obartra/ssim

import { Matrix } from '../types'

/**
 * Generates a matrix based on input `mx` but excluding items based on their modulo and their
 * position in the original matrix.
 *
 * It's a crude implementation of Matlab's `A(1:f:end,1:f:end)` syntax where the first parameter
 * is the matrix, the next one is an array describing the rows to skip [start position, every `f`
 * elements an end position] and the last one follows the same syntax for columns.
 *
 * @example
 * ```
 * img1(1:f:end,1:f:end)
 *
 * ```
 *
 * becomes:
 *
 * ```
 * skip2d(img1, [0, f, img1.length], [0, f, img1[0].length])
 * ```
 *
 * Note that the start index is 0 since, unlike Matlab's, arrays start at 0. Also, unlike in Matlab,
 * `f` must be an integer greater than or equal to 1.
 *
 * @method skip2d
 * @param {Matrix} A - The input matrix
 * @param {Array<number>} - start row, every row, end row
 * @param {Array<number>} - start col, every col, end col
 * @returns {Matrix} B - The downsized matrix
 * @public
 * @memberOf matlab
 * @since 0.0.2
 */
export function skip2d(
  A: Matrix,
  [startRow, everyRow, endRow]: [number, number, number],
  [startCol, everyCol, endCol]: [number, number, number]
): Matrix {
  const width = Math.ceil((endCol - startCol) / everyCol)
  const height = Math.ceil((endRow - startRow) / everyRow)
  const data = new Array(width * height)

  for (let i = 0; i < height; i++) {
    for (let j = 0; j < width; j++) {
      const Ai = startRow + i * everyRow
      const Aj = startCol + j * everyCol

      data[i * width + j] = A.data[Ai * A.width + Aj]
    }
  }

  return {
    data,
    width,
    height,
  }
}