// TFT_eSPI library demo, principally for STM32F processors with DMA:
// https://en.wikipedia.org/wiki/Direct_memory_access

// Tested with ESP32, Nucleo 64 STM32F446RE and Nucleo 144 STM32F767ZI
// TFT's with SPI can use DMA, the sketch also works with 8 bit
// parallel TFT's (tested with ILI9341 and ILI9481)

// The sketch will run on processors without DMA and also parallel
// interface TFT's. Comment out line 29 for no DMA.

// Library here:
// https://github.com/Bodmer/TFT_eSPI

// Adapted by Bodmer 18/12/19 from "RotatingCube" by Daniel Eichhorn.
// See MIT License at end of sketch.

// The rotating cube is drawn into a 128 x 128 Sprite and then this is
// rendered to screen. The Sprite need 32Kbytes of RAM and DMA buffer the same
// so processors with at least >64Kbytes RAM free will be required.

// Define to use DMA for Sprite transfer to SPI TFT, comment out to use no DMA
// (Tested with Nucleo 64 STM32F446RE and Nucleo 144 STM32F767ZI)
// STM32F767 27MHz SPI 50% processor load: Non DMA  52 fps, with DMA 101 fps
// STM32F767 27MHz SPI  0% processor load: Non DMA  97 fps, with DMA 102 fps

// ESP32     27MHz SPI  0% processor load: Non DMA  90 fps, with DMA 101 fps
// ESP32     40MHz SPI  0% processor load: Non DMA 127 fps, with DMA 145 fps
// NOTE: FOR SPI DISPLAYS ONLY
#define USE_DMA_TO_TFT

// Show a processing load does not impact rendering performance
// Processing load is simple algorithm to calculate prime numbers
//#define PRIME_NUMBER_PROCESSOR_LOAD 491 // 241 = 50% CPU load for 128 * 128 and STM32F466 Nucleo 64
                                          // 491 = 50% CPU load for 128 * 128 and STM32F767 Nucleo 144

// Color depth has to be 16 bits if DMA is used to render image
#define COLOR_DEPTH 16

// 128x128 for a 16 bit colour Sprite (32Kbytes RAM)
// Maximum is 181x181 (64Kbytes) for DMA -  restricted by processor design
#define IWIDTH  128
#define IHEIGHT 128

// Size of cube image
// 358 is max for 128x128 sprite, too big and pixel trails are drawn...
#define CUBE_SIZE 358

#include <TFT_eSPI.h>

// Library instance
TFT_eSPI    tft = TFT_eSPI();         // Declare object "tft"

// Create two sprites for a DMA toggle buffer
TFT_eSprite spr[2] = {TFT_eSprite(&tft), TFT_eSprite(&tft) };

// Toggle buffer selection
bool sprSel = 0;

// Pointers to start of Sprites in RAM
uint16_t* sprPtr[2];

// Define the cube face colors
uint16_t palette[] = {TFT_WHITE,  // 1
                      TFT_GREENYELLOW,    // 2
                      TFT_YELLOW, // 3
                      TFT_PINK,  // 4
                      TFT_MAGENTA, // 5
                      TFT_CYAN  // 6
                     };

// Used for fps measuring
uint16_t counter = 0;
long startMillis = millis();
uint16_t interval = 100;

// size / 2 of cube edge
float d = 15;
float px[] = {
  -d,  d,  d, -d, -d,  d,  d, -d
};
float py[] = {
  -d, -d,  d,  d, -d, -d,  d,  d
};
float pz[] = {
  -d, -d, -d, -d,  d,  d,  d,  d
};

// Define the triangles
// The order of the vertices MUST be CCW or the
// shoelace method won't work to detect visible edges
int  faces[12][3] = {
  {0, 1, 4},
  {1, 5, 4},
  {1, 2, 5},
  {2, 6, 5},
  {5, 7, 4},
  {6, 7, 5},
  {3, 4, 7},
  {4, 3, 0},
  {0, 3, 1},
  {1, 3, 2},
  {2, 3, 6},
  {6, 3, 7}
};

// mapped coordinates on screen
float p2x[] = {
  0, 0, 0, 0, 0, 0, 0, 0
};
float p2y[] = {
  0, 0, 0, 0, 0, 0, 0, 0
};

// rotation angle in radians
float r[] = {
  0, 0, 0
};

// Frames per second
String fps = "0fps";

// Sprite draw position
int16_t xpos = 0;
int16_t ypos = 0;

// Prime number initial value
int prime_max = 2;

// 3 axis spin control
bool spinX = true;
bool spinY = true;
bool spinZ = true;

// Min and max of cube edges, "int" type used for compatibility with original sketch min() function
int xmin,ymin,xmax,ymax;

/////////////////////////////////////////////////////////// setup ///////////////////////////////////////////////////
void setup() {

  Serial.begin(115200);

  tft.init();

  tft.fillScreen(TFT_BLACK);

  xpos = 0;
  ypos = (tft.height() - IHEIGHT) / 2;

  // Define cprite colour depth
  spr[0].setColorDepth(COLOR_DEPTH);
  spr[1].setColorDepth(COLOR_DEPTH);

  // Create the 2 sprites
  sprPtr[0] = (uint16_t*)spr[0].createSprite(IWIDTH, IHEIGHT);
  sprPtr[1] = (uint16_t*)spr[1].createSprite(IWIDTH, IHEIGHT);

  // Define text datum and text colour for Sprites
  spr[0].setTextColor(TFT_BLACK);
  spr[0].setTextDatum(MC_DATUM);
  spr[1].setTextColor(TFT_BLACK);
  spr[1].setTextDatum(MC_DATUM);

#ifdef USE_DMA_TO_TFT
  // DMA - should work with ESP32, STM32F2xx/F4xx/F7xx processors
  // NOTE: >>>>>> DMA IS FOR SPI DISPLAYS ONLY <<<<<<
  tft.initDMA(); // Initialise the DMA engine (tested with STM32F446 and STM32F767)
#endif

  // Animation control timer
  startMillis = millis();

}

/////////////////////////////////////////////////////////// loop ///////////////////////////////////////////////////
void loop() {
  uint32_t updateTime = 0;       // time for next update
  bool bounce = false;
  int wait = 0; //random (20);

  // Random movement direction
  int dx = 1; if (random(2)) dx = -1;
  int dy = 1; if (random(2)) dy = -1;

  // Grab exclusive use of the SPI bus
  tft.startWrite();

  // Loop forever
  while (1) {

    // Pull it back onto screen if it wanders off
    if (xpos < -xmin) {
      dx = 1;
      bounce = true;
    }
    if (xpos >= tft.width() - xmax) {
      dx = -1;
      bounce = true;
    }
    if (ypos < -ymin) {
      dy = 1;
      bounce = true;
    }
    if (ypos >= tft.height() - ymax) {
      dy = -1;
      bounce = true;
    }

    if (bounce) {
      // Randomise spin
      if (random(2)) spinX = true;
      else spinX = false;
      if (random(2)) spinY = true;
      else spinY = false;
      if (random(2)) spinZ = true;
      else spinZ = false;
      bounce = false;
      //wait = random (20);
    }

    if (updateTime <= millis())
    {
      // Use time delay so sprtie does not move fast when not all on screen
      updateTime = millis() + wait;
      xmin = IWIDTH / 2; xmax = IWIDTH / 2; ymin = IHEIGHT / 2; ymax = IHEIGHT / 2;
      drawCube();

#ifdef USE_DMA_TO_TFT
      if (tft.dmaBusy()) prime_max++; // Increase processing load until just not busy
      tft.pushImageDMA(xpos, ypos, IWIDTH, IHEIGHT, sprPtr[sprSel]);
      sprSel = !sprSel;
#else
  #ifdef PRIME_NUMBER_PROCESSOR_LOAD
      prime_max = PRIME_NUMBER_PROCESSOR_LOAD;
  #endif
      spr[sprSel].pushSprite(xpos, ypos); // Blocking write (no DMA) 115fps
#endif

      counter++;
      // only calculate the fps every <interval> iterations.
      if (counter % interval == 0) {
        long millisSinceUpdate = millis() - startMillis;
        fps = String((int)(interval * 1000.0 / (millisSinceUpdate))) + " fps";
        Serial.println(fps);
        startMillis = millis();
      }
#ifdef PRIME_NUMBER_PROCESSOR_LOAD
      // Add a processor task
      uint32_t pr = computePrimeNumbers(prime_max);
      Serial.print("Biggest = "); Serial.println(pr);
#endif
      // Change coord for next loop
      xpos += dx;
      ypos += dy;
    }
  } // End of forever loop

  // Release exclusive use of SPI bus ( here as a reminder... forever loop prevents execution)
  tft.endWrite();
}

/**
  Detected visible triangles. If calculated area > 0 the triangle
  is rendered facing towards the viewer, since the vertices are CCW.
  If the area is negative the triangle is CW and thus facing away from us.
*/
int shoelace(int x1, int y1, int x2, int y2, int x3, int y3) {
  // (x1y2 - y1x2) + (x2y3 - y2x3)
  return x1 * y2 - y1 * x2 + x2 * y3 - y2 * x3 + x3 * y1 - y3 * x1;
}

/**
  Rotates and renders the cube.
**/
void drawCube()
{
  double speed = 90;
  if (spinX) r[0] = r[0] + PI / speed; // Add a degree
  if (spinY) r[1] = r[1] + PI / speed; // Add a degree
  if (spinZ) r[2] = r[2] + PI / speed; // Add a degree

  if (r[0] >= 360.0 * PI / 90.0) r[0] = 0;
  if (r[1] >= 360.0 * PI / 90.0) r[1] = 0;
  if (r[2] >= 360.0 * PI / 90.0) r[2] = 0;

  float ax[8] = {0, 0, 0, 0, 0, 0, 0, 0};
  float ay[8] = {0, 0, 0, 0, 0, 0, 0, 0};
  float az[8] = {0, 0, 0, 0, 0, 0, 0, 0};

  // Calculate all vertices of the cube
  for (int i = 0; i < 8; i++)
  {
    float px2 = px[i];
    float py2 = cos(r[0]) * py[i] - sin(r[0]) * pz[i];
    float pz2 = sin(r[0]) * py[i] + cos(r[0]) * pz[i];

    float px3 = cos(r[1]) * px2 + sin(r[1]) * pz2;
    float py3 = py2;
    float pz3 = -sin(r[1]) * px2 + cos(r[1]) * pz2;

    ax[i] = cos(r[2]) * px3 - sin(r[2]) * py3;
    ay[i] = sin(r[2]) * px3 + cos(r[2]) * py3;
    az[i] = pz3 - 150;

    p2x[i] = IWIDTH / 2 + ax[i] * CUBE_SIZE / az[i];
    p2y[i] = IHEIGHT / 2 + ay[i] * CUBE_SIZE / az[i];
  }

  // Fill the buffer with colour 0 (Black)
  spr[sprSel].fillSprite(TFT_BLACK);

  for (int i = 0; i < 12; i++) {

    if (shoelace(p2x[faces[i][0]], p2y[faces[i][0]], p2x[faces[i][1]], p2y[faces[i][1]], p2x[faces[i][2]], p2y[faces[i][2]]) > 0) {
      int x0 = p2x[faces[i][0]];
      int y0 = p2y[faces[i][0]];
      int x1 = p2x[faces[i][1]];
      int y1 = p2y[faces[i][1]];
      int x2 = p2x[faces[i][2]];
      int y2 = p2y[faces[i][2]];

      xmin = min(xmin, x0);
      ymin = min(ymin, y0);
      xmin = min(xmin, x1);
      ymin = min(ymin, y1);
      xmin = min(xmin, x2);
      ymin = min(ymin, y2);
      xmax = max(xmax, x0);
      ymax = max(ymax, y0);
      xmax = max(xmax, x1);
      ymax = max(ymax, y1);
      xmax = max(xmax, x2);
      ymax = max(ymax, y2);

      spr[sprSel].fillTriangle(x0, y0, x1, y1, x2, y2, palette[i / 2]);
      if (i % 2) {
        int avX = 0;
        int avY = 0;
        for (int v = 0; v < 3; v++) {
          avX += p2x[faces[i][v]];
          avY += p2y[faces[i][v]];
        }
        avX = avX / 3;
        avY = avY / 3;
      }
    }
  }

  //spr[sprSel].drawString(fps, IWIDTH / 2, IHEIGHT / 2, 4);
  //delay(100);
}

// This is to provide a processing load to see the improvement DMA gives
uint32_t computePrimeNumbers(int32_t n) {
  if (n<2) return 1;

  int32_t i, fact, j, p = 0;

  //Serial.print("\nPrime Numbers are: \n");
  for (i = 1; i <= n; i++)
  {
    fact = 0;
    for (j = 1; j <= n; j++)
    {
      if (i % j == 0)
        fact++;
    }
    if (fact == 2) {
      p = i;//Serial.print(i); Serial.print(", ");
    }
  }
  //Serial.println();
  return p; // Biggest
}

/*
  Original licence:
  The MIT License (MIT)
  Copyright (c) 2017 by Daniel Eichhorn
  Permission is hereby granted, free of charge, to any person obtaining a copy
  of this software and associated documentation files (the "Software"), to deal
  in the Software without restriction, including without limitation the rights
  to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
  copies of the Software, and to permit persons to whom the Software is
  furnished to do so, subject to the following conditions:
  The above copyright notice and this permission notice shall be included in all
  copies or substantial portions of the Software.
  THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
  AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
  LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
  OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
  SOFTWARE.
*/