Visualize data layers in TypeScript

export interface GeoTiff {
  width: number;
  height: number;
  rasters: Array<number>[];
  bounds: Bounds;
}
solar.ts

/**
 * Renders an RGB GeoTiff image into an HTML canvas.
 *
 * The GeoTiff image must include 3 rasters (bands) which
 * correspond to [Red, Green, Blue] in that order.
 *
 * @param  {GeoTiff} rgb   GeoTiff with RGB values of the image.
 * @param  {GeoTiff} mask  Optional mask for transparency, defaults to opaque.
 * @return {HTMLCanvasElement}  Canvas element with the rendered image.
 */
export function renderRGB(rgb: GeoTiff, mask?: GeoTiff): HTMLCanvasElement {
  // Create an HTML canvas to draw the image.
  // https://www.w3schools.com/tags/canvas_createimagedata.asp
  const canvas = document.createElement('canvas');

  // Set the canvas size to the mask size if it's available,
  // otherwise set it to the RGB data layer size.
  canvas.width = mask ? mask.width : rgb.width;
  canvas.height = mask ? mask.height : rgb.height;

  // Since the mask size can be different than the RGB data layer size,
  // we calculate the "delta" between the RGB layer size and the canvas/mask
  // size. For example, if the RGB layer size is the same as the canvas size,
  // the delta is 1. If the RGB layer size is smaller than the canvas size,
  // the delta would be greater than 1.
  // This is used to translate the index from the canvas to the RGB layer.
  const dw = rgb.width / canvas.width;
  const dh = rgb.height / canvas.height;

  // Get the canvas image data buffer.
  const ctx = canvas.getContext('2d')!;
  const img = ctx.getImageData(0, 0, canvas.width, canvas.height);

  // Fill in every pixel in the canvas with the corresponding RGB layer value.
  // Since Javascript doesn't support multidimensional arrays or tensors,
  // everything is stored in flat arrays and we have to keep track of the
  // indices for each row and column ourselves.
  for (let y = 0; y < canvas.height; y++) {
    for (let x = 0; x < canvas.width; x++) {
      // RGB index keeps track of the RGB layer position.
      // This is multiplied by the deltas since it might be a different
      // size than the image size.
      const rgbIdx = Math.floor(y * dh) * rgb.width + Math.floor(x * dw);
      // Mask index keeps track of the mask layer position.
      const maskIdx = y * canvas.width + x;

      // Image index keeps track of the canvas image position.
      // HTML canvas expects a flat array with consecutive RGBA values.
      // Each value in the image buffer must be between 0 and 255.
      // The Alpha value is the transparency of that pixel,
      // if a mask was not provided, we default to 255 which is opaque.
      const imgIdx = y * canvas.width * 4 + x * 4;
      img.data[imgIdx + 0] = rgb.rasters[0][rgbIdx]; // Red
      img.data[imgIdx + 1] = rgb.rasters[1][rgbIdx]; // Green
      img.data[imgIdx + 2] = rgb.rasters[2][rgbIdx]; // Blue
      img.data[imgIdx + 3] = mask // Alpha
        ? mask.rasters[0][maskIdx] * 255
        : 255;
    }
  }

  // Draw the image data buffer into the canvas context.
  ctx.putImageData(img, 0, 0);
  return canvas;
}
visualize.ts

/**
 * Renders a single value GeoTiff image into an HTML canvas.
 *
 * The GeoTiff image must include 1 raster (band) which contains
 * the values we want to display.
 *
 * @param  {GeoTiff}  data    GeoTiff with the values of interest.
 * @param  {GeoTiff}  mask    Optional mask for transparency, defaults to opaque.
 * @param  {string[]} colors  Hex color palette, defaults to ['000000', 'ffffff'].
 * @param  {number}   min     Minimum value of the data range, defaults to 0.
 * @param  {number}   max     Maximum value of the data range, defaults to 1.
 * @param  {number}   index   Raster index for the data, defaults to 0.
 * @return {HTMLCanvasElement}  Canvas element with the rendered image.
 */
export function renderPalette({
  data,
  mask,
  colors,
  min,
  max,
  index,
}: {
  data: GeoTiff;
  mask?: GeoTiff;
  colors?: string[];
  min?: number;
  max?: number;
  index?: number;
}): HTMLCanvasElement {
  // First create a palette from a list of hex colors.
  const palette = createPalette(colors ?? ['000000', 'ffffff']);
  // Normalize each value of our raster/band of interest into indices,
  // such that they always map into a value within the palette.
  const indices = data.rasters[index ?? 0]
    .map((x) => normalize(x, max ?? 1, min ?? 0))
    .map((x) => Math.round(x * (palette.length - 1)));
  return renderRGB(
    {
      ...data,
      // Map each index into the corresponding RGB values.
      rasters: [
        indices.map((i: number) => palette[i].r),
        indices.map((i: number) => palette[i].g),
        indices.map((i: number) => palette[i].b),
      ],
    },
    mask,
  );
}

/**
 * Creates an {r, g, b} color palette from a hex list of colors.
 *
 * Each {r, g, b} value is a number between 0 and 255.
 * The created palette is always of size 256, regardless of the number of
 * hex colors passed in. Inbetween values are interpolated.
 *
 * @param  {string[]} hexColors  List of hex colors for the palette.
 * @return {{r, g, b}[]}         RGB values for the color palette.
 */
export function createPalette(hexColors: string[]): { r: number; g: number; b: number }[] {
  // Map each hex color into an RGB value.
  const rgb = hexColors.map(colorToRGB);
  // Create a palette with 256 colors derived from our rgb colors.
  const size = 256;
  const step = (rgb.length - 1) / (size - 1);
  return Array(size)
    .fill(0)
    .map((_, i) => {
      // Get the lower and upper indices for each color.
      const index = i * step;
      const lower = Math.floor(index);
      const upper = Math.ceil(index);
      // Interpolate between the colors to get the shades.
      return {
        r: lerp(rgb[lower].r, rgb[upper].r, index - lower),
        g: lerp(rgb[lower].g, rgb[upper].g, index - lower),
        b: lerp(rgb[lower].b, rgb[upper].b, index - lower),
      };
    });
}

/**
 * Convert a hex color into an {r, g, b} color.
 *
 * @param  {string} color  Hex color like 0099FF or #0099FF.
 * @return {{r, g, b}}     RGB values for that color.
 */
export function colorToRGB(color: string): { r: number; g: number; b: number } {
  const hex = color.startsWith('#') ? color.slice(1) : color;
  return {
    r: parseInt(hex.substring(0, 2), 16),
    g: parseInt(hex.substring(2, 4), 16),
    b: parseInt(hex.substring(4, 6), 16),
  };
}

/**
 * Normalizes a number to a given data range.
 *
 * @param  {number} x    Value of interest.
 * @param  {number} max  Maximum value in data range, defaults to 1.
 * @param  {number} min  Minimum value in data range, defaults to 0.
 * @return {number}      Normalized value.
 */
export function normalize(x: number, max: number = 1, min: number = 0): number {
  const y = (x - min) / (max - min);
  return clamp(y, 0, 1);
}

/**
 * Calculates the linear interpolation for a value within a range.
 *
 * @param  {number} x  Lower value in the range, when `t` is 0.
 * @param  {number} y  Upper value in the range, when `t` is 1.
 * @param  {number} t  "Time" between 0 and 1.
 * @return {number}    Inbetween value for that "time".
 */
export function lerp(x: number, y: number, t: number): number {
  return x + t * (y - x);
}

/**
 * Clamps a value to always be within a range.
 *
 * @param  {number} x    Value to clamp.
 * @param  {number} min  Minimum value in the range.
 * @param  {number} max  Maximum value in the range.
 * @return {number}      Clamped value.
 */
export function clamp(x: number, min: number, max: number): number {
  return Math.min(Math.max(x, min), max);
}
visualize.ts