Add push to talk voice assistant (#4648)

* Add push to talk voice assistant

* Refactor most code into voice_assistant

* Make voice_assistant the component and remove push_to_talk (can be done in yaml)

* Fix component setup

* Always AF_INET to match serverside

* Fix microphone and media player co-existence

* Format

* Update codeowners

* Update test file

* Fix endifs

* nullptr not NULL

* clang-tidy

* Format

* fixup: Add VA event data

* Generate proto

* Parse and log events

* Add default to switch

* Fix

* Add mic/va to test5

esphome/components/i2s_audio/microphone/__init__.py

@@ -0,0 +1,41 @@
+import esphome.config_validation as cv
+import esphome.codegen as cg
+
+from esphome import pins
+from esphome.const import CONF_ID
+from esphome.components import microphone
+
+from .. import (
+    i2s_audio_ns,
+    I2SAudioComponent,
+    I2SAudioIn,
+    CONF_I2S_AUDIO_ID,
+    CONF_I2S_DIN_PIN,
+)
+
+CODEOWNERS = ["@jesserockz"]
+DEPENDENCIES = ["i2s_audio"]
+
+I2SAudioMicrophone = i2s_audio_ns.class_(
+    "I2SAudioMicrophone", I2SAudioIn, microphone.Microphone, cg.Component
+)
+
+CONFIG_SCHEMA = microphone.MICROPHONE_SCHEMA.extend(
+    {
+        cv.GenerateID(): cv.declare_id(I2SAudioMicrophone),
+        cv.GenerateID(CONF_I2S_AUDIO_ID): cv.use_id(I2SAudioComponent),
+        cv.Required(CONF_I2S_DIN_PIN): pins.internal_gpio_output_pin_number,
+    }
+).extend(cv.COMPONENT_SCHEMA)
+
+
+async def to_code(config):
+    var = cg.new_Pvariable(config[CONF_ID])
+    await cg.register_component(var, config)
+
+    parent = await cg.get_variable(config[CONF_I2S_AUDIO_ID])
+    cg.add(parent.register_audio_in(var))
+
+    cg.add(var.set_din_pin(config[CONF_I2S_DIN_PIN]))
+
+    await microphone.register_microphone(var, config)

esphome/components/i2s_audio/microphone/i2s_audio_microphone.cpp

@@ -0,0 +1,101 @@
+#include "i2s_audio_microphone.h"
+
+#ifdef USE_ESP32
+
+#include <driver/i2s.h>
+
+#include "esphome/core/hal.h"
+#include "esphome/core/log.h"
+
+namespace esphome {
+namespace i2s_audio {
+
+static const size_t BUFFER_SIZE = 512;
+
+static const char *const TAG = "i2s_audio.microphone";
+
+void I2SAudioMicrophone::setup() {
+  ESP_LOGCONFIG(TAG, "Setting up I2S Audio Microphone...");
+  this->buffer_.resize(BUFFER_SIZE);
+}
+
+void I2SAudioMicrophone::start() { this->state_ = microphone::STATE_STARTING; }
+void I2SAudioMicrophone::start_() {
+  if (!this->parent_->try_lock()) {
+    return;  // Waiting for another i2s to return lock
+  }
+  i2s_driver_config_t config = {
+      .mode = (i2s_mode_t) (I2S_MODE_MASTER | I2S_MODE_RX | I2S_MODE_PDM),
+      .sample_rate = 16000,
+      .bits_per_sample = I2S_BITS_PER_SAMPLE_16BIT,
+      .channel_format = I2S_CHANNEL_FMT_ONLY_RIGHT,
+      .communication_format = I2S_COMM_FORMAT_STAND_I2S,
+      .intr_alloc_flags = ESP_INTR_FLAG_LEVEL1,
+      .dma_buf_count = 4,
+      .dma_buf_len = 256,
+      .use_apll = false,
+      .tx_desc_auto_clear = false,
+      .fixed_mclk = 0,
+      .mclk_multiple = I2S_MCLK_MULTIPLE_DEFAULT,
+      .bits_per_chan = I2S_BITS_PER_CHAN_DEFAULT,
+  };
+
+  i2s_driver_install(this->parent_->get_port(), &config, 0, nullptr);
+
+  i2s_pin_config_t pin_config = this->parent_->get_pin_config();
+  pin_config.data_in_num = this->din_pin_;
+
+  i2s_set_pin(this->parent_->get_port(), &pin_config);
+  this->state_ = microphone::STATE_RUNNING;
+  this->high_freq_.start();
+}
+
+void I2SAudioMicrophone::stop() {
+  if (this->state_ == microphone::STATE_STOPPED)
+    return;
+  this->state_ = microphone::STATE_STOPPING;
+}
+
+void I2SAudioMicrophone::stop_() {
+  i2s_stop(this->parent_->get_port());
+  i2s_driver_uninstall(this->parent_->get_port());
+  this->parent_->unlock();
+  this->state_ = microphone::STATE_STOPPED;
+  this->high_freq_.stop();
+}
+
+void I2SAudioMicrophone::read_() {
+  size_t bytes_read = 0;
+  esp_err_t err =
+      i2s_read(this->parent_->get_port(), this->buffer_.data(), BUFFER_SIZE, &bytes_read, (100 / portTICK_PERIOD_MS));
+  if (err != ESP_OK) {
+    ESP_LOGW(TAG, "Error reading from I2S microphone: %s", esp_err_to_name(err));
+    this->status_set_warning();
+    return;
+  }
+
+  this->status_clear_warning();
+
+  this->data_callbacks_.call(this->buffer_);
+}
+
+void I2SAudioMicrophone::loop() {
+  switch (this->state_) {
+    case microphone::STATE_STOPPED:
+      break;
+    case microphone::STATE_STARTING:
+      this->start_();
+      break;
+    case microphone::STATE_RUNNING:
+      this->read_();
+      break;
+    case microphone::STATE_STOPPING:
+      this->stop_();
+      break;
+  }
+}
+
+}  // namespace i2s_audio
+}  // namespace esphome
+
+#endif  // USE_ESP32

esphome/components/i2s_audio/microphone/i2s_audio_microphone.h

@@ -0,0 +1,37 @@
+#pragma once
+
+#ifdef USE_ESP32
+
+#include "../i2s_audio.h"
+
+#include "esphome/components/microphone/microphone.h"
+#include "esphome/core/component.h"
+
+namespace esphome {
+namespace i2s_audio {
+
+class I2SAudioMicrophone : public I2SAudioIn, public microphone::Microphone, public Component {
+ public:
+  void setup() override;
+  void start() override;
+  void stop() override;
+
+  void loop() override;
+
+  void set_din_pin(uint8_t pin) { this->din_pin_ = pin; }
+
+ protected:
+  void start_();
+  void stop_();
+  void read_();
+
+  uint8_t din_pin_{0};
+  std::vector<uint8_t> buffer_;
+
+  HighFrequencyLoopRequester high_freq_;
+};
+
+}  // namespace i2s_audio
+}  // namespace esphome
+
+#endif  // USE_ESP32