Fix SSE timeout issues with long-running stream discovery

Problem:
- WriteTimeout was 30 seconds
- Progress only sent when values changed
- Long ffprobe tests (7-8s each) could cause 30+ seconds without writes
- Result: "curl: (18) transfer closed with outstanding read data remaining"

Solution:
- Increase WriteTimeout from 30s to 5 minutes
- Send progress every 1 second (instead of 3 seconds)
- Always send progress, even if values unchanged
- Guarantees write every second, preventing timeout

Changes:
- internal/config/config.go: WriteTimeout 30s → 5min
- internal/camera/discovery/scanner.go:
  - Progress ticker 3s → 1s
  - Remove "only if changed" check
  - Always send progress to keep connection alive

Testing:
- HiWatch camera with 591 streams: Previously timed out at ~338/591
- Should now complete all 591 streams without timeout

internal/config/config.go

@@ -73,7 +73,7 @@ func Load() *Config {
 		Server: ServerConfig{
 			Listen:       ":4567", // Default listen address
 			ReadTimeout:  30 * time.Second,
-			WriteTimeout: 30 * time.Second,
+			WriteTimeout: 5 * time.Minute, // Increased for SSE long-polling
 		},
 		Database: DatabaseConfig{
 			DataPath:       dataPath,