chore: add new supports for multisegmented sensor formats

Author: mwfarb

AllSpark-ios/AppConstants.swift

@@ -92,10 +92,8 @@ enum AppConstants {
     }
 
     enum Video {
-        static let defaultChunkDurationMs: Int = 10000
         static let dimensionHigh: Int = 1920
         static let dimensionLow: Int = 1080
-        static let defaultFPS: Double = 30.0
     }
 
     enum Audio {
@@ -109,7 +107,17 @@ enum AppConstants {
     }
 
     enum Storage {
-        static let defaultVideoBufferMaxMB: Int = 16000
         static let bytesPerMB: Int64 = 1024 * 1024
     }
+
+    enum ClientConfig {
+        static let defaultChunkDurationMs: Int = 10000
+        static let defaultFPS: Double = 30.0
+        static let defaultVideoBufferMaxMB: Int = 16000
+        static let defaultVideoFormat: String = "mp4"
+        static let defaultAudioFormat: String = "wav"
+        static let defaultDepthFormat: String = "png"
+        static let defaultPoseFormat: String = "json"
+        static let defaultTimestampFormat: String = "txt"
+    }
 }

AllSpark-ios/CameraViewController.swift

@@ -51,7 +51,7 @@ class CameraViewController: UIViewController, UINavigationControllerDelegate {
     private var sessionAtSourceTime: CMTime?
     private var videoURL: URL?
     private var videoFormat: AVFileType = .mp4 // Default format
-    private var recordingDurationMs: Int = AppConstants.Video.defaultChunkDurationMs
+    private var recordingDurationMs: Int = AppConstants.ClientConfig.defaultChunkDurationMs
     private var autoStopTimer: Timer?
     private var shouldUploadAfterRecording = false
     private let recordingStateLock = NSLock()
@@ -335,7 +335,7 @@ class CameraViewController: UIViewController, UINavigationControllerDelegate {
                     // We upload if there is ANY overlap.
 
                     // Get current chunk duration from config
-                    var chunkDuration = Double(AppConstants.Video.defaultChunkDurationMs) / 1000.0
+                    var chunkDuration = Double(AppConstants.ClientConfig.defaultChunkDurationMs) / 1000.0
                     if let config = ConnectionManager.shared.clientConfig,
                        let durationMs = config["videoChunkDurationMs"] as? Int {
                         chunkDuration = Double(durationMs) / 1000.0
@@ -545,7 +545,7 @@ class CameraViewController: UIViewController, UINavigationControllerDelegate {
                 }
 
                 // Chunk Timer
-                var chunkMs = AppConstants.Video.defaultChunkDurationMs
+                var chunkMs = AppConstants.ClientConfig.defaultChunkDurationMs
                 if let config = ConnectionManager.shared.clientConfig,
                    let ms = config["videoChunkDurationMs"] as? Int {
                     chunkMs = ms
@@ -689,7 +689,7 @@ class CameraViewController: UIViewController, UINavigationControllerDelegate {
                    }
 
                    let asset = AVURLAsset(url: urlToProcess)
-                   var fps: Double = AppConstants.Video.defaultFPS
+                   var fps: Double = AppConstants.ClientConfig.defaultFPS
                    var width: Double = 0
                    var height: Double = 0
 

AllSpark-ios/ConnectionManager.swift

@@ -525,7 +525,7 @@ class ConnectionManager: NSObject, ObservableObject {
             let documentsPath = FileManager.default.urls(for: .documentDirectory, in: .userDomainMask)[0]
 
             // Get limit from config
-            var limitMB = AppConstants.Storage.defaultVideoBufferMaxMB
+            var limitMB = AppConstants.ClientConfig.defaultVideoBufferMaxMB
             if let config = self?.clientConfig,
                let configLimit = config["videoBufferMaxMB"] as? Int {
                 limitMB = configLimit

AllSpark-ios/SettingsView.swift

@@ -129,6 +129,7 @@ struct SettingsView: View {
                     Picker("Mode", selection: $privacyMode) {
                         Text("Person Segmentation (Default)").tag("segmentation")
                         Text("Body Pose Detection (Limbs)").tag("pose")
+                        Text("No Privacy Filter").tag("none")
                     }
                 }
             }

README.md

@@ -20,6 +20,37 @@ See also: **[CHANGELOG.md](CHANGELOG.md)** · **[RELEASE.md](RELEASE.md)**
 - Server-managed communications policy with app-level enforcement
 - Configurable video format (MP4/MOV) synced from server
 
+## Client Configuration
+
+The iOS application receives a `clientConfig` JSON object from the Edge Server to remotely configure capture formats and sensor streams. The new architecture separates sensor streams to allow for multimodal VLM/LLM analysis on the edge server.
+
+### Default Sensor Streams
+
+Here are the default expectations and supported formats:
+
+```json
+"clientConfig": {
+  "fps": 30,
+  "videoChunkDurationMs": 10000,
+  "videoBufferMaxMB": 16000,
+  "videoFormat": "mp4",
+  "audioFormat": "wav",
+  "depthFormat": "png",
+  "poseFormat": "json",
+  "timestampFormat": "txt"
+}
+```
+
+### Format Options
+- **`videoFormat`** (rgb only, no audio): `"mp4"`, `"mov"`, `"none"`
+- **`audioFormat`**: `"wav"`, `"m4a"`, `"none"`
+- **`depthFormat`**: `"png"`, `"exr"`, `"none"`
+- **`poseFormat`**: `"json"`, `"none"`
+- **`timestampFormat`**: `"txt"`, `"none"`
+
+> [!NOTE]
+> Setting any format to `"none"` will remotely disable collection from that specific sensor. This allows operators to save bandwidth and compute resources when certain streams are not needed for their agentic pipelines.
+
 ## API & WebSocket Communication
 
 See the **[Endpoints Documentation](https://github.com/WiseLabCMU/AllSpark-edge-server/blob/master/docs/endpoints.md)** in the AllSpark Edge Server repository.