Continuous voice_assistant and silence detection (#4892)

esphome/components/api/api.proto

@@ -1397,6 +1397,7 @@ message VoiceAssistantRequest {
   option (ifdef) = "USE_VOICE_ASSISTANT";
 
   bool start = 1;
+  string conversation_id = 2;
 }
 
 message VoiceAssistantResponse {

esphome/components/api/api_connection.cpp

@@ -895,11 +895,12 @@ BluetoothConnectionsFreeResponse APIConnection::subscribe_bluetooth_connections_
 #endif
 
 #ifdef USE_VOICE_ASSISTANT
-bool APIConnection::request_voice_assistant(bool start) {
+bool APIConnection::request_voice_assistant(bool start, const std::string &conversation_id) {
   if (!this->voice_assistant_subscription_)
     return false;
   VoiceAssistantRequest msg;
   msg.start = start;
+  msg.conversation_id = conversation_id;
   return this->send_voice_assistant_request(msg);
 }
 void APIConnection::on_voice_assistant_response(const VoiceAssistantResponse &msg) {

esphome/components/api/api_connection.h

@@ -128,7 +128,7 @@ class APIConnection : public APIServerConnection {
   void subscribe_voice_assistant(const SubscribeVoiceAssistantRequest &msg) override {
     this->voice_assistant_subscription_ = msg.subscribe;
   }
-  bool request_voice_assistant(bool start);
+  bool request_voice_assistant(bool start, const std::string &conversation_id);
   void on_voice_assistant_response(const VoiceAssistantResponse &msg) override;
   void on_voice_assistant_event_response(const VoiceAssistantEventResponse &msg) override;
 #endif

esphome/components/api/api_pb2.cpp

@@ -6187,7 +6187,20 @@ bool VoiceAssistantRequest::decode_varint(uint32_t field_id, ProtoVarInt value)
       return false;
   }
 }
-void VoiceAssistantRequest::encode(ProtoWriteBuffer buffer) const { buffer.encode_bool(1, this->start); }
+bool VoiceAssistantRequest::decode_length(uint32_t field_id, ProtoLengthDelimited value) {
+  switch (field_id) {
+    case 2: {
+      this->conversation_id = value.as_string();
+      return true;
+    }
+    default:
+      return false;
+  }
+}
+void VoiceAssistantRequest::encode(ProtoWriteBuffer buffer) const {
+  buffer.encode_bool(1, this->start);
+  buffer.encode_string(2, this->conversation_id);
+}
 #ifdef HAS_PROTO_MESSAGE_DUMP
 void VoiceAssistantRequest::dump_to(std::string &out) const {
   __attribute__((unused)) char buffer[64];
@@ -6195,6 +6208,10 @@ void VoiceAssistantRequest::dump_to(std::string &out) const {
   out.append("  start: ");
   out.append(YESNO(this->start));
   out.append("\n");
+
+  out.append("  conversation_id: ");
+  out.append("'").append(this->conversation_id).append("'");
+  out.append("\n");
   out.append("}");
 }
 #endif

esphome/components/api/api_pb2.h

@@ -1604,12 +1604,14 @@ class SubscribeVoiceAssistantRequest : public ProtoMessage {
 class VoiceAssistantRequest : public ProtoMessage {
  public:
   bool start{false};
+  std::string conversation_id{};
   void encode(ProtoWriteBuffer buffer) const override;
 #ifdef HAS_PROTO_MESSAGE_DUMP
   void dump_to(std::string &out) const override;
 #endif
 
  protected:
+  bool decode_length(uint32_t field_id, ProtoLengthDelimited value) override;
   bool decode_varint(uint32_t field_id, ProtoVarInt value) override;
 };
 class VoiceAssistantResponse : public ProtoMessage {

esphome/components/api/api_server.cpp

@@ -428,16 +428,16 @@ void APIServer::on_shutdown() {
 }
 
 #ifdef USE_VOICE_ASSISTANT
-bool APIServer::start_voice_assistant() {
+bool APIServer::start_voice_assistant(const std::string &conversation_id) {
   for (auto &c : this->clients_) {
-    if (c->request_voice_assistant(true))
+    if (c->request_voice_assistant(true, conversation_id))
       return true;
   }
   return false;
 }
 void APIServer::stop_voice_assistant() {
   for (auto &c : this->clients_) {
-    if (c->request_voice_assistant(false))
+    if (c->request_voice_assistant(false, ""))
       return;
   }
 }

esphome/components/api/api_server.h

@@ -96,7 +96,7 @@ class APIServer : public Component, public Controller {
 #endif
 
 #ifdef USE_VOICE_ASSISTANT
-  bool start_voice_assistant();
+  bool start_voice_assistant(const std::string &conversation_id);
   void stop_voice_assistant();
 #endif