: string) {
console.log('App received stream URL:', streamUrl);
// Convert backend URL to browser-accessible URL using environment config
const browserUrl = streamUrl.replace(/^http:\/\/[^\/]+/, environment.hlsBaseUrl);
this.selectedStreamUrl = browserUrl;
console.log('Converted to browser URL:', browserUrl);
// Extract filename from backend URL, then construct a browser-resolvable URL
const filename = streamUrl.split('/').pop() || '';
this.selectedStreamUrl = `/streaming/${filename}`;
console.log('Using HLS URL:', this.selectedStreamUrl);
// Extract stream ID from filename: 476c0bd7-d037-4b6c-a29d-0773c19a76c5.m3u8
const streamIdMatch = streamUrl.match(/([0-9a-f-]+)\.m3u8/);
// Retry function to get active stream (with small delays to allow DB update)
const getActiveStreamWithRetry = (attempt = 1, maxAttempts = 3) => {
this.streamService.getStreams().subscribe({
next: (response) => {
const activeStream = response.streams.find(stream => stream.status === 'active');
if (activeStream) {
this.currentStreamId = activeStream.stream_key;
console.log('Found active stream with key:', this.currentStreamId);
// Connect to WebSocket for this stream
this.analysisService.connectToStream(this.currentStreamId);
} else if (attempt < maxAttempts) {
console.log(`No active stream found (attempt ${attempt}/${maxAttempts}), retrying in 1s...`);
setTimeout(() => getActiveStreamWithRetry(attempt + 1, maxAttempts), 1000);
} else {
console.log('No active stream found after retries, falling back to filename parsing');
this.fallbackToFilenameExtraction(filename);
}
},
error: (error) => {
console.error('Failed to get streams from API:', error);
this.fallbackToFilenameExtraction(filename);
}
});
};
// Start the retry process
getActiveStreamWithRetry();
}
private fallbackToFilenameExtraction(filename: string) {
const streamIdMatch = filename.match(/^([a-zA-Z0-9-]+)\.m3u8$/);
if (streamIdMatch) {
this.currentStreamId = streamIdMatch[1];
console.log('Extracted stream ID:', this.currentStreamId);
// Connect to WebSocket for this stream
console.log('Fallback: Extracted stream ID from filename:', this.currentStreamId);
this.analysisService.connectToStream(this.currentStreamId);
} else {
console.error('Could not extract stream ID from URL:', streamUrl);
console.error('Could not extract stream ID from filename:', filename);
}
}
onStreamStopped() {
console.log('Stream stopped - clearing player');
// Clear the stream from player
if (this.streamViewer) {
this.streamViewer.clearStream();
}
// Clear app state
this.selectedStreamUrl = '';
this.currentStreamId = '';
this.currentDetections = [];
this.currentVisual = undefined;
// Disconnect from WebSocket
this.analysisService.disconnect();
}
}

3
media-analyzer/frontend/src/app/app.config.ts
View File

@@ -1,8 +1,9 @@
import { ApplicationConfig } from '@angular/core';
import { provideRouter } from '@angular/router';
import { provideHttpClient } from '@angular/common/http';
import { routes } from './app.routes';
export const appConfig: ApplicationConfig = {
providers: [provideRouter(routes)]
providers: [provideRouter(routes), provideHttpClient()]
};

112
media-analyzer/frontend/src/app/components/stream-control/stream-control.component.html
View File

@@ -1,38 +1,94 @@
<div class="stream-control">
<h2>Stream Control</h2>
<!-- Create New Stream -->
<div class="create-stream">
<h3>Create Stream</h3>
<input
type="text"
[(ngModel)]="newStreamName"
placeholder="Stream name"
class="input">
<button (click)="createStream()" class="btn btn-primary">Create Stream</button>
<!-- Tabs -->
<div class="tabs">
<button
class="tab"
[class.active]="activeTab === 'webcam'"
(click)="switchTab('webcam')">
📹 Webcam
</button>
<button
class="tab"
[class.active]="activeTab === 'rtmp'"
(click)="switchTab('rtmp')">
📡 RTMP Stream
</button>
</div>
<!-- Stream List -->
<div class="streams-list">
<h3>Streams</h3>
<div class="stream-item" *ngFor="let stream of streams">
<div class="stream-info">
<h4>{{ stream.name }}</h4>
<p class="status">Status: <span [class]="'status-' + stream.status">{{ stream.status }}</span></p>
<p class="rtmp-url">RTMP URL: <code>{{ stream.rtmp_ingest_url }}</code></p>
<p class="stream-key">Stream Key: <code>{{ stream.stream_key }}</code></p>
<!-- Webcam Tab -->
<div class="tab-content" *ngIf="activeTab === 'webcam'">
<div class="webcam-section">
<h3>Webcam Stream</h3>
<p class="description">Start your webcam for real-time logo detection</p>
<button (click)="startWebcam()" class="btn btn-webcam">📹 Start Webcam</button>
<!-- Webcam Stream List -->
<div class="streams-list" *ngIf="webcamStreams.length > 0">
<h4>Active Webcam Stream</h4>
<div class="stream-item" *ngFor="let stream of webcamStreams">
<div class="stream-info">
<h4>{{ stream.name }}</h4>
<p class="status">Status: <span [class]="'status-' + stream.status">{{ stream.status }}</span></p>
<p class="stream-key">Stream Key: <code>{{ stream.stream_key }}</code></p>
</div>
<div class="stream-actions">
<button (click)="startStream(stream)"
[disabled]="stream.status === 'active'"
class="btn btn-success">Start</button>
<button (click)="stopStream(stream)"
[disabled]="stream.status === 'inactive'"
class="btn btn-danger">Stop</button>
<button (click)="selectStream(stream)"
*ngIf="stream.hls_playlist_url"
class="btn btn-info">Load</button>
</div>
</div>
</div>
</div>
</div>
<!-- RTMP Tab -->
<div class="tab-content" *ngIf="activeTab === 'rtmp'">
<div class="rtmp-section">
<!-- Create New Stream -->
<div class="create-stream">
<h3>Create RTMP Stream</h3>
<p class="description">Create a stream for OBS or other RTMP sources</p>
<input
type="text"
[(ngModel)]="newStreamName"
placeholder="Stream name"
class="input">
<button (click)="createStream()" class="btn btn-primary">Create Stream</button>
</div>
<div class="stream-actions">
<button (click)="startStream(stream)"
[disabled]="stream.status === 'active'"
class="btn btn-success">Start</button>
<button (click)="stopStream(stream)"
[disabled]="stream.status === 'inactive'"
class="btn btn-danger">Stop</button>
<button (click)="selectStream(stream)"
*ngIf="stream.hls_playlist_url"
class="btn btn-info">Load</button>
<!-- RTMP Stream List -->
<div class="streams-list" *ngIf="rtmpStreams.length > 0">
<h4>RTMP Streams</h4>
<div class="stream-item" *ngFor="let stream of rtmpStreams">
<div class="stream-info">
<h4>{{ stream.name }}</h4>
<p class="status">Status: <span [class]="'status-' + stream.status">{{ stream.status }}</span></p>
<p class="rtmp-url">RTMP URL: <code>{{ stream.rtmp_ingest_url }}</code></p>
<p class="stream-key">Stream Key: <code>{{ stream.stream_key }}</code></p>
</div>
<div class="stream-actions">
<button (click)="startStream(stream)"
[disabled]="stream.status === 'active'"
class="btn btn-success">Start</button>
<button (click)="stopStream(stream)"
[disabled]="stream.status === 'inactive'"
class="btn btn-danger">Stop</button>
<button (click)="selectStream(stream)"
*ngIf="stream.hls_playlist_url"
class="btn btn-info">Load</button>
</div>
</div>
</div>
</div>
</div>

134
media-analyzer/frontend/src/app/components/stream-control/stream-control.component.scss
View File

@@ -0,0 +1,134 @@
.stream-control {
padding: 20px;
}
.tabs {
display: flex;
margin-bottom: 20px;
border-bottom: 2px solid #ddd;
}
.tab {
padding: 12px 24px;
border: none;
background: transparent;
cursor: pointer;
font-size: 16px;
font-weight: 500;
border-bottom: 3px solid transparent;
margin-right: 8px;
transition: all 0.2s;
}
.tab:hover {
background: #f8f9fa;
}
.tab.active {
color: #007bff;
border-bottom-color: #007bff;
background: #f8f9fa;
}
.tab-content {
margin-top: 20px;
}
.webcam-section, .rtmp-section {
min-height: 300px;
}
.description {
color: #666;
margin-bottom: 15px;
font-style: italic;
}
.btn-webcam {
background: #4CAF50;
color: white;
border: none;
padding: 12px 24px;
border-radius: 8px;
cursor: pointer;
font-size: 16px;
font-weight: 500;
}
.btn-webcam:hover {
background: #45a049;
}
.create-stream {
margin-bottom: 20px;
padding: 15px;
border: 1px solid #ddd;
border-radius: 8px;
}
.input {
padding: 8px;
margin-right: 10px;
border: 1px solid #ddd;
border-radius: 4px;
min-width: 200px;
}
.btn {
padding: 8px 16px;
border: none;
border-radius: 4px;
cursor: pointer;
margin-right: 8px;
}
.btn-primary { background: #007bff; color: white; }
.btn-success { background: #28a745; color: white; }
.btn-danger { background: #dc3545; color: white; }
.btn-info { background: #17a2b8; color: white; }
.btn:disabled {
background: #6c757d !important;
cursor: not-allowed;
}
.streams-list {
border: 1px solid #ddd;
border-radius: 8px;
padding: 15px;
}
.stream-item {
display: flex;
justify-content: space-between;
align-items: center;
padding: 15px;
border-bottom: 1px solid #eee;
}
.stream-item:last-child {
border-bottom: none;
}
.stream-info h4 {
margin: 0 0 8px 0;
}
.source-type {
font-weight: normal;
font-size: 0.8em;
color: #666;
}
.status-active { color: #28a745; font-weight: bold; }
.status-inactive { color: #6c757d; }
.status-starting { color: #ffc107; }
.status-stopping { color: #fd7e14; }
.status-error { color: #dc3545; font-weight: bold; }
code {
background: #f8f9fa;
padding: 2px 4px;
border-radius: 3px;
font-size: 0.9em;
}

76
media-analyzer/frontend/src/app/components/stream-control/stream-control.component.ts
View File

@@ -1,20 +1,8 @@
import { Component, EventEmitter, Output } from '@angular/core';
import { HttpClient } from '@angular/common/http';
import { CommonModule } from '@angular/common';
import { FormsModule } from '@angular/forms';
import { environment } from '../../../environments/environment';
interface Stream {
id: number;
name: string;
source_type: string;
processing_mode: string;
status: string;
stream_key: string;
hls_playlist_url: string | null;
rtmp_ingest_url: string;
created_at: string;
}
import { StreamService } from '../../services/stream.service';
import { Stream } from '../../models/stream';
@Component({
selector: 'app-stream-control',
@@ -25,17 +13,19 @@ interface Stream {
})
export class StreamControlComponent {
@Output() streamSelected = new EventEmitter<string>();
@Output() streamStopped = new EventEmitter<void>();
streams: Stream[] = [];
newStreamName = '';
selectedStream: Stream | null = null;
activeTab: 'rtmp' | 'webcam' = 'webcam';
constructor(private http: HttpClient) {
constructor(private streamService: StreamService) {
this.loadStreams();
}
loadStreams() {
this.http.get<{streams: Stream[]}>(`${environment.apiUrl}/streams/`).subscribe({
this.streamService.getStreams().subscribe({
next: (response) => {
this.streams = response.streams;
},
@@ -46,7 +36,7 @@ export class StreamControlComponent {
createStream() {
if (!this.newStreamName) return;
this.http.post<Stream>(`${environment.apiUrl}/streams/create/`, {
this.streamService.createStream({
name: this.newStreamName,
source_type: 'rtmp',
processing_mode: 'live'
@@ -60,8 +50,12 @@ export class StreamControlComponent {
}
startStream(stream: Stream) {
this.http.post(`${environment.apiUrl}/streams/${stream.id}/start/`, {}).subscribe({
next: () => {
this.streamService.startStream(stream.stream_key).subscribe({
next: (response) => {
console.log('Stream started, HLS URL:', response.hls_playlist_url);
// Emit the stream selection immediately with the HLS URL from response
this.streamSelected.emit(response.hls_playlist_url);
// Then reload streams to get updated status
this.loadStreams();
},
error: (error) => console.error('Error starting stream:', error)
@@ -69,14 +63,56 @@ export class StreamControlComponent {
}
stopStream(stream: Stream) {
this.http.post(`${environment.apiUrl}/streams/${stream.id}/stop/`, {}).subscribe({
this.streamService.stopStream(stream.stream_key).subscribe({
next: () => {
this.loadStreams();
// Emit event to clear the player
this.streamStopped.emit();
},
error: (error) => console.error('Error stopping stream:', error)
});
}
startWebcam() {
this.streamService.startWebcamStream().subscribe({
next: (stream) => {
this.loadStreams();
// Backend now waits for HLS to be ready, so we can directly select
this.selectStream(stream);
},
error: (error) => {
console.error('Error starting webcam:', error);
if (error.status === 409) {
const activeStreamKey = error.error.active_stream_key;
if (activeStreamKey) {
console.log(`Stopping active stream ${activeStreamKey} before retrying webcam`);
this.streamService.stopStream(activeStreamKey).subscribe({
next: () => this.startWebcam(),
error: (stopError) => {
console.error('Error stopping active stream:', stopError);
alert(`Cannot start webcam: ${error.error.error}`);
}
});
} else {
alert(`Cannot start webcam: ${error.error.error}`);
}
}
}
});
}
switchTab(tab: 'rtmp' | 'webcam') {
this.activeTab = tab;
}
get rtmpStreams() {
return this.streams.filter(stream => stream.source_type === 'rtmp');
}
get webcamStreams() {
return this.streams.filter(stream => stream.source_type === 'webcam');
}
selectStream(stream: Stream) {
this.selectedStream = stream;
if (stream.hls_playlist_url) {

20
media-analyzer/frontend/src/app/components/stream-viewer/stream-viewer.component.ts
View File

@@ -188,4 +188,24 @@ export class StreamViewerComponent implements AfterViewInit, OnDestroy, OnChange
const canvas = this.overlayElement.nativeElement;
this.ctx.clearRect(0, 0, canvas.width, canvas.height);
}
clearStream() {
const video = this.videoElement.nativeElement;
// Stop and clear HLS
if (this.hls) {
this.hls.destroy();
this.hls = undefined;
}
// Clear video source and stop playback
video.src = '';
video.srcObject = null;
video.load(); // Reset video element
// Clear overlay
this.clearOverlay();
console.log('Stream cleared');
}
}

9
media-analyzer/frontend/src/app/models/stream.ts
View File

@@ -1,2 +1,11 @@
export interface Stream {
id: number;
name: string;
source_type: 'rtmp' | 'webcam' | 'file' | 'url';
processing_mode: 'live' | 'batch';
status: 'inactive' | 'starting' | 'active' | 'stopping' | 'error';
stream_key: string;
hls_playlist_url?: string;
rtmp_ingest_url?: string;
created_at: string;
}

21
media-analyzer/frontend/src/app/services/analysis.service.ts
View File

@@ -10,6 +10,7 @@ export class AnalysisService {
private currentDetections = new BehaviorSubject<DetectionResult[]>([]);
private currentVisual = new BehaviorSubject<VisualAnalysis | null>(null);
private recentAnalyses = new BehaviorSubject<Analysis[]>([]);
private streamStartTime: Date | null = null;
public detections$ = this.currentDetections.asObservable();
public visual$ = this.currentVisual.asObservable();
@@ -23,16 +24,34 @@ export class AnalysisService {
}
connectToStream(streamId: string) {
this.websocketService.connect(streamId);
this.streamStartTime = new Date();
this.websocketService.subscribe(streamId);
}
disconnect() {
this.websocketService.unsubscribe();
this.websocketService.disconnect();
this.currentDetections.next([]);
this.currentVisual.next(null);
this.streamStartTime = null;
}
private handleAnalysisUpdate(analysis: Analysis) {
console.log('Received analysis update:', analysis);
// Filter out analysis from before stream started (with 30 second buffer for recent analysis)
if (this.streamStartTime && analysis.timestamp) {
const analysisTime = new Date(analysis.timestamp);
const bufferTime = new Date(this.streamStartTime.getTime() - 30000); // 30 seconds before stream start
if (analysisTime < bufferTime) {
console.log('Ignoring old analysis from before stream start:', {
analysisTime: analysisTime.toISOString(),
streamStart: this.streamStartTime.toISOString()
});
return;
}
}
// Update recent analyses list
const current = this.recentAnalyses.value;
const updated = [analysis, ...current.slice(0, 9)]; // Keep last 10

27
media-analyzer/frontend/src/app/services/stream.service.ts
View File

@@ -1,9 +1,34 @@
import { Injectable } from '@angular/core';
import { HttpClient } from '@angular/common/http';
import { Observable } from 'rxjs';
import { Stream } from '../models/stream';
import { environment } from '../../environments/environment';
@Injectable({
providedIn: 'root'
})
export class StreamService {
private apiUrl = `${environment.apiUrl}/streaming`;
constructor() { }
constructor(private http: HttpClient) { }
getStreams(): Observable<{streams: Stream[]}> {
return this.http.get<{streams: Stream[]}>(`${this.apiUrl}/streams/`);
}
createStream(streamData: any): Observable<Stream> {
return this.http.post<Stream>(`${this.apiUrl}/streams/create/`, streamData);
}
startWebcamStream(): Observable<Stream> {
return this.http.post<Stream>(`${this.apiUrl}/streams/webcam/start/`, {});
}
startStream(streamKey: string): Observable<{message: string, hls_playlist_url: string}> {
return this.http.post<{message: string, hls_playlist_url: string}>(`${this.apiUrl}/streams/${streamKey}/start/`, {});
}
stopStream(streamKey: string): Observable<{message: string}> {
return this.http.post<{message: string}>(`${this.apiUrl}/streams/${streamKey}/stop/`, {});
}
}

19
media-analyzer/frontend/src/app/services/websocket.service.ts
View File

@@ -7,6 +7,7 @@ import { Analysis } from '../models/analysis';
})
export class WebsocketService {
private socket?: WebSocket;
private currentStreamId?: string;
private analysisSubject = new Subject<Analysis>();
private connectionStatus = new BehaviorSubject<boolean>(false);
@@ -15,13 +16,13 @@ export class WebsocketService {
constructor() { }
connect(streamId: string) {
connect() {
if (this.socket?.readyState === WebSocket.OPEN) {
return;
}
const protocol = window.location.protocol === 'https:' ? 'wss:' : 'ws:';
const wsUrl = `${protocol}//${window.location.host}/ws/stream/${streamId}/`;
const wsUrl = `${protocol}//${window.location.host}/ws/stream/`;
console.log('Connecting to WebSocket:', wsUrl);
this.socket = new WebSocket(wsUrl);
@@ -63,6 +64,7 @@ export class WebsocketService {
if (this.socket) {
this.socket.close();
this.socket = undefined;
this.currentStreamId = undefined;
this.connectionStatus.next(false);
}
}
@@ -79,4 +81,17 @@ export class WebsocketService {
timestamp: Date.now()
});
}
subscribe(streamId: string) {
this.currentStreamId = streamId;
this.connect();
this.send({ type: 'subscribe', stream_id: streamId });
}
unsubscribe() {
if (this.currentStreamId) {
this.send({ type: 'unsubscribe', stream_id: this.currentStreamId });
this.currentStreamId = undefined;
}
}
}

13
thevideo.MD
View File

@@ -18,19 +18,20 @@ video structure
keyboards (early days)
music vs coding (the gap gets wider)
recurrent back to basics
behind the scenes
the setup
deskmeter
timelapses/ffmpeg
demo
phase 1
phase 2
phase 3
extras
behind the scenes
the setup
deskmeter
timelapses/ffmpeg
make your own path
opinions
bootcamps
pimp-up-your-profile new trend
for seenka
show the current state of my use of AI tools
motivations
im not in it (just) for the money
video processing is my passion (?