Add ESP32 camera support (#544)

* Add ESP32 Camera

* Add methods

* Lint

* Fixes

* clang-format

* clang-tidy

* Update total_daily_energy.cpp

* Updates

* ESP8266 override pin mode INPUT_PULLUP with INPUT for GPIO16

Fixes esphome/issues#118

* Add servo

* Run clang-format

* Better error message

src/esphome/api/api.proto

@@ -356,3 +356,20 @@ message ExecuteServiceRequest {
   fixed32 key = 1;
   repeated ExecuteServiceArgument args = 2;
 }
+
+message ListEntitiesCameraResponse {
+  string object_id = 1;
+  fixed32 key = 2;
+  string name = 3;
+  string unique_id = 4;
+}
+
+message CameraImageResponse {
+  fixed32 key = 1;
+  bytes data = 2;
+  bool done = 3;
+}
+message CameraImageRequest {
+  bool single = 1;
+  bool stream = 2;
+}

src/esphome/api/api_message.h

@@ -33,6 +33,7 @@ enum class APIMessageType {
   LIST_ENTITIES_SWITCH_RESPONSE = 17,
   LIST_ENTITIES_TEXT_SENSOR_RESPONSE = 18,
   LIST_ENTITIES_SERVICE_RESPONSE = 41,
+  LIST_ENTITIES_CAMERA_RESPONSE = 43,
   LIST_ENTITIES_DONE_RESPONSE = 19,
 
   SUBSCRIBE_STATES_REQUEST = 20,
@@ -43,6 +44,7 @@ enum class APIMessageType {
   SENSOR_STATE_RESPONSE = 25,
   SWITCH_STATE_RESPONSE = 26,
   TEXT_SENSOR_STATE_RESPONSE = 27,
+  CAMERA_IMAGE_RESPONSE = 44,
 
   SUBSCRIBE_LOGS_REQUEST = 28,
   SUBSCRIBE_LOGS_RESPONSE = 29,
@@ -51,6 +53,7 @@ enum class APIMessageType {
   FAN_COMMAND_REQUEST = 31,
   LIGHT_COMMAND_REQUEST = 32,
   SWITCH_COMMAND_REQUEST = 33,
+  CAMERA_IMAGE_REQUEST = 45,
 
   SUBSCRIBE_SERVICE_CALLS_REQUEST = 34,
   SERVICE_CALL_RESPONSE = 35,

src/esphome/api/api_server.cpp

@@ -53,6 +53,16 @@ void APIServer::setup() {
   });
 
   this->last_connected_ = millis();
+
+#ifdef USE_ESP32_CAMERA
+  if (global_esp32_camera != nullptr) {
+    global_esp32_camera->add_image_callback([this](std::shared_ptr<CameraImage> image) {
+      for (auto *c : this->clients_)
+        if (!c->remove_)
+          c->send_camera_state(image);
+    });
+  }
+#endif
 }
 void APIServer::loop() {
   // Partition clients into remove and active
@@ -385,6 +395,7 @@ void APIConnection::read_message_(uint32_t size, uint32_t type, uint8_t *msg) {
     case APIMessageType::LIST_ENTITIES_SWITCH_RESPONSE:
     case APIMessageType::LIST_ENTITIES_TEXT_SENSOR_RESPONSE:
     case APIMessageType::LIST_ENTITIES_SERVICE_RESPONSE:
+    case APIMessageType::LIST_ENTITIES_CAMERA_RESPONSE:
     case APIMessageType::LIST_ENTITIES_DONE_RESPONSE:
       // Invalid
       break;
@@ -401,6 +412,7 @@ void APIConnection::read_message_(uint32_t size, uint32_t type, uint8_t *msg) {
     case APIMessageType::SENSOR_STATE_RESPONSE:
     case APIMessageType::SWITCH_STATE_RESPONSE:
     case APIMessageType::TEXT_SENSOR_STATE_RESPONSE:
+    case APIMessageType::CAMERA_IMAGE_RESPONSE:
       // Invalid
       break;
     case APIMessageType::SUBSCRIBE_LOGS_REQUEST: {
@@ -484,6 +496,14 @@ void APIConnection::read_message_(uint32_t size, uint32_t type, uint8_t *msg) {
       this->on_execute_service_(req);
       break;
     }
+    case APIMessageType::CAMERA_IMAGE_REQUEST: {
+#ifdef USE_ESP32_CAMERA
+      CameraImageRequest req;
+      req.decode(msg, size);
+      this->on_camera_image_request_(req);
+#endif
+      break;
+    }
   }
 }
 void APIConnection::on_hello_request_(const HelloRequest &req) {
@@ -714,6 +734,31 @@ void APIConnection::loop() {
     this->sent_ping_ = true;
     this->send_ping_request();
   }
+
+#ifdef USE_ESP32_CAMERA
+  if (this->image_reader_.available()) {
+    uint32_t space = this->client_->space();
+    // reserve 15 bytes for metadata, and at least 64 bytes of data
+    if (space >= 15 + 64) {
+      uint32_t to_send = std::min(space - 15, this->image_reader_.available());
+      auto buffer = this->get_buffer();
+      // fixed32 key = 1;
+      buffer.encode_fixed32(1, global_esp32_camera->get_object_id_hash());
+      // bytes data = 2;
+      buffer.encode_bytes(2, this->image_reader_.peek_data_buffer(), to_send);
+      // bool done = 3;
+      bool done = this->image_reader_.available() == to_send;
+      buffer.encode_bool(3, done);
+      bool success = this->send_buffer(APIMessageType::CAMERA_IMAGE_RESPONSE);
+      if (success) {
+        this->image_reader_.consume_data(to_send);
+      }
+      if (success && done) {
+        this->image_reader_.return_image();
+      }
+    }
+  }
+#endif
 }
 
 #ifdef USE_BINARY_SENSOR
@@ -1014,6 +1059,31 @@ APIBuffer APIConnection::get_buffer() {
 void APIConnection::send_time_request() { this->send_empty_message(APIMessageType::GET_TIME_REQUEST); }
 #endif
 
+#ifdef USE_ESP32_CAMERA
+void APIConnection::send_camera_state(std::shared_ptr<CameraImage> image) {
+  if (!this->state_subscription_)
+    return;
+  if (this->image_reader_.available())
+    return;
+  this->image_reader_.set_image(image);
+}
+#endif
+
+#ifdef USE_ESP32_CAMERA
+void APIConnection::on_camera_image_request_(const CameraImageRequest &req) {
+  if (global_esp32_camera == nullptr)
+    return;
+
+  ESP_LOGV(TAG, "on_camera_image_request_ stream=%s single=%s", YESNO(req.get_stream()), YESNO(req.get_single()));
+  if (req.get_single()) {
+    global_esp32_camera->request_image();
+  }
+  if (req.get_stream()) {
+    global_esp32_camera->request_stream();
+  }
+}
+#endif
+
 }  // namespace api
 
 ESPHOME_NAMESPACE_END

src/esphome/api/api_server.h

@@ -63,6 +63,9 @@ class APIConnection {
 #endif
 #ifdef USE_TEXT_SENSOR
   bool send_text_sensor_state(text_sensor::TextSensor *text_sensor, std::string state);
+#endif
+#ifdef USE_ESP32_CAMERA
+  void send_camera_state(std::shared_ptr<CameraImage> image);
 #endif
   bool send_log_message(int level, const char *tag, const char *line);
   bool send_disconnect_request(const char *reason);
@@ -111,6 +114,9 @@ class APIConnection {
   void on_subscribe_home_assistant_states_request_(const SubscribeHomeAssistantStatesRequest &req);
   void on_home_assistant_state_response_(const HomeAssistantStateResponse &req);
   void on_execute_service_(const ExecuteServiceRequest &req);
+#ifdef USE_ESP32_CAMERA
+  void on_camera_image_request_(const CameraImageRequest &req);
+#endif
 
   enum class ConnectionState {
     WAITING_FOR_HELLO,
@@ -128,6 +134,9 @@ class APIConnection {
   std::string client_info_;
   ListEntitiesIterator list_entities_iterator_;
   InitialStateIterator initial_state_iterator_;
+#ifdef USE_ESP32_CAMERA
+  CameraImageReader image_reader_;
+#endif
 
   bool state_subscription_{false};
   int log_subscription_{ESPHOME_LOG_LEVEL_NONE};

src/esphome/api/command_messages.cpp

@@ -283,6 +283,26 @@ uint32_t SwitchCommandRequest::get_key() const { return this->key_; }
 bool SwitchCommandRequest::get_state() const { return this->state_; }
 #endif
 
+#ifdef USE_ESP32_CAMERA
+bool CameraImageRequest::get_single() const { return this->single_; }
+bool CameraImageRequest::get_stream() const { return this->stream_; }
+bool CameraImageRequest::decode_varint(uint32_t field_id, uint32_t value) {
+  switch (field_id) {
+    case 1:
+      // bool single = 1;
+      this->single_ = value;
+      return true;
+    case 2:
+      // bool stream = 2;
+      this->stream_ = value;
+      return true;
+    default:
+      return false;
+  }
+}
+APIMessageType CameraImageRequest::message_type() const { return APIMessageType::CAMERA_IMAGE_REQUEST; }
+#endif
+
 }  // namespace api
 
 ESPHOME_NAMESPACE_END

src/esphome/api/command_messages.h

@@ -107,6 +107,20 @@ class SwitchCommandRequest : public APIMessage {
 };
 #endif
 
+#ifdef USE_ESP32_CAMERA
+class CameraImageRequest : public APIMessage {
+ public:
+  bool decode_varint(uint32_t field_id, uint32_t value) override;
+  bool get_single() const;
+  bool get_stream() const;
+  APIMessageType message_type() const override;
+
+ protected:
+  bool single_{false};
+  bool stream_{false};
+};
+#endif
+
 }  // namespace api
 
 ESPHOME_NAMESPACE_END

src/esphome/api/list_entities.cpp

@@ -136,6 +136,16 @@ bool ListEntitiesIterator::on_service(UserServiceDescriptor *service) {
   return this->client_->send_buffer(APIMessageType::LIST_ENTITIES_SERVICE_RESPONSE);
 }
 
+#ifdef USE_ESP32_CAMERA
+bool ListEntitiesIterator::on_camera(ESP32Camera *camera) {
+  auto buffer = this->client_->get_buffer();
+  buffer.encode_nameable(camera);
+  // string unique_id = 4;
+  buffer.encode_string(4, get_default_unique_id("camera", camera));
+  return this->client_->send_buffer(APIMessageType::LIST_ENTITIES_CAMERA_RESPONSE);
+}
+#endif
+
 APIMessageType ListEntitiesRequest::message_type() const { return APIMessageType::LIST_ENTITIES_REQUEST; }
 
 }  // namespace api

src/esphome/api/list_entities.h

@@ -44,6 +44,9 @@ class ListEntitiesIterator : public ComponentIterator {
   bool on_text_sensor(text_sensor::TextSensor *text_sensor) override;
 #endif
   bool on_service(UserServiceDescriptor *service) override;
+#ifdef USE_ESP32_CAMERA
+  bool on_camera(ESP32Camera *camera) override;
+#endif
   bool on_end() override;
 
  protected:

src/esphome/api/util.cpp

@@ -34,6 +34,9 @@ void APIBuffer::encode_bool(uint32_t field, bool value) {
 void APIBuffer::encode_string(uint32_t field, const std::string &value) {
   this->encode_string(field, value.data(), value.size());
 }
+void APIBuffer::encode_bytes(uint32_t field, const uint8_t *data, size_t len) {
+  this->encode_string(field, reinterpret_cast<const char *>(data), len);
+}
 void APIBuffer::encode_string(uint32_t field, const char *string, size_t len) {
   if (len == 0)
     return;
@@ -299,6 +302,20 @@ void ComponentIterator::advance() {
         success = this->on_service(service);
       }
       break;
+#ifdef USE_ESP32_CAMERA
+    case IteratorState::CAMERA:
+      if (global_esp32_camera == nullptr) {
+        advance_platform = true;
+      } else {
+        if (global_esp32_camera->is_internal()) {
+          advance_platform = success = true;
+          break;
+        } else {
+          advance_platform = success = this->on_camera(global_esp32_camera);
+        }
+      }
+      break;
+#endif
     case IteratorState::MAX:
       if (this->on_end()) {
         this->state_ = IteratorState::NONE;
@@ -316,6 +333,9 @@ void ComponentIterator::advance() {
 bool ComponentIterator::on_end() { return true; }
 bool ComponentIterator::on_begin() { return true; }
 bool ComponentIterator::on_service(UserServiceDescriptor *service) { return true; }
+#ifdef USE_ESP32_CAMERA
+bool ComponentIterator::on_camera(ESP32Camera *camera) { return true; }
+#endif
 
 }  // namespace api
 

src/esphome/api/util.h

@@ -8,6 +8,7 @@
 #include "esphome/helpers.h"
 #include "esphome/component.h"
 #include "esphome/controller.h"
+#include "esphome/esp32_camera.h"
 
 ESPHOME_NAMESPACE_BEGIN
 
@@ -26,6 +27,7 @@ class APIBuffer {
   void encode_bool(uint32_t field, bool value);
   void encode_string(uint32_t field, const std::string &value);
   void encode_string(uint32_t field, const char *string, size_t len);
+  void encode_bytes(uint32_t field, const uint8_t *data, size_t len);
   void encode_fixed32(uint32_t field, uint32_t value);
   void encode_float(uint32_t field, float value);
   void encode_nameable(Nameable *nameable);
@@ -78,6 +80,9 @@ class ComponentIterator {
   virtual bool on_text_sensor(text_sensor::TextSensor *text_sensor) = 0;
 #endif
   virtual bool on_service(UserServiceDescriptor *service);
+#ifdef USE_ESP32_CAMERA
+  virtual bool on_camera(ESP32Camera *camera);
+#endif
   virtual bool on_end();
 
  protected:
@@ -106,6 +111,9 @@ class ComponentIterator {
     TEXT_SENSOR,
 #endif
     SERVICE,
+#ifdef USE_ESP32_CAMERA
+    CAMERA,
+#endif
     MAX,
   } state_{IteratorState::NONE};
   size_t at_{0};

src/esphome/application.h

@@ -21,6 +21,7 @@
 #include "esphome/log_component.h"
 #include "esphome/ota_component.h"
 #include "esphome/power_supply_component.h"
+#include "esphome/servo.h"
 #include "esphome/spi_component.h"
 #include "esphome/status_led.h"
 #include "esphome/uart_component.h"

src/esphome/defines.h

@@ -84,6 +84,7 @@
 #define USE_TEMPLATE_COVER
 #ifdef ARDUINO_ARCH_ESP32
 #define USE_ESP32_HALL_SENSOR
+#define USE_ESP32_CAMERA
 #endif
 #define USE_DUTY_CYCLE_SENSOR
 #define USE_STATUS_LED
@@ -138,6 +139,7 @@
 #define USE_MQTT
 #define USE_COPY_OUTPUT
 #define USE_WIFI_INFO_TEXT_SENSOR
+#define USE_SERVO
 #endif
 
 #ifdef USE_REMOTE_RECEIVER