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/microphone/__init__.py

@@ -0,0 +1,91 @@
+from esphome import automation
+import esphome.config_validation as cv
+import esphome.codegen as cg
+
+from esphome.automation import maybe_simple_id
+from esphome.const import CONF_ID, CONF_TRIGGER_ID
+from esphome.core import CORE
+from esphome.coroutine import coroutine_with_priority
+
+
+CODEOWNERS = ["@jesserockz"]
+
+IS_PLATFORM_COMPONENT = True
+
+CONF_ON_DATA = "on_data"
+
+microphone_ns = cg.esphome_ns.namespace("microphone")
+
+Microphone = microphone_ns.class_("Microphone")
+
+CaptureAction = microphone_ns.class_(
+    "CaptureAction", automation.Action, cg.Parented.template(Microphone)
+)
+StopCaptureAction = microphone_ns.class_(
+    "StopCaptureAction", automation.Action, cg.Parented.template(Microphone)
+)
+
+
+DataTrigger = microphone_ns.class_(
+    "DataTrigger",
+    automation.Trigger.template(cg.std_vector.template(cg.int16).operator("ref")),
+)
+
+IsCapturingCondition = microphone_ns.class_(
+    "IsCapturingCondition", automation.Condition
+)
+
+
+async def setup_microphone_core_(var, config):
+    for conf in config.get(CONF_ON_DATA, []):
+        trigger = cg.new_Pvariable(conf[CONF_TRIGGER_ID], var)
+        await automation.build_automation(
+            trigger,
+            [(cg.std_vector.template(cg.uint8).operator("ref").operator("const"), "x")],
+            conf,
+        )
+
+
+async def register_microphone(var, config):
+    if not CORE.has_id(config[CONF_ID]):
+        var = cg.Pvariable(config[CONF_ID], var)
+    await setup_microphone_core_(var, config)
+
+
+MICROPHONE_SCHEMA = cv.Schema(
+    {
+        cv.Optional(CONF_ON_DATA): automation.validate_automation(
+            {
+                cv.GenerateID(CONF_TRIGGER_ID): cv.declare_id(DataTrigger),
+            }
+        ),
+    }
+)
+
+
+MICROPHONE_ACTION_SCHEMA = maybe_simple_id({cv.GenerateID(): cv.use_id(Microphone)})
+
+
+async def media_player_action(config, action_id, template_arg, args):
+    var = cg.new_Pvariable(action_id, template_arg)
+    await cg.register_parented(var, config[CONF_ID])
+    return var
+
+
+automation.register_action(
+    "microphone.capture", CaptureAction, MICROPHONE_ACTION_SCHEMA
+)(media_player_action)
+
+automation.register_action(
+    "microphone.stop_capture", StopCaptureAction, MICROPHONE_ACTION_SCHEMA
+)(media_player_action)
+
+automation.register_condition(
+    "microphone.is_capturing", IsCapturingCondition, MICROPHONE_ACTION_SCHEMA
+)(media_player_action)
+
+
+@coroutine_with_priority(100.0)
+async def to_code(config):
+    cg.add_global(microphone_ns.using)
+    cg.add_define("USE_MICROPHONE")

esphome/components/microphone/automation.h

@@ -0,0 +1,32 @@
+#pragma once
+
+#include "esphome/core/automation.h"
+#include "microphone.h"
+
+#include <vector>
+
+namespace esphome {
+namespace microphone {
+
+template<typename... Ts> class CaptureAction : public Action<Ts...>, public Parented<Microphone> {
+  void play(Ts... x) override { this->parent_->start(); }
+};
+
+template<typename... Ts> class StopCaptureAction : public Action<Ts...>, public Parented<Microphone> {
+  void play(Ts... x) override { this->parent_->stop(); }
+};
+
+class DataTrigger : public Trigger<const std::vector<uint8_t> &> {
+ public:
+  explicit DataTrigger(Microphone *mic) {
+    mic->add_data_callback([this](const std::vector<uint8_t> &data) { this->trigger(data); });
+  }
+};
+
+template<typename... Ts> class IsCapturingActon : public Condition<Ts...>, public Parented<Microphone> {
+ public:
+  bool check(Ts... x) override { return this->parent_->is_running(); }
+};
+
+}  // namespace microphone
+}  // namespace esphome

esphome/components/microphone/microphone.h

@@ -0,0 +1,33 @@
+#pragma once
+
+#include "esphome/core/entity_base.h"
+#include "esphome/core/helpers.h"
+
+namespace esphome {
+namespace microphone {
+
+enum State : uint8_t {
+  STATE_STOPPED = 0,
+  STATE_STARTING,
+  STATE_RUNNING,
+  STATE_STOPPING,
+};
+
+class Microphone {
+ public:
+  virtual void start() = 0;
+  virtual void stop() = 0;
+  void add_data_callback(std::function<void(const std::vector<uint8_t> &)> &&data_callback) {
+    this->data_callbacks_.add(std::move(data_callback));
+  }
+
+  bool is_running() const { return this->state_ == STATE_RUNNING; }
+
+ protected:
+  State state_{STATE_STOPPED};
+
+  CallbackManager<void(const std::vector<uint8_t> &)> data_callbacks_{};
+};
+
+}  // namespace microphone
+}  // namespace esphome