Merge branch 'camera-platform' into integration

CODEOWNERS

@@ -87,6 +87,7 @@ esphome/components/bp1658cj/* @Cossid
 esphome/components/bp5758d/* @Cossid
 esphome/components/button/* @esphome/core
 esphome/components/bytebuffer/* @clydebarrow
+esphome/components/camera/* @DT-art1 @bdraco
 esphome/components/canbus/* @danielschramm @mvturnho
 esphome/components/cap1188/* @mreditor97
 esphome/components/captive_portal/* @OttoWinter

esphome/components/api/api.proto

@@ -836,7 +836,7 @@ message ListEntitiesCameraResponse {
   option (id) = 43;
   option (base_class) = "InfoResponseProtoMessage";
   option (source) = SOURCE_SERVER;
-  option (ifdef) = "USE_ESP32_CAMERA";
+  option (ifdef) = "USE_CAMERA";
 
   string object_id = 1;
   fixed32 key = 2;
@@ -851,7 +851,7 @@ message ListEntitiesCameraResponse {
 message CameraImageResponse {
   option (id) = 44;
   option (source) = SOURCE_SERVER;
-  option (ifdef) = "USE_ESP32_CAMERA";
+  option (ifdef) = "USE_CAMERA";
 
   fixed32 key = 1;
   bytes data = 2;
@@ -860,7 +860,7 @@ message CameraImageResponse {
 message CameraImageRequest {
   option (id) = 45;
   option (source) = SOURCE_CLIENT;
-  option (ifdef) = "USE_ESP32_CAMERA";
+  option (ifdef) = "USE_CAMERA";
   option (no_delay) = true;
 
   bool single = 1;

esphome/components/api/api_connection.cpp

@@ -38,8 +38,8 @@ static constexpr uint16_t PING_RETRY_INTERVAL = 1000;
 static constexpr uint32_t KEEPALIVE_DISCONNECT_TIMEOUT = (KEEPALIVE_TIMEOUT_MS * 5) / 2;
 
 static const char *const TAG = "api.connection";
-#ifdef USE_ESP32_CAMERA
-static const int ESP32_CAMERA_STOP_STREAM = 5000;
+#ifdef USE_CAMERA
+static const int CAMERA_STOP_STREAM = 5000;
 #endif
 
 APIConnection::APIConnection(std::unique_ptr<socket::Socket> sock, APIServer *parent)
@@ -58,6 +58,11 @@ APIConnection::APIConnection(std::unique_ptr<socket::Socket> sock, APIServer *pa
 #else
 #error "No frame helper defined"
 #endif
+#ifdef USE_CAMERA
+  if (camera::Camera::instance() != nullptr) {
+    this->image_reader_ = std::unique_ptr<camera::CameraImageReader>{camera::Camera::instance()->create_image_reader()};
+  }
+#endif
 }
 
 uint32_t APIConnection::get_batch_delay_ms_() const { return this->parent_->get_batch_delay(); }
@@ -180,10 +185,10 @@ void APIConnection::loop() {
     }
   }
 
-#ifdef USE_ESP32_CAMERA
-  if (this->image_reader_.available() && this->helper_->can_write_without_blocking()) {
-    uint32_t to_send = std::min((size_t) MAX_PACKET_SIZE, this->image_reader_.available());
-    bool done = this->image_reader_.available() == to_send;
+#ifdef USE_CAMERA
+  if (this->image_reader_ && this->image_reader_->available() && this->helper_->can_write_without_blocking()) {
+    uint32_t to_send = std::min((size_t) MAX_PACKET_SIZE, this->image_reader_->available());
+    bool done = this->image_reader_->available() == to_send;
     uint32_t msg_size = 0;
     ProtoSize::add_fixed_field<4>(msg_size, 1, true);
     // partial message size calculated manually since its a special case
@@ -193,18 +198,18 @@ void APIConnection::loop() {
 
     auto buffer = this->create_buffer(msg_size);
     // fixed32 key = 1;
-    buffer.encode_fixed32(1, esp32_camera::global_esp32_camera->get_object_id_hash());
+    buffer.encode_fixed32(1, camera::Camera::instance()->get_object_id_hash());
     // bytes data = 2;
-    buffer.encode_bytes(2, this->image_reader_.peek_data_buffer(), to_send);
+    buffer.encode_bytes(2, this->image_reader_->peek_data_buffer(), to_send);
     // bool done = 3;
     buffer.encode_bool(3, done);
 
     bool success = this->send_buffer(buffer, CameraImageResponse::MESSAGE_TYPE);
 
     if (success) {
-      this->image_reader_.consume_data(to_send);
+      this->image_reader_->consume_data(to_send);
       if (done) {
-        this->image_reader_.return_image();
+        this->image_reader_->return_image();
       }
     }
   }
@@ -1112,36 +1117,36 @@ void APIConnection::media_player_command(const MediaPlayerCommandRequest &msg) {
 }
 #endif
 
-#ifdef USE_ESP32_CAMERA
-void APIConnection::set_camera_state(std::shared_ptr<esp32_camera::CameraImage> image) {
+#ifdef USE_CAMERA
+void APIConnection::set_camera_state(std::shared_ptr<camera::CameraImage> image) {
   if (!this->flags_.state_subscription)
     return;
-  if (this->image_reader_.available())
+  if (!this->image_reader_)
     return;
-  if (image->was_requested_by(esphome::esp32_camera::API_REQUESTER) ||
-      image->was_requested_by(esphome::esp32_camera::IDLE))
-    this->image_reader_.set_image(std::move(image));
+  if (this->image_reader_->available())
+    return;
+  if (image->was_requested_by(esphome::camera::API_REQUESTER) || image->was_requested_by(esphome::camera::IDLE))
+    this->image_reader_->set_image(std::move(image));
 }
 uint16_t APIConnection::try_send_camera_info(EntityBase *entity, APIConnection *conn, uint32_t remaining_size,
                                              bool is_single) {
-  auto *camera = static_cast<esp32_camera::ESP32Camera *>(entity);
+  auto *camera = static_cast<camera::Camera *>(entity);
   ListEntitiesCameraResponse msg;
   msg.unique_id = get_default_unique_id("camera", camera);
   fill_entity_info_base(camera, msg);
   return encode_message_to_buffer(msg, ListEntitiesCameraResponse::MESSAGE_TYPE, conn, remaining_size, is_single);
 }
 void APIConnection::camera_image(const CameraImageRequest &msg) {
-  if (esp32_camera::global_esp32_camera == nullptr)
+  if (camera::Camera::instance() == nullptr)
     return;
 
   if (msg.single)
-    esp32_camera::global_esp32_camera->request_image(esphome::esp32_camera::API_REQUESTER);
+    camera::Camera::instance()->request_image(esphome::camera::API_REQUESTER);
   if (msg.stream) {
-    esp32_camera::global_esp32_camera->start_stream(esphome::esp32_camera::API_REQUESTER);
+    camera::Camera::instance()->start_stream(esphome::camera::API_REQUESTER);
 
-    App.scheduler.set_timeout(this->parent_, "api_esp32_camera_stop_stream", ESP32_CAMERA_STOP_STREAM, []() {
-      esp32_camera::global_esp32_camera->stop_stream(esphome::esp32_camera::API_REQUESTER);
-    });
+    App.scheduler.set_timeout(this->parent_, "api_camera_stop_stream", CAMERA_STOP_STREAM,
+                              []() { camera::Camera::instance()->stop_stream(esphome::camera::API_REQUESTER); });
   }
 }
 #endif

esphome/components/api/api_connection.h

@@ -60,8 +60,8 @@ class APIConnection : public APIServerConnection {
 #ifdef USE_TEXT_SENSOR
   bool send_text_sensor_state(text_sensor::TextSensor *text_sensor);
 #endif
-#ifdef USE_ESP32_CAMERA
-  void set_camera_state(std::shared_ptr<esp32_camera::CameraImage> image);
+#ifdef USE_CAMERA
+  void set_camera_state(std::shared_ptr<camera::CameraImage> image);
   void camera_image(const CameraImageRequest &msg) override;
 #endif
 #ifdef USE_CLIMATE
@@ -425,7 +425,7 @@ class APIConnection : public APIServerConnection {
   static uint16_t try_send_update_info(EntityBase *entity, APIConnection *conn, uint32_t remaining_size,
                                        bool is_single);
 #endif
-#ifdef USE_ESP32_CAMERA
+#ifdef USE_CAMERA
   static uint16_t try_send_camera_info(EntityBase *entity, APIConnection *conn, uint32_t remaining_size,
                                        bool is_single);
 #endif
@@ -455,8 +455,8 @@ class APIConnection : public APIServerConnection {
   // These contain vectors/pointers internally, so putting them early ensures good alignment
   InitialStateIterator initial_state_iterator_;
   ListEntitiesIterator list_entities_iterator_;
-#ifdef USE_ESP32_CAMERA
-  esp32_camera::CameraImageReader image_reader_;
+#ifdef USE_CAMERA
+  std::unique_ptr<camera::CameraImageReader> image_reader_;
 #endif
 
   // Group 3: Strings (12 bytes each on 32-bit, 4-byte aligned)

esphome/components/api/api_pb2.cpp

@@ -2216,7 +2216,7 @@ void ExecuteServiceRequest::calculate_size(uint32_t &total_size) const {
   ProtoSize::add_fixed_field<4>(total_size, 1, this->key != 0, false);
   ProtoSize::add_repeated_message(total_size, 1, this->args);
 }
-#ifdef USE_ESP32_CAMERA
+#ifdef USE_CAMERA
 bool ListEntitiesCameraResponse::decode_varint(uint32_t field_id, ProtoVarInt value) {
   switch (field_id) {
     case 5: {

esphome/components/api/api_pb2.h

@@ -1273,7 +1273,7 @@ class ExecuteServiceRequest : public ProtoMessage {
   bool decode_32bit(uint32_t field_id, Proto32Bit value) override;
   bool decode_length(uint32_t field_id, ProtoLengthDelimited value) override;
 };
-#ifdef USE_ESP32_CAMERA
+#ifdef USE_CAMERA
 class ListEntitiesCameraResponse : public InfoResponseProtoMessage {
  public:
   static constexpr uint16_t MESSAGE_TYPE = 43;

esphome/components/api/api_pb2_dump.cpp

@@ -1890,7 +1890,7 @@ void ExecuteServiceRequest::dump_to(std::string &out) const {
   }
   out.append("}");
 }
-#ifdef USE_ESP32_CAMERA
+#ifdef USE_CAMERA
 void ListEntitiesCameraResponse::dump_to(std::string &out) const {
   __attribute__((unused)) char buffer[64];
   out.append("ListEntitiesCameraResponse {\n");

esphome/components/api/api_pb2_service.cpp

@@ -204,7 +204,7 @@ void APIServerConnectionBase::read_message(uint32_t msg_size, uint32_t msg_type,
       this->on_execute_service_request(msg);
       break;
     }
-#ifdef USE_ESP32_CAMERA
+#ifdef USE_CAMERA
     case 45: {
       CameraImageRequest msg;
       msg.decode(msg_data, msg_size);
@@ -682,7 +682,7 @@ void APIServerConnection::on_button_command_request(const ButtonCommandRequest &
   }
 }
 #endif
-#ifdef USE_ESP32_CAMERA
+#ifdef USE_CAMERA
 void APIServerConnection::on_camera_image_request(const CameraImageRequest &msg) {
   if (this->check_authenticated_()) {
     this->camera_image(msg);

esphome/components/api/api_pb2_service.h

@@ -71,7 +71,7 @@ class APIServerConnectionBase : public ProtoService {
 
   virtual void on_execute_service_request(const ExecuteServiceRequest &value){};
 
-#ifdef USE_ESP32_CAMERA
+#ifdef USE_CAMERA
   virtual void on_camera_image_request(const CameraImageRequest &value){};
 #endif
 
@@ -223,7 +223,7 @@ class APIServerConnection : public APIServerConnectionBase {
 #ifdef USE_BUTTON
   virtual void button_command(const ButtonCommandRequest &msg) = 0;
 #endif
-#ifdef USE_ESP32_CAMERA
+#ifdef USE_CAMERA
   virtual void camera_image(const CameraImageRequest &msg) = 0;
 #endif
 #ifdef USE_CLIMATE
@@ -340,7 +340,7 @@ class APIServerConnection : public APIServerConnectionBase {
 #ifdef USE_BUTTON
   void on_button_command_request(const ButtonCommandRequest &msg) override;
 #endif
-#ifdef USE_ESP32_CAMERA
+#ifdef USE_CAMERA
   void on_camera_image_request(const CameraImageRequest &msg) override;
 #endif
 #ifdef USE_CLIMATE

esphome/components/api/api_server.cpp

@@ -111,10 +111,9 @@ void APIServer::setup() {
   }
 #endif
 
-#ifdef USE_ESP32_CAMERA
-  if (esp32_camera::global_esp32_camera != nullptr && !esp32_camera::global_esp32_camera->is_internal()) {
-    esp32_camera::global_esp32_camera->add_image_callback(
-        [this](const std::shared_ptr<esp32_camera::CameraImage> &image) {
+#ifdef USE_CAMERA
+  if (camera::Camera::instance() != nullptr && !camera::Camera::instance()->is_internal()) {
+    camera::Camera::instance()->add_image_callback([this](const std::shared_ptr<camera::CameraImage> &image) {
       for (auto &c : this->clients_) {
         if (!c->flags_.remove)
           c->set_camera_state(image);

esphome/components/api/list_entities.cpp

@@ -40,8 +40,8 @@ LIST_ENTITIES_HANDLER(lock, lock::Lock, ListEntitiesLockResponse)
 #ifdef USE_VALVE
 LIST_ENTITIES_HANDLER(valve, valve::Valve, ListEntitiesValveResponse)
 #endif
-#ifdef USE_ESP32_CAMERA
-LIST_ENTITIES_HANDLER(camera, esp32_camera::ESP32Camera, ListEntitiesCameraResponse)
+#ifdef USE_CAMERA
+LIST_ENTITIES_HANDLER(camera, camera::Camera, ListEntitiesCameraResponse)
 #endif
 #ifdef USE_CLIMATE
 LIST_ENTITIES_HANDLER(climate, climate::Climate, ListEntitiesClimateResponse)

esphome/components/api/list_entities.h

@@ -45,8 +45,8 @@ class ListEntitiesIterator : public ComponentIterator {
   bool on_text_sensor(text_sensor::TextSensor *entity) override;
 #endif
   bool on_service(UserServiceDescriptor *service) override;
-#ifdef USE_ESP32_CAMERA
-  bool on_camera(esp32_camera::ESP32Camera *entity) override;
+#ifdef USE_CAMERA
+  bool on_camera(camera::Camera *entity) override;
 #endif
 #ifdef USE_CLIMATE
   bool on_climate(climate::Climate *entity) override;

esphome/components/camera/__init__.py

@@ -0,0 +1 @@
+CODEOWNERS = ["@DT-art1", "@bdraco"]

esphome/components/camera/camera.cpp

@@ -0,0 +1,22 @@
+#include "camera.h"
+
+namespace esphome {
+namespace camera {
+
+// NOLINTNEXTLINE(cppcoreguidelines-avoid-non-const-global-variables)
+Camera *Camera::global_camera = nullptr;
+
+Camera::Camera() {
+  if (global_camera != nullptr) {
+    this->status_set_error("Multiple cameras are configured, but only one is supported.");
+    this->mark_failed();
+    return;
+  }
+
+  global_camera = this;
+}
+
+Camera *Camera::instance() { return global_camera; }
+
+}  // namespace camera
+}  // namespace esphome

esphome/components/camera/camera.h

@@ -0,0 +1,80 @@
+#pragma once
+
+#include "esphome/core/automation.h"
+#include "esphome/core/component.h"
+#include "esphome/core/entity_base.h"
+#include "esphome/core/helpers.h"
+
+namespace esphome {
+namespace camera {
+
+/** Different sources for filtering.
+ *  IDLE: Camera requests to send an image to the API.
+ *  API_REQUESTER: API requests a new image.
+ *  WEB_REQUESTER: ESP32 web server request an image. Ignored by API.
+ */
+enum CameraRequester : uint8_t { IDLE, API_REQUESTER, WEB_REQUESTER };
+
+/** Abstract camera image base class.
+ *  Encapsulates the JPEG encoded data and it is shared among
+ *  all connected clients.
+ */
+class CameraImage {
+ public:
+  virtual uint8_t *get_data_buffer() = 0;
+  virtual size_t get_data_length() = 0;
+  virtual bool was_requested_by(CameraRequester requester) const = 0;
+  virtual ~CameraImage() {}
+};
+
+/** Abstract image reader base class.
+ *  Keeps track of the data offset of the camera image and
+ *  how many bytes are remaining to read. When the image
+ *  is returned, the shared_ptr is reset and the camera can
+ *  reuse the memory of the camera image.
+ */
+class CameraImageReader {
+ public:
+  virtual void set_image(std::shared_ptr<CameraImage> image) = 0;
+  virtual size_t available() const = 0;
+  virtual uint8_t *peek_data_buffer() = 0;
+  virtual void consume_data(size_t consumed) = 0;
+  virtual void return_image() = 0;
+  virtual ~CameraImageReader() {}
+};
+
+/** Abstract camera base class. Collaborates with API.
+ *  1) API server starts and installs callback (add_image_callback)
+ *     which is called by the camera when a new image is available.
+ *  2) New API client connects and creates a new image reader (create_image_reader).
+ *  3) API connection receives protobuf CameraImageRequest and calls request_image.
+ *  3.a) API connection receives protobuf CameraImageRequest and calls start_stream.
+ *  4) Camera implementation provides JPEG data in the CameraImage and calls callback.
+ *  5) API connection sets the image in the image reader.
+ *  6) API connection consumes data from the image reader and returns the image when finished.
+ *  7.a) Camera captures a new image and continues with 4) until start_stream is called.
+ */
+class Camera : public EntityBase, public Component {
+ public:
+  Camera();
+  // Camera implementation invokes callback to publish a new image.
+  virtual void add_image_callback(std::function<void(std::shared_ptr<CameraImage>)> &&callback) = 0;
+  /// Returns a new camera image reader that keeps track of the JPEG data in the camera image.
+  virtual CameraImageReader *create_image_reader() = 0;
+  // Connection, camera or web server requests one new JPEG image.
+  virtual void request_image(CameraRequester requester) = 0;
+  // Connection, camera or web server requests a stream of images.
+  virtual void start_stream(CameraRequester requester) = 0;
+  // Connection or web server stops the previously started stream.
+  virtual void stop_stream(CameraRequester requester) = 0;
+  virtual ~Camera() {}
+  /// The singleton instance of the camera implementation.
+  static Camera *instance();
+
+ protected:
+  // NOLINTNEXTLINE(cppcoreguidelines-avoid-non-const-global-variables)
+  static Camera *global_camera;
+};
+
+}  // namespace camera
+}  // namespace esphome

esphome/components/esp32_camera/__init__.py

@@ -23,7 +23,7 @@ from esphome.core.entity_helpers import setup_entity
 
 DEPENDENCIES = ["esp32"]
 
-AUTO_LOAD = ["psram"]
+AUTO_LOAD = ["camera", "psram"]
 
 esp32_camera_ns = cg.esphome_ns.namespace("esp32_camera")
 ESP32Camera = esp32_camera_ns.class_("ESP32Camera", cg.PollingComponent, cg.EntityBase)
@@ -283,6 +283,7 @@ SETTERS = {
 
 
 async def to_code(config):
+    cg.add_define("USE_CAMERA")
     var = cg.new_Pvariable(config[CONF_ID])
     await setup_entity(var, config, "camera")
     await cg.register_component(var, config)

esphome/components/esp32_camera/esp32_camera.cpp

@@ -14,8 +14,6 @@ static const char *const TAG = "esp32_camera";
 
 /* ---------------- public API (derivated) ---------------- */
 void ESP32Camera::setup() {
-  global_esp32_camera = this;
-
 #ifdef USE_I2C
   if (this->i2c_bus_ != nullptr) {
     this->config_.sccb_i2c_port = this->i2c_bus_->get_port();
@@ -43,7 +41,7 @@ void ESP32Camera::setup() {
   xTaskCreatePinnedToCore(&ESP32Camera::framebuffer_task,
                           "framebuffer_task",  // name
                           1024,                // stack size
-                          nullptr,             // task pv params
+                          this,                // task pv params
                           1,                   // priority
                           nullptr,             // handle
                           1                    // core
@@ -176,7 +174,7 @@ void ESP32Camera::loop() {
   const uint32_t now = App.get_loop_component_start_time();
   if (this->idle_update_interval_ != 0 && now - this->last_idle_request_ > this->idle_update_interval_) {
     this->last_idle_request_ = now;
-    this->request_image(IDLE);
+    this->request_image(camera::IDLE);
   }
 
   // Check if we should fetch a new image
@@ -202,7 +200,7 @@ void ESP32Camera::loop() {
     xQueueSend(this->framebuffer_return_queue_, &fb, portMAX_DELAY);
     return;
   }
-  this->current_image_ = std::make_shared<CameraImage>(fb, this->single_requesters_ | this->stream_requesters_);
+  this->current_image_ = std::make_shared<ESP32CameraImage>(fb, this->single_requesters_ | this->stream_requesters_);
 
   ESP_LOGD(TAG, "Got Image: len=%u", fb->len);
   this->new_image_callback_.call(this->current_image_);
@@ -225,8 +223,6 @@ ESP32Camera::ESP32Camera() {
   this->config_.fb_count = 1;
   this->config_.grab_mode = CAMERA_GRAB_WHEN_EMPTY;
   this->config_.fb_location = CAMERA_FB_IN_PSRAM;
-
-  global_esp32_camera = this;
 }
 
 /* ---------------- setters ---------------- */
@@ -356,7 +352,7 @@ void ESP32Camera::set_frame_buffer_count(uint8_t fb_count) {
 }
 
 /* ---------------- public API (specific) ---------------- */
-void ESP32Camera::add_image_callback(std::function<void(std::shared_ptr<CameraImage>)> &&callback) {
+void ESP32Camera::add_image_callback(std::function<void(std::shared_ptr<camera::CameraImage>)> &&callback) {
   this->new_image_callback_.add(std::move(callback));
 }
 void ESP32Camera::add_stream_start_callback(std::function<void()> &&callback) {
@@ -365,15 +361,16 @@ void ESP32Camera::add_stream_start_callback(std::function<void()> &&callback) {
 void ESP32Camera::add_stream_stop_callback(std::function<void()> &&callback) {
   this->stream_stop_callback_.add(std::move(callback));
 }
-void ESP32Camera::start_stream(CameraRequester requester) {
+void ESP32Camera::start_stream(camera::CameraRequester requester) {
   this->stream_start_callback_.call();
   this->stream_requesters_ |= (1U << requester);
 }
-void ESP32Camera::stop_stream(CameraRequester requester) {
+void ESP32Camera::stop_stream(camera::CameraRequester requester) {
   this->stream_stop_callback_.call();
   this->stream_requesters_ &= ~(1U << requester);
 }
-void ESP32Camera::request_image(CameraRequester requester) { this->single_requesters_ |= (1U << requester); }
+void ESP32Camera::request_image(camera::CameraRequester requester) { this->single_requesters_ |= (1U << requester); }
+camera::CameraImageReader *ESP32Camera::create_image_reader() { return new ESP32CameraImageReader; }
 void ESP32Camera::update_camera_parameters() {
   sensor_t *s = esp_camera_sensor_get();
   /* update image */
@@ -402,39 +399,39 @@ void ESP32Camera::update_camera_parameters() {
 bool ESP32Camera::has_requested_image_() const { return this->single_requesters_ || this->stream_requesters_; }
 bool ESP32Camera::can_return_image_() const { return this->current_image_.use_count() == 1; }
 void ESP32Camera::framebuffer_task(void *pv) {
+  ESP32Camera *that = (ESP32Camera *) pv;
   while (true) {
     camera_fb_t *framebuffer = esp_camera_fb_get();
-    xQueueSend(global_esp32_camera->framebuffer_get_queue_, &framebuffer, portMAX_DELAY);
+    xQueueSend(that->framebuffer_get_queue_, &framebuffer, portMAX_DELAY);
     // return is no-op for config with 1 fb
-    xQueueReceive(global_esp32_camera->framebuffer_return_queue_, &framebuffer, portMAX_DELAY);
+    xQueueReceive(that->framebuffer_return_queue_, &framebuffer, portMAX_DELAY);
     esp_camera_fb_return(framebuffer);
   }
 }
 
-ESP32Camera *global_esp32_camera;  // NOLINT(cppcoreguidelines-avoid-non-const-global-variables)
-
-/* ---------------- CameraImageReader class ---------------- */
-void CameraImageReader::set_image(std::shared_ptr<CameraImage> image) {
-  this->image_ = std::move(image);
+/* ---------------- ESP32CameraImageReader class ----------- */
+void ESP32CameraImageReader::set_image(std::shared_ptr<camera::CameraImage> image) {
+  this->image_ = std::static_pointer_cast<ESP32CameraImage>(image);
   this->offset_ = 0;
 }
-size_t CameraImageReader::available() const {
+size_t ESP32CameraImageReader::available() const {
   if (!this->image_)
     return 0;
 
   return this->image_->get_data_length() - this->offset_;
 }
-void CameraImageReader::return_image() { this->image_.reset(); }
-void CameraImageReader::consume_data(size_t consumed) { this->offset_ += consumed; }
-uint8_t *CameraImageReader::peek_data_buffer() { return this->image_->get_data_buffer() + this->offset_; }
+void ESP32CameraImageReader::return_image() { this->image_.reset(); }
+void ESP32CameraImageReader::consume_data(size_t consumed) { this->offset_ += consumed; }
+uint8_t *ESP32CameraImageReader::peek_data_buffer() { return this->image_->get_data_buffer() + this->offset_; }
 
-/* ---------------- CameraImage class ---------------- */
-CameraImage::CameraImage(camera_fb_t *buffer, uint8_t requesters) : buffer_(buffer), requesters_(requesters) {}
+/* ---------------- ESP32CameraImage class ----------- */
+ESP32CameraImage::ESP32CameraImage(camera_fb_t *buffer, uint8_t requesters)
+    : buffer_(buffer), requesters_(requesters) {}
 
-camera_fb_t *CameraImage::get_raw_buffer() { return this->buffer_; }
-uint8_t *CameraImage::get_data_buffer() { return this->buffer_->buf; }
-size_t CameraImage::get_data_length() { return this->buffer_->len; }
-bool CameraImage::was_requested_by(CameraRequester requester) const {
+camera_fb_t *ESP32CameraImage::get_raw_buffer() { return this->buffer_; }
+uint8_t *ESP32CameraImage::get_data_buffer() { return this->buffer_->buf; }
+size_t ESP32CameraImage::get_data_length() { return this->buffer_->len; }
+bool ESP32CameraImage::was_requested_by(camera::CameraRequester requester) const {
   return (this->requesters_ & (1 << requester)) != 0;
 }
 

esphome/components/esp32_camera/esp32_camera.h

@@ -7,7 +7,7 @@
 #include <freertos/queue.h>
 #include "esphome/core/automation.h"
 #include "esphome/core/component.h"
-#include "esphome/core/entity_base.h"
+#include "esphome/components/camera/camera.h"
 #include "esphome/core/helpers.h"
 
 #ifdef USE_I2C
@@ -19,9 +19,6 @@ namespace esp32_camera {
 
 class ESP32Camera;
 
-/* ---------------- enum classes ---------------- */
-enum CameraRequester { IDLE, API_REQUESTER, WEB_REQUESTER };
-
 enum ESP32CameraFrameSize {
   ESP32_CAMERA_SIZE_160X120,    // QQVGA
   ESP32_CAMERA_SIZE_176X144,    // QCIF
@@ -77,13 +74,13 @@ enum ESP32SpecialEffect {
 };
 
 /* ---------------- CameraImage class ---------------- */
-class CameraImage {
+class ESP32CameraImage : public camera::CameraImage {
  public:
-  CameraImage(camera_fb_t *buffer, uint8_t requester);
+  ESP32CameraImage(camera_fb_t *buffer, uint8_t requester);
   camera_fb_t *get_raw_buffer();
-  uint8_t *get_data_buffer();
-  size_t get_data_length();
-  bool was_requested_by(CameraRequester requester) const;
+  uint8_t *get_data_buffer() override;
+  size_t get_data_length() override;
+  bool was_requested_by(camera::CameraRequester requester) const override;
 
  protected:
   camera_fb_t *buffer_;
@@ -96,21 +93,21 @@ struct CameraImageData {
 };
 
 /* ---------------- CameraImageReader class ---------------- */
-class CameraImageReader {
+class ESP32CameraImageReader : public camera::CameraImageReader {
  public:
-  void set_image(std::shared_ptr<CameraImage> image);
-  size_t available() const;
-  uint8_t *peek_data_buffer();
-  void consume_data(size_t consumed);
-  void return_image();
+  void set_image(std::shared_ptr<camera::CameraImage> image) override;
+  size_t available() const override;
+  uint8_t *peek_data_buffer() override;
+  void consume_data(size_t consumed) override;
+  void return_image() override;
 
  protected:
-  std::shared_ptr<CameraImage> image_;
+  std::shared_ptr<ESP32CameraImage> image_;
   size_t offset_{0};
 };
 
 /* ---------------- ESP32Camera class ---------------- */
-class ESP32Camera : public EntityBase, public Component {
+class ESP32Camera : public camera::Camera {
  public:
   ESP32Camera();
 
@@ -162,14 +159,15 @@ class ESP32Camera : public EntityBase, public Component {
   void dump_config() override;
   float get_setup_priority() const override;
   /* public API (specific) */
-  void start_stream(CameraRequester requester);
-  void stop_stream(CameraRequester requester);
-  void request_image(CameraRequester requester);
+  void start_stream(camera::CameraRequester requester) override;
+  void stop_stream(camera::CameraRequester requester) override;
+  void request_image(camera::CameraRequester requester) override;
   void update_camera_parameters();
 
-  void add_image_callback(std::function<void(std::shared_ptr<CameraImage>)> &&callback);
+  void add_image_callback(std::function<void(std::shared_ptr<camera::CameraImage>)> &&callback) override;
   void add_stream_start_callback(std::function<void()> &&callback);
   void add_stream_stop_callback(std::function<void()> &&callback);
+  camera::CameraImageReader *create_image_reader() override;
 
  protected:
   /* internal methods */
@@ -206,12 +204,12 @@ class ESP32Camera : public EntityBase, public Component {
   uint32_t idle_update_interval_{15000};
 
   esp_err_t init_error_{ESP_OK};
-  std::shared_ptr<CameraImage> current_image_;
+  std::shared_ptr<ESP32CameraImage> current_image_;
   uint8_t single_requesters_{0};
   uint8_t stream_requesters_{0};
   QueueHandle_t framebuffer_get_queue_;
   QueueHandle_t framebuffer_return_queue_;
-  CallbackManager<void(std::shared_ptr<CameraImage>)> new_image_callback_{};
+  CallbackManager<void(std::shared_ptr<camera::CameraImage>)> new_image_callback_{};
   CallbackManager<void()> stream_start_callback_{};
   CallbackManager<void()> stream_stop_callback_{};
 
@@ -222,13 +220,10 @@ class ESP32Camera : public EntityBase, public Component {
 #endif  // USE_I2C
 };
 
-// NOLINTNEXTLINE(cppcoreguidelines-avoid-non-const-global-variables)
-extern ESP32Camera *global_esp32_camera;
-
 class ESP32CameraImageTrigger : public Trigger<CameraImageData> {
  public:
   explicit ESP32CameraImageTrigger(ESP32Camera *parent) {
-    parent->add_image_callback([this](const std::shared_ptr<esp32_camera::CameraImage> &image) {
+    parent->add_image_callback([this](const std::shared_ptr<camera::CameraImage> &image) {
       CameraImageData camera_image_data{};
       camera_image_data.length = image->get_data_length();
       camera_image_data.data = image->get_data_buffer();

esphome/components/esp32_camera_web_server/__init__.py

@@ -3,7 +3,8 @@ import esphome.config_validation as cv
 from esphome.const import CONF_ID, CONF_MODE, CONF_PORT
 
 CODEOWNERS = ["@ayufan"]
-DEPENDENCIES = ["esp32_camera", "network"]
+AUTO_LOAD = ["camera"]
+DEPENDENCIES = ["network"]
 MULTI_CONF = True
 
 esp32_camera_web_server_ns = cg.esphome_ns.namespace("esp32_camera_web_server")

esphome/components/esp32_camera_web_server/camera_web_server.cpp

@@ -40,7 +40,7 @@ CameraWebServer::CameraWebServer() {}
 CameraWebServer::~CameraWebServer() {}
 
 void CameraWebServer::setup() {
-  if (!esp32_camera::global_esp32_camera || esp32_camera::global_esp32_camera->is_failed()) {
+  if (!camera::Camera::instance() || camera::Camera::instance()->is_failed()) {
     this->mark_failed();
     return;
   }
@@ -67,8 +67,8 @@ void CameraWebServer::setup() {
 
   httpd_register_uri_handler(this->httpd_, &uri);
 
-  esp32_camera::global_esp32_camera->add_image_callback([this](std::shared_ptr<esp32_camera::CameraImage> image) {
-    if (this->running_ && image->was_requested_by(esp32_camera::WEB_REQUESTER)) {
+  camera::Camera::instance()->add_image_callback([this](std::shared_ptr<camera::CameraImage> image) {
+    if (this->running_ && image->was_requested_by(camera::WEB_REQUESTER)) {
       this->image_ = std::move(image);
       xSemaphoreGive(this->semaphore_);
     }
@@ -108,8 +108,8 @@ void CameraWebServer::loop() {
   }
 }
 
-std::shared_ptr<esphome::esp32_camera::CameraImage> CameraWebServer::wait_for_image_() {
-  std::shared_ptr<esphome::esp32_camera::CameraImage> image;
+std::shared_ptr<esphome::camera::CameraImage> CameraWebServer::wait_for_image_() {
+  std::shared_ptr<esphome::camera::CameraImage> image;
   image.swap(this->image_);
 
   if (!image) {
@@ -172,7 +172,7 @@ esp_err_t CameraWebServer::streaming_handler_(struct httpd_req *req) {
   uint32_t last_frame = millis();
   uint32_t frames = 0;
 
-  esp32_camera::global_esp32_camera->start_stream(esphome::esp32_camera::WEB_REQUESTER);
+  camera::Camera::instance()->start_stream(esphome::camera::WEB_REQUESTER);
 
   while (res == ESP_OK && this->running_) {
     auto image = this->wait_for_image_();
@@ -205,7 +205,7 @@ esp_err_t CameraWebServer::streaming_handler_(struct httpd_req *req) {
     res = httpd_send_all(req, STREAM_ERROR, strlen(STREAM_ERROR));
   }
 
-  esp32_camera::global_esp32_camera->stop_stream(esphome::esp32_camera::WEB_REQUESTER);
+  camera::Camera::instance()->stop_stream(esphome::camera::WEB_REQUESTER);
 
   ESP_LOGI(TAG, "STREAM: closed. Frames: %" PRIu32, frames);
 
@@ -215,7 +215,7 @@ esp_err_t CameraWebServer::streaming_handler_(struct httpd_req *req) {
 esp_err_t CameraWebServer::snapshot_handler_(struct httpd_req *req) {
   esp_err_t res = ESP_OK;
 
-  esp32_camera::global_esp32_camera->request_image(esphome::esp32_camera::WEB_REQUESTER);
+  camera::Camera::instance()->request_image(esphome::camera::WEB_REQUESTER);
 
   auto image = this->wait_for_image_();
 

esphome/components/esp32_camera_web_server/camera_web_server.h

@@ -6,7 +6,7 @@
 #include <freertos/FreeRTOS.h>
 #include <freertos/semphr.h>
 
-#include "esphome/components/esp32_camera/esp32_camera.h"
+#include "esphome/components/camera/camera.h"
 #include "esphome/core/component.h"
 #include "esphome/core/helpers.h"
 #include "esphome/core/preferences.h"
@@ -32,7 +32,7 @@ class CameraWebServer : public Component {
   void loop() override;
 
  protected:
-  std::shared_ptr<esphome::esp32_camera::CameraImage> wait_for_image_();
+  std::shared_ptr<camera::CameraImage> wait_for_image_();
   esp_err_t handler_(struct httpd_req *req);
   esp_err_t streaming_handler_(struct httpd_req *req);
   esp_err_t snapshot_handler_(struct httpd_req *req);
@@ -40,7 +40,7 @@ class CameraWebServer : public Component {
   uint16_t port_{0};
   void *httpd_{nullptr};
   SemaphoreHandle_t semaphore_;
-  std::shared_ptr<esphome::esp32_camera::CameraImage> image_;
+  std::shared_ptr<camera::CameraImage> image_;
   bool running_{false};
   Mode mode_{STREAM};
 };

esphome/core/component_iterator.cpp

@@ -158,16 +158,16 @@ void ComponentIterator::advance() {
       }
       break;
 #endif
-#ifdef USE_ESP32_CAMERA
+#ifdef USE_CAMERA
     case IteratorState::CAMERA:
-      if (esp32_camera::global_esp32_camera == nullptr) {
+      if (camera::Camera::instance() == nullptr) {
         advance_platform = true;
       } else {
-        if (esp32_camera::global_esp32_camera->is_internal() && !this->include_internal_) {
+        if (camera::Camera::instance()->is_internal() && !this->include_internal_) {
           advance_platform = success = true;
           break;
         } else {
-          advance_platform = success = this->on_camera(esp32_camera::global_esp32_camera);
+          advance_platform = success = this->on_camera(camera::Camera::instance());
         }
       }
       break;
@@ -386,8 +386,8 @@ bool ComponentIterator::on_begin() { return true; }
 #ifdef USE_API
 bool ComponentIterator::on_service(api::UserServiceDescriptor *service) { return true; }
 #endif
-#ifdef USE_ESP32_CAMERA
-bool ComponentIterator::on_camera(esp32_camera::ESP32Camera *camera) { return true; }
+#ifdef USE_CAMERA
+bool ComponentIterator::on_camera(camera::Camera *camera) { return true; }
 #endif
 #ifdef USE_MEDIA_PLAYER
 bool ComponentIterator::on_media_player(media_player::MediaPlayer *media_player) { return true; }

esphome/core/component_iterator.h

@@ -4,8 +4,8 @@
 #include "esphome/core/controller.h"
 #include "esphome/core/helpers.h"
 
-#ifdef USE_ESP32_CAMERA
-#include "esphome/components/esp32_camera/esp32_camera.h"
+#ifdef USE_CAMERA
+#include "esphome/components/camera/camera.h"
 #endif
 
 namespace esphome {
@@ -48,8 +48,8 @@ class ComponentIterator {
 #ifdef USE_API
   virtual bool on_service(api::UserServiceDescriptor *service);
 #endif
-#ifdef USE_ESP32_CAMERA
-  virtual bool on_camera(esp32_camera::ESP32Camera *camera);
+#ifdef USE_CAMERA
+  virtual bool on_camera(camera::Camera *camera);
 #endif
 #ifdef USE_CLIMATE
   virtual bool on_climate(climate::Climate *climate) = 0;
@@ -125,7 +125,7 @@ class ComponentIterator {
 #ifdef USE_API
     SERVICE,
 #endif
-#ifdef USE_ESP32_CAMERA
+#ifdef USE_CAMERA
     CAMERA,
 #endif
 #ifdef USE_CLIMATE

esphome/core/defines.h

@@ -23,6 +23,7 @@
 #define USE_AREAS
 #define USE_BINARY_SENSOR
 #define USE_BUTTON
+#define USE_CAMERA
 #define USE_CLIMATE
 #define USE_COVER
 #define USE_DATETIME
@@ -144,7 +145,6 @@
 #define USE_ESP32_BLE
 #define USE_ESP32_BLE_CLIENT
 #define USE_ESP32_BLE_SERVER
-#define USE_ESP32_CAMERA
 #define USE_I2C
 #define USE_IMPROV
 #define USE_MICROPHONE

tests/components/camera/common.yaml

@@ -0,0 +1,18 @@
+esphome:
+  includes:
+    - ../../../esphome/components/camera/
+
+script:
+  - id: interface_compile_check
+    then:
+      - lambda: |-
+            using namespace esphome::camera;
+            class MockCamera : public Camera {
+              public:
+                void add_image_callback(std::function<void(std::shared_ptr<CameraImage>)> &&callback) override {}
+                CameraImageReader *create_image_reader() override { return 0; }
+                void request_image(CameraRequester requester) override {}
+                void start_stream(CameraRequester requester) override {}
+                void stop_stream(CameraRequester requester) override {}
+            };
+            MockCamera* camera = new MockCamera();

tests/components/camera/test.esp32-ard.yaml

@@ -0,0 +1 @@
+<<: !include common.yaml

tests/components/camera/test.esp32-idf.yaml

@@ -0,0 +1 @@
+<<: !include common.yaml