feat(mixer): enhance mixer functionality with interaction and tool legs

- Updated mixer to handle participant and isolated leg roles, allowing for IVR and consent interactions.
- Introduced commands for starting and canceling interactions, managing tool legs for recording and transcription.
- Implemented per-source audio handling for tool legs, enabling separate audio processing.
- Enhanced DTMF handling to forward events between participant legs only.
- Added support for PCM recording directly from tool legs, with WAV file generation.
- Updated TypeScript definitions and functions to support new interaction and tool leg features.

rust/crates/proxy-engine/src/call.rs

@@ -5,6 +5,7 @@
 
 use crate::mixer::{MixerCommand, RtpPacket};
 use crate::sip_leg::SipLeg;
+use sip_proto::message::SipMessage;
 use std::collections::HashMap;
 use std::net::SocketAddr;
 use std::sync::Arc;
@@ -59,6 +60,7 @@ pub enum LegKind {
     SipDevice,
     WebRtc,
     Media, // voicemail playback, IVR, recording
+    Tool,  // observer leg for recording, transcription, etc.
 }
 
 impl LegKind {
@@ -68,6 +70,7 @@ impl LegKind {
             Self::SipDevice => "sip-device",
             Self::WebRtc => "webrtc",
             Self::Media => "media",
+            Self::Tool => "tool",
         }
     }
 }
@@ -113,6 +116,10 @@ pub struct LegInfo {
     pub remote_media: Option<SocketAddr>,
     /// SIP signaling address (provider or device).
     pub signaling_addr: Option<SocketAddr>,
+
+    /// Flexible key-value metadata (consent state, tool config, etc.).
+    /// Persisted into call history on call end.
+    pub metadata: HashMap<String, serde_json::Value>,
 }
 
 /// A multiparty call with N legs and a central mixer.
@@ -127,6 +134,10 @@ pub struct Call {
     pub callee_number: Option<String>,
     pub provider_id: String,
 
+    /// Original INVITE from the device (for device-originated outbound calls).
+    /// Used to construct proper 180/200/error responses back to the device.
+    pub device_invite: Option<SipMessage>,
+
     /// All legs in this call, keyed by leg ID.
     pub legs: HashMap<LegId, LegInfo>,
 
@@ -153,6 +164,7 @@ impl Call {
             caller_number: None,
             callee_number: None,
             provider_id,
+            device_invite: None,
             legs: HashMap::new(),
             mixer_cmd_tx,
             mixer_task: Some(mixer_task),
@@ -207,6 +219,13 @@ impl Call {
             .values()
             .filter(|l| l.state != LegState::Terminated)
             .map(|l| {
+                let metadata: serde_json::Value = if l.metadata.is_empty() {
+                    serde_json::json!({})
+                } else {
+                    serde_json::Value::Object(
+                        l.metadata.iter().map(|(k, v)| (k.clone(), v.clone())).collect(),
+                    )
+                };
                 serde_json::json!({
                     "id": l.id,
                     "type": l.kind.as_str(),
@@ -214,6 +233,7 @@ impl Call {
                     "codec": sip_proto::helpers::codec_name(l.codec_pt),
                     "rtpPort": l.rtp_port,
                     "remoteMedia": l.remote_media.map(|a| format!("{}:{}", a.ip(), a.port())),
+                    "metadata": metadata,
                 })
             })
             .collect();