feat(voice): migrate to push2talk and serialize websocket text sends

components/domain/internal/voice_interaction_internal.h

@@ -8,6 +8,7 @@
 #include "esp_opus_enc.h"
 #include "esp_websocket_client.h"
 #include "freertos/FreeRTOS.h"
+#include "freertos/queue.h"
 #include "freertos/semphr.h"
 #include "freertos/task.h"
 #include "domain.h"
@@ -21,7 +22,7 @@
 
 #define VOICE_SAMPLE_BYTES 2
 #define VOICE_STATUS_LOCK_TIMEOUT_MS 1000
-#define VOICE_WS_SEND_TIMEOUT_MS 1000
+#define VOICE_WS_SEND_TIMEOUT_MS 3000
 #define VOICE_WS_TASK_STACK 24576
 #define VOICE_UPLINK_TASK_STACK 32768
 #define VOICE_HEARTBEAT_TASK_STACK 4096
@@ -31,15 +32,25 @@
 #define VOICE_AUDIO_DRAIN_QUIET_MS 180
 #define VOICE_AUDIO_DRAIN_WAIT_MS 2200
 #define VOICE_AUDIO_DRAIN_WAIT_ON_STOP_MS 3500
+#define VOICE_WS_TEXT_QUEUE_LEN 16
+#define VOICE_WS_TEXT_QUEUE_RX_TIMEOUT_MS 100
+
+typedef struct {
+    char *text;
+    int len;
+} voice_ws_text_msg_t;
 
 typedef struct {
     SemaphoreHandle_t lock;
     esp_websocket_client_handle_t ws;
+    QueueHandle_t ws_text_queue;
 
     TaskHandle_t uplink_task;
     TaskHandle_t heartbeat_task;
+    TaskHandle_t ws_text_task;
     bool uplink_task_with_caps;
     bool heartbeat_task_with_caps;
+    bool ws_text_task_with_caps;
 
     void *opus_enc;
     void *opus_dec;
@@ -63,7 +74,10 @@ typedef struct {
     bool ws_connected;
     bool started;
     bool tap_active;
+    bool speech_active;
     bool stop_requested;
+    bool ws_text_shutdown;
+    bool ws_text_busy;
     bool ws_low_stack_warned;
     bool response_first_sentence_checked;
 
@@ -129,6 +143,10 @@ void voice_websocket_event_handler(void *handler_args,
                                    void *event_data);
 
 void voice_interaction_ws_init_cjson_hooks(void);
+esp_err_t voice_interaction_try_send_speech(char *err, size_t err_len);
 
 void voice_uplink_task(void *arg);
 void voice_heartbeat_task(void *arg);
+void voice_ws_text_task(void *arg);
+
+esp_err_t voice_interaction_mic_key_init(void);