BitmapGrayscaleInput.py

'''**
 * @author EricN
 * February 2, 2009
 * A "short" code segment to open bitmaps and
 * extract the bits as an array of integers. If the array is small (less than 30 x 30)
 * it will print the hex values to the console.
 * The code subsequently saves the array as a 32-bit true color bitmap. The default input file name is test1.bmp
 * the default output name is test2.bmp. You can override these defaults by passing
 * different names as arguments. This file is not meant to be used "as is". You should create your own class
 * and extract what you need from here to populate it.
 *
 * This code has a lot of magic numbers. I suggest you figure out what they are for and make properly named constants for them
 *
 * Rev: 2/18/09 - case 1: for 2 colors was missing
 *                case 2: had 2 not 4 colors.
 *                The mask for 16 colors was 1 and should have been 0x0F.
 *                case 16: for 2^16 colors was not decoding the 5 bit colors properly and did not read the padded bytes. It should work properly now. Not tested.
 *                Updated the comment on biSizeImage and all the image color depths
 *                Decoding for color table images was incorrect. All image types are padded so that the number of bytes read per
 *                   scan line is a multiple of 4. Added the code to read in the "dead bytes" along with updating the comments. Additionally
 *                   the most significant bit, half-nibble or nibble is on the left side of the least significant parts. The ordering was
 *                   reversed which scrambled the images.
 *                256 Color images now works correctly.
 *                16 Color images now works correctly.
 *                4 Color images should work, but is not tested.
 *                2 Color images now works correctly.
 * 
 * Rev: 2/19/09 - The color table was not correctly read when biClrUsed was non-zero. Added one line (and comments) just prior to reading the color table
 *                   to account for this field being non-zero.
 * Rev: 2/20/09 - Added RgbQuad class
 *                Added pelToRGB(), rgbToPel() and colorToGrayscale() to BitmapInput class. These use the new RgbQuad class.
 *                Added peltoRGBQ(), rgbqToPel() (these handle the reserved byte in 32-bit images)
 *                Did NOT implement pelToRGB and rgbToPel in BitmapInput overall.
 * Rev: 2/21/09   The array index values for passing arguments in main() were 1 and 2, should have been 0 and 1 (at least according to Conrad). Not tested.
 * Rev: 11/12/14  Added the topDownDIB flag to deal with negative biHeight values which means image is stored rightside up. All loops depending on the
 *                biHeight value were modified to accommodate both inverted (normal) and top down images. The image is stored in the normal manner
 *                regardless of how it was read in.
 * Rev: 01/10/17  Was using the term 24-bit color when it was 32-bit in the comments. Fixed the documentation to be correct.
 *
 * Classes in the file:
 *  RgbQuad
 *  BitmapInput
 *  
 * Methods in this file:
 *  int     swapInt(int v)
 *  int     swapShort(int v)
 *  RgbQuad pelToRGBQ(int pel)
 *  int     rgbqToPel(int red, int green, int blue, int reserved)
 *  RgbQuad pelToRGB(int pel)
 *  int     rgbToPel(int red, int green, int blue)
 *  int     colorToGrayscale(int pel)
 *  void    main(String[] args)
 *  
 * There is a lot of cutting and pasting from various
 * documents dealing with bitmaps and I have not taken the
 * time to clean up the formatting in the comments. The C syntax is
 * included for reference. The types are declared in windows.h. The C
 * structures and data arrays are predefined static so that they don't
 * ever fall out of scope.
 *
 * I have not "javafied" this file. Much of it needs to be broken out into
 * various specialty methods. These modifications are left as an exercise
 * for the reader.
 *
 * Notes on reading bitmaps:
 *
 * The BMP format assumes an Intel integer type (little endian), however, the Java virtual machine
 * uses the Motorola integer type (big endian), so we have to do a bunch of byte swaps to get things
 * to read and write correctly. Also note that many of the values in a bitmap header are unsigned
 * integers of some kind and Java does not know about unsigned values, except for reading in
 * unsigned byte and unsigned short, but the unsigned int still poses a problem.
 * We don't do any math with the unsigned int values, so we won't see a problem.
 *
 * Bitmaps on disk have the following basic structure
 *  BITMAPFILEHEADER (may be missing if file is not saved properly by the creating application)
 *  BITMAPINFO -
 *        BITMAPINFOHEADER
 *        RGBQUAD - Color Table Array (not present for true color images)
 *  Bitmap Bits in one of many coded formats
 *
 *  The BMP image is stored from bottom to top, meaning that the first scan line in the file is the last scan line in the image.
 *
 *  For ALL images types, each scan line is padded to an even 4-byte boundary.
 *  
 *  For images where there are multiple pels per byte, the left side is the high order element and the right is the
 *  low order element.
 *
 *  in Windows on a 32 bit processor...
 *  DWORD is an unsigned 4 byte integer
 *  WORD is an unsigned 2 byte integer
 *  LONG is a 4 byte signed integer
 *
 *  in Java we have the following sizes:
 *
 * byte
 *   1 signed byte (two's complement). Covers values from -128 to 127.
 *
 * short
 *   2 bytes, signed (two's complement), -32,768 to 32,767
 *
 * int
 *   4 bytes, signed (two's complement). -2,147,483,648 to 2,147,483,647.
 *   Like all numeric types ints may be cast into other numeric types (byte, short, long, float, double).
 *   When lossy casts are done (e.g. int to byte) the conversion is done modulo the length of the smaller type.
 *'''
from final_class import final
import math
import numpy as np
import sys
import argparse
import struct
'''*
 * A member-variable-only class for holding the RGBQUAD C structure elements.
 *'''
@final
class RgbQuad(object):
   def __init__(self):
      self.red = None
      self.green = None
      self.blue = None
      self.reserved = None
class BitmapInput():
   def __init__(self):
      #  BITMAPFILEHEADER  
      self.bmpFileHeader_bfType = None         # WORD
      self.bmpFileHeader_bfSize = None         # DWORD
      self.bmpFileHeader_bfReserved1 = None    # WORD
      self.bmpFileHeader_bfReserved2 = None    # WORD
      self.bmpFileHeader_bfOffBits = None      # DWORD
      # BITMAPINFOHEADER
      self.bmpInfoHeader_biSize = None          # DWORD
      self.bmpInfoHeader_biWidth = None         # LONG
      self.bmpInfoHeader_biHeight = None        # LONG
      self.bmpInfoHeader_biPlanes = None        # WORD
      self.bmpInfoHeader_biBitCount = None      # WORD
      self.bmpInfoHeader_biCompression = None   # DWORD
      self.bmpInfoHeader_biSizeImage = None     # DWORD
      self.bmpInfoHeader_biXPelsPerMeter = None # LONG
      self.bmpInfoHeader_biYPelsPerMeter = None # LONG
      self.bmpInfoHeader_biClrUsed = None       # DWORD
      self.bmpInfoHeader_biClrImportant = None  # DWORD
      # The true color pels
      self.imageArray = None

      # if bmpInfoHeader_biHeight is negative then the image is a top down DIB. This flag is used to
      # identify it as such. Note that when the image is saved, it will be written out in the usual
      # inverted format with a positive bmpInfoHeader_biHeight value.
      self.topDownDIB = False
   '''*
   * Methods to go between little and big endian integer formats.
   *'''
   def swapInt(self, v):
      return  (v >> 24) | (v << 24) | ((v << 8) & 0x00FF0000) | ((v >> 8) & 0x0000FF00)

   def swapShort(self, v):
      return  ((v << 8) & 0xFF00) | ((v >> 8) & 0x00FF)
   '''*
   * Method pelToRGBQ accepts an integer (32 bit) picture element and returns the red, green and blue colors.
   * Unlike pelToRGB, this method also extracts the most significant byte and populates the reserved element of RgbQuad.
   * It returns an RgbQuad object. See rgbqToPel(int red, int green, int blue, int reserved) to go the the other way. 
   *'''
   def pelToRGBQ(self, pel):
      rgbq = RgbQuad()

      rgbq.blue     =  pel        & 0x00FF
      rgbq.green    = (pel >> 8)  & 0x00FF
      rgbq.red      = (pel >> 16) & 0x00FF
      rgbq.reserved = (pel >> 24) & 0x00FF
            
      return rgbq
   '''*
   * The rgbqToPel method takes red, green and blue color values plus an additional byte and returns a single 32-bit integer color.
   * See pelToRGBQ(int pel) to go the other way.
   *'''
   def rgbqToPel(self, red, green, blue, reserved):
      return (reserved << 24) | (red << 16) | (green << 8) | blue

   '''*
   * Method pelToRGB accepts an integer (32 bit) picture element and returns the red, green and blue colors
   * as an RgbQuad object. See rgbToPel(int red, int green, int blue) to go the the other way. 
   *'''
   def pelToRGB(self, pel):
      rgb = RgbQuad()

      rgb.reserved = 0

      rgb.blue  =  pel        & 0x00FF
      rgb.green = (pel >> 8)  & 0x00FF
      rgb.red   = (pel >> 16) & 0x00FF
        
      return rgb
      

   '''*
   * The rgbToPel method takes red, green and blue color values and returns a single 32-bit integer color.
   * See pelToRGB(int pel) to go the other way.
   *'''
   def rgbToPel(self, red, green, blue):
      return (red << 16) | (green << 8) | blue

   '''*
   * Y = 0.3RED+0.59GREEN+0.11Blue
   * The colorToGrayscale method takes a color picture element (pel) and returns the gray scale pel using just one of may possible formulas
   *'''
   def colorToGrayscale(self, pel):
      rgb = self.pelToRGB(pel)
      lum = round(0.3 * float(rgb.red) + 0.589 * float(rgb.green) + 0.11 * float(rgb.blue))
      lum = 256 - lum
      return lum - 1
   '''*
   *
   * ---- MAIN ----
   *
   *'''
   @staticmethod
   def main():
      inFileName, outFileName = None, None
      i, j, k = None, None, None
      numberOfColors = None
      pel = None
      iByteVal, iColumn, iBytesPerRow, iPelsPerRow, iTrailingBits, iDeadBytes = None, None, None, None, None, None
      # RBGQUAD
      rgbQuad_rgbBlue = None
      rgbQuad_rgbGreen = None
      rgbQuad_rgbRed = None
      rgbQuad_rgbReserved = None           # not used in this method

      commandLineArgs = argparse.ArgumentParser()
      commandLineArgs.add_argument("--input_path", type = str, default = "test1.bmp")
      commandLineArgs.add_argument("--output_path", type = str, default = "test1.txt")
      arguments = commandLineArgs.parse_args() 
      # The color table
      colorPallet = np.zeros(256, np.int32)  # reserve space for the largest possible color table

      dibdumper = BitmapInput() # needed to get to the byte swapping methods
      inFileName = arguments.input_path

      outFileName = arguments.output_path

      try: # lots of things can go wrong when doing file i'''o
         # Open the file that is the first command line parameter
         fstream = open(inFileName, 'rb')

         '''*
         *  Read in BITMAPFILEHEADER
         *
         *              typedef struct tagBITMAPFILEHEADER {
                           WORD    bfType
                           DWORD   bfSize
                           WORD    bfReserved1
                           WORD    bfReserved2
                           DWORD   bfOffBits
                     } BITMAPFILEHEADER, FAR *LPBITMAPFILEHEADER, *PBITMAPFILEHEADER

         bfType
            Specifies the file type. It must be set to the signature word BM (0x4D42) to indicate bitmap.
         bfSize
            Specifies the size, in bytes, of the bitmap file.
         bfReserved1
            Reserved set to zero
         bfReserved2
            Reserved set to zero
         bfOffBits
            Specifies the offset, in bytes, from the BITMAPFILEHEADER structure to the bitmap bits
         *'''

         # Read and Convert to big endian
         dibdumper.bmpFileHeader_bfType      = struct.unpack("<H", fstream.read(2))[0]      # WORD
         dibdumper.bmpFileHeader_bfSize      = struct.unpack("<i", fstream.read(4))[0]      # DWORD
         dibdumper.bmpFileHeader_bfReserved1 = struct.unpack("<H", fstream.read(2))[0]      # WORD
         dibdumper.bmpFileHeader_bfReserved2 = struct.unpack("<H", fstream.read(2))[0]      # WORD
         dibdumper.bmpFileHeader_bfOffBits   = struct.unpack("<i", fstream.read(4))[0]      # DWORD

         print(
            "bfType = " + str(dibdumper.bmpFileHeader_bfType) + "\n" +
            "bfSize = " + str(dibdumper.bmpFileHeader_bfSize) + "\n" +
            "bfReserved1 = " + str(dibdumper.bmpFileHeader_bfReserved1) + "\n" +
            "bfReserved2 = " + str(dibdumper.bmpFileHeader_bfReserved2) + "\n" +
            "bfOffBits = " + str(dibdumper.bmpFileHeader_bfOffBits) + "\n"
         )

         '''*
         Read in BITMAPINFOHEADER

                        typedef struct tagBITMAPINFOHEADER{
                              DWORD      biSize
                              LONG       biWidth
                              LONG       biHeight
                              WORD       biPlanes
                              WORD       biBitCount
                              DWORD      biCompression
                              DWORD      biSizeImage
                              LONG       biXPelsPerMeter
                              LONG       biYPelsPerMeter
                              DWORD      biClrUsed
                              DWORD      biClrImportant
                        } BITMAPINFOHEADER, FAR *LPBITMAPINFOHEADER, *PBITMAPINFOHEADER


         biSize
            Specifies the size of the structure, in bytes.
            This size does not include the color table or the masks mentioned in the biClrUsed member.
            See the Remarks section for more information.
         biWidth
            Specifies the width of the bitmap, in pixels.
         biHeight
            Specifies the height of the bitmap, in pixels.
            If biHeight is positive, the bitmap is a bottom-up DIB and its origin is the lower left corner.
            If biHeight is negative, the bitmap is a top-down DIB and its origin is the upper left corner.
            If biHeight is negative, indicating a top-down DIB, biCompression must be either BI_RGB or BI_BITFIELDS. Top-down DIBs cannot be compressed.
         biPlanes
            Specifies the number of planes for the target device.
            This value must be set to 1.
         biBitCount
            Specifies the number of bits per pixel.
            The biBitCount member of the BITMAPINFOHEADER structure determines the number of bits that define each pixel and the maximum number of colors in the bitmap.
            This member must be one of the following values.
            Value     Description
            1       The bitmap is monochrome, and the bmiColors member contains two entries.
                     Each bit in the bitmap array represents a pixel. The most significant bit is to the left in the image. 
                     If the bit is clear, the pixel is displayed with the color of the first entry in the bmiColors table.
                     If the bit is set, the pixel has the color of the second entry in the table.
            2       The bitmap has four possible color values.  The most significant half-nibble is to the left in the image.
            4       The bitmap has a maximum of 16 colors, and the bmiColors member contains up to 16 entries.
                     Each pixel in the bitmap is represented by a 4-bit index into the color table. The most significant nibble is to the left in the image.
                     For example, if the first byte in the bitmap is 0x1F, the byte represents two pixels. The first pixel contains the color in the second table entry, and the second pixel contains the color in the sixteenth table entry.
            8       The bitmap has a maximum of 256 colors, and the bmiColors member contains up to 256 entries. In this case, each byte in the array represents a single pixel.
            16      The bitmap has a maximum of 2^16 colors.
                     If the biCompression member of the BITMAPINFOHEADER is BI_RGB, the bmiColors member is NULL.
                     Each WORD in the bitmap array represents a single pixel. The relative intensities of red, green, and blue are represented with 5 bits for each color component.
                     The value for blue is in the least significant 5 bits, followed by 5 bits each for green and red.
                     The most significant bit is not used. The bmiColors color table is used for optimizing colors used on palette-based devices, and must contain the number of entries specified by the biClrUsed member of the BITMAPINFOHEADER.
            24      The bitmap has a maximum of 2^24 colors, and the bmiColors member is NULL.
                     Each 3-byte triplet in the bitmap array represents the relative intensities of blue, green, and red, respectively, for a pixel.
                     The bmiColors color table is used for optimizing colors used on palette-based devices, and must contain the number of entries specified by the biClrUsed member of the BITMAPINFOHEADER.
            32      The bitmap has a maximum of 2^32 colors. If the biCompression member of the BITMAPINFOHEADER is BI_RGB, the bmiColors member is NULL. Each DWORD in the bitmap array represents the relative intensities of blue, green, and red, respectively, for a pixel. The high byte in each DWORD is not used. The bmiColors color table is
                     used for optimizing colors used on palette-based devices, and must contain the number of entries specified by the biClrUsed member of the BITMAPINFOHEADER.
                     If the biCompression member of the BITMAPINFOHEADER is BI_BITFIELDS, the bmiColors member contains three DWORD color masks that specify the red, green, and blue components, respectively, of each pixel.
                     Each DWORD in the bitmap array represents a single pixel.
         biCompression
            Specifies the type of compression for a compressed bottom-up bitmap (top-down DIBs cannot be compressed). This member can be one of the following values.
            Value               Description
            BI_RGB              An uncompressed format.
            BI_BITFIELDS        Specifies that the bitmap is not compressed and that the color table consists of three DWORD color masks that specify the red, green, and blue components of each pixel.
                                 This is valid when used with 16- and 32-bpp bitmaps.
                                 This value is valid in Windows Embedded CE versions 2.0 and later.
            BI_ALPHABITFIELDS   Specifies that the bitmap is not compressed and that the color table consists of four DWORD color masks that specify the red, green, blue, and alpha components of each pixel.
                                 This is valid when used with 16- and 32-bpp bitmaps.
                                 This value is valid in Windows CE .NET 4.0 and later.
                                 You can OR any of the values in the above table with BI_SRCPREROTATE to specify that the source DIB section has the same rotation angle as the destination.
         biSizeImage
            Specifies the size, in bytes, of the image. This value will be the number of bytes in each scan line which must be padded to
            insure the line is a multiple of 4 bytes (it must align on a DWORD boundary) times the number of rows.
            This value may be set to zero for BI_RGB bitmaps (so you cannot be sure it will be set).
         biXPelsPerMeter
            Specifies the horizontal resolution, in pixels per meter, of the target device for the bitmap.
            An application can use this value to select a bitmap from a resource group that best matches the characteristics of the current device.
         biYPelsPerMeter
            Specifies the vertical resolution, in pixels per meter, of the target device for the bitmap
         biClrUsed
            Specifies the number of color indexes in the color table that are actually used by the bitmap.
            If this value is zero, the bitmap uses the maximum number of colors corresponding to the value of the biBitCount member for the compression mode specified by biCompression.
            If biClrUsed is nonzero and the biBitCount member is less than 16, the biClrUsed member specifies the actual number of colors the graphics engine or device driver accesses.
            If biBitCount is 16 or greater, the biClrUsed member specifies the size of the color table used to optimize performance of the system color palettes.
            If biBitCount equals 16 or 32, the optimal color palette starts immediately following the three DWORD masks.
            If the bitmap is a packed bitmap (a bitmap in which the bitmap array immediately follows the BITMAPINFO header and is referenced by a single pointer), the biClrUsed member must be either zero or the actual size of the color table.
         biClrImportant
            Specifies the number of color indexes required for displaying the bitmap.
            If this value is zero, all colors are required.
         Remarks

         The BITMAPINFO structure combines the BITMAPINFOHEADER structure and a color table to provide a complete definition of the dimensions and colors of a DIB.
         An application should use the information stored in the biSize member to locate the color table in a BITMAPINFO structure, as follows.

         pColor = ((LPSTR)pBitmapInfo + (WORD)(pBitmapInfo->bmiHeader.biSize))
         *'''

         # Read and convert to big endian
         dibdumper.bmpInfoHeader_biSize           = struct.unpack("<i", fstream.read(4))[0]                                   # DWORD
         dibdumper.bmpInfoHeader_biWidth          = struct.unpack("<i", fstream.read(4))[0]                                   # LONG
         dibdumper.bmpInfoHeader_biHeight         = struct.unpack("<i", fstream.read(4))[0]                                   # LONG
         dibdumper.bmpInfoHeader_biPlanes         = struct.unpack("<H", fstream.read(2))[0]                                   # WORD
         dibdumper.bmpInfoHeader_biBitCount       = struct.unpack("<H", fstream.read(2))[0]                                   # WORD
         dibdumper.bmpInfoHeader_biCompression    = struct.unpack("<i", fstream.read(4))[0]                                   # DWORD
         dibdumper.bmpInfoHeader_biSizeImage      = struct.unpack("<i", fstream.read(4))[0]                                   # DWORD
         dibdumper.bmpInfoHeader_biXPelsPerMeter  = struct.unpack("<i", fstream.read(4))[0]                                   # LONG
         dibdumper.bmpInfoHeader_biYPelsPerMeter  = struct.unpack("<i", fstream.read(4))[0]                                   # LONG
         dibdumper.bmpInfoHeader_biClrUsed        = struct.unpack("<i", fstream.read(4))[0]                                   # DWORD
         dibdumper.bmpInfoHeader_biClrImportant   = struct.unpack("<i", fstream.read(4))[0]                                   # DWORD

         print(
             "biSize = " + str(dibdumper.bmpInfoHeader_biSize) + "\n" +
             "biWidth = " + str(dibdumper.bmpInfoHeader_biWidth) + "\n" +
             "biHeight = " + str(dibdumper.bmpInfoHeader_biHeight) + "\n" +
             "biPlanes = " + str(dibdumper.bmpInfoHeader_biPlanes) + "\n" +
             "biBitCount = " + str(dibdumper.bmpInfoHeader_biBitCount) + "\n" +
             "biCompression = " + str(dibdumper.bmpInfoHeader_biCompression) + "\n" +
             "biSizeImage = " + str(dibdumper.bmpInfoHeader_biSizeImage) + "\n" +
             "biXPelsPerMeter = " + str(dibdumper.bmpInfoHeader_biXPelsPerMeter) + "\n" +
             "biYPelsPerMeter = " + str(dibdumper.bmpInfoHeader_biYPelsPerMeter) + "\n" + 
             "biClrUsed = " + str(dibdumper.bmpInfoHeader_biClrUsed) + "\n" +
             "biClrImportant = " + str(dibdumper.bmpInfoHeader_biClrImportant) + "\n"
         )

         # Since we use the height to crate arrays, it cannot have a negative a value. If the height field is
         # less than zero, then make it positive and set the topDownDIB flag to TRUE so we know that the image is
         # stored on disc upsidedown (which means it is actually rightside up).
         if dibdumper.bmpInfoHeader_biHeight < 0:
            dibdumper.topDownDIB = True
            dibdumper.bmpInfoHeader_biHeight = -dibdumper.bmpInfoHeader_biHeight
         '''*
         Now for the color table. For true color images, there isn't one.

         typedef struct tagRGBQUAD {
               BYTE    rgbBlue
               BYTE    rgbGreen
               BYTE    rgbRed
               BYTE    rgbReserved
               } RGBQUAD

         typedef RGBQUAD FAR* LPRGBQUAD
         *'''
         if dibdumper.bmpInfoHeader_biBitCount == 1:
            numberOfColors = 2
         elif dibdumper.bmpInfoHeader_biBitCount == 2:
            numberOfColors = 4
         elif dibdumper.bmpInfoHeader_biBitCount == 4:
            numberOfColors = 16
         elif dibdumper.bmpInfoHeader_biBitCount == 8:
            numberOfColors = 256
         else:
            numberOfColors = 0

         print("Color Depth = " + str(dibdumper.bmpInfoHeader_biBitCount) + "," +
            str(numberOfColors)
         )
         '''*
         * biClrUsed -  Specifies the number of color indexes in the color table that are actually used by the bitmap.
         *     If this value is zero, the bitmap uses the maximum number of colors corresponding to the value of the biBitCount member for the compression mode specified by biCompression.
         *     If biClrUsed is nonzero and the biBitCount member is less than 16, the biClrUsed member specifies the actual number of colors the graphics engine or device driver accesses.
         *     If biBitCount is 16 or greater, the biClrUsed member specifies the size of the color table used to optimize performance of the system color palettes.
         *     If biBitCount equals 16 or 32, the optimal color palette starts immediately following the three DWORD masks.
         *     If the bitmap is a packed bitmap (a bitmap in which the bitmap array immediately follows the BITMAPINFO header and is referenced by a single pointer), the biClrUsed member must be either zero or the actual size of the color table.
         *'''
         if dibdumper.bmpInfoHeader_biClrUsed > 0:
            numberOfColors = dibdumper.bmpInfoHeader_biClrUsed
               
         for i in range(numberOfColors): # Read in the color table (or not if numberOfColors is zero)
            rgbQuad_rgbBlue      = struct.unpack(">B", fstream.read(1))[0] # lowest byte in the color
            rgbQuad_rgbGreen     = struct.unpack(">B", fstream.read(1))[0]
            rgbQuad_rgbRed       = struct.unpack(">B", fstream.read(1))[0] # highest byte in the color
            rgbQuad_rgbReserved  = struct.unpack(">B", fstream.read(1))[0]

            # Build the color from the RGB values. Since we declared the rgbQuad values to be int, we can shift and then OR the values
            # to build up the color. Since we are reading one byte at a time, there are no "endian" issues.

            colorPallet[i] = (rgbQuad_rgbRed << 16) | (rgbQuad_rgbGreen << 8) | rgbQuad_rgbBlue
            # System.out.printf("DEBUG: Color Table = %d, %06X\n", i, colorPallet[i])
            # for (i = 0 i < numberOfColors ++i)

            '''*
            * Now for the fun part. We need to read in the rest of the bit map, but how we interpret the values depends on the color depth.
            *
            * numberOfColors = 2:   Each bit is a pel, so there are 8 pels per byte. The Color Table has only two values for "black" and "white"
            * numberOfColors = 4:   Each pair of bits is a pel, so there are 4 pels per byte. The Color Table has only four values
            * numberOfColors = 16  Each nibble (4 bits) is a pel, so there are 2 pels per byte. The Color Table has 16 entries.
            * numberOfColors = 256 Each byte is a pel and the value maps into the 256 byte Color Table.
            *
            * Any other value is read in as "true" color.
            *
            * The BMP image is stored from bottom to top, meaning that the first scan line is the last scan line in the image.
            *
            * The rest is the bitmap. Use the height and width information to read it in. And as I mentioned before....
            * In the 32-bit format, each pixel in the image is represented by a series of four bytes of RGB stored as xBRG,
            * where the 'x' is an unused byte. For ALL image types each scan line is padded to an even 4-byte boundary.
            *
            *'''
         dibdumper.imageArray = np.zeros((dibdumper.bmpInfoHeader_biHeight, dibdumper.bmpInfoHeader_biWidth), np.int64) # Create the array for the pels
         '''*
         * I use the same loop structure for each case for clarity so you can see the similarities and differences.
         * The outer loop is over the rows (in reverse), the inner loop over the columns. 
         *'''
         if dibdumper.bmpInfoHeader_biBitCount == 1: # each bit is a color, so there are 8 pels per byte.  Works
            '''*
            * Each byte read in is 8 columns, so we need to break them out. We also have to deal with the case
            * where the image width is not an integer multiple of 8, in which case we will
            * have bits from part of the remaining byte. Each color is 1 bit which is masked with 0x01.
            * The screen ordering of the pels is High-Bit to Low-Bit, so the most significant element is first in the array of pels.
            *'''
            iBytesPerRow = dibdumper.bmpInfoHeader_biWidth // 8
            iTrailingBits = dibdumper.bmpInfoHeader_biWidth % 8

            iDeadBytes = iBytesPerRow
            if iTrailingBits > 0:
               iDeadBytes += 1
               
            iDeadBytes = (4 - iDeadBytes % 4) % 4

            for row in range(dibdumper.bmpInfoHeader_biHeight): # read over the rows
               if dibdumper.topDownDIB:
                  i = row 
               else:
                  i = dibdumper.bmpInfoHeader_biHeight - 1 - row

               for j in range(iBytesPerRow):
                  iByteVal = struct.unpack(">B", fstream.read(1))[0]

                  for k in range(8):     # Get 8 pels from the one byte
                     iColumn = j * 8 + k
                     pel = colorPallet[(iByteVal >> (7 - k)) & 0x01]
                     dibdumper.imageArray[i][iColumn] = pel
               if iTrailingBits > 0: # pick up the trailing bits for images that are not mod 8 columns wide
                  iByteVal = struct.unpack(">B", fstream.read(1))[0]

                  for k in range(iTrailingBits):
                     iColumn = iBytesPerRow * 8 + k
                     pel = colorPallet[(iByteVal >> (7 - k)) & 0x01]
                     dibdumper.imageArray[i][iColumn] = pel
               for j in range(iDeadBytes):
                  struct.unpack(">B", fstream.read(1))[0] # Now read in the "dead bytes" to pad to a 4 byte boundary

         elif dibdumper.bmpInfoHeader_biBitCount == 2: # 4 colors, Each byte is 4 pels (2 bits each),  Should work, not tested.
            '''*
            * Each byte read in is 4 columns, so we need to break them out. We also have to deal with the case
            * where the image width is not an integer multiple of 4, in which case we will
            * have from 2 to 6 bits of the remaining byte. Each color is 2 bits which is masked with 0x03.
            * The screen ordering of the pels is High-Half-Nibble to Low-Half-Nibble, so the most significant element is first in the array of pels.
            *'''
            iBytesPerRow = dibdumper.bmpInfoHeader_biWidth // 4
            iTrailingBits = dibdumper.bmpInfoHeader_biWidth % 4 # 0, 1, 2 or 3

            iDeadBytes = iBytesPerRow
            if iTrailingBits > 0:
               iDeadBytes += 1
            iDeadBytes = (4 - iDeadBytes % 4) % 4

            for row in range(dibdumper.bmpInfoHeader_biHeight): # read over the rows
               if dibdumper.topDownDIB:
                  i = row
               else:
                  i = dibdumper.bmpInfoHeader_biHeight - 1 - row

               for j in range(iBytesPerRow):
                  iByteVal = struct.unpack(">B", fstream.read(1))[0]

                  for k in range(4): # Get 4 pels from one byte
                     iColumn = j * 4 + k
                     pel = colorPallet[(iByteVal >> ((3 - k) * 2)) & 0x03] # shift 2 bits at a time and reverse order
                     dibdumper.imageArray[i][iColumn] = pel

               if iTrailingBits > 0: # pick up the trailing nibble for images that are not mod 2 columns wide
                  iByteVal = struct.unpack(">B", fstream.read(1))[0]

                  for k in range(iTrailingBits):
                     iColumn = iBytesPerRow * 4 + k  
                     pel = colorPallet[(iByteVal >> ((3 - k) * 2)) & 0x03]
                     dibdumper.imageArray[i][iColumn] = pel

               for j in range(iDeadBytes):
                  struct.unpack(">B", fstream.read(1))[0] # Now read in the "dead bytes" to pad to a 4 byte boundary

         elif dibdumper.bmpInfoHeader_biBitCount == 4: # 16 colors, Each byte is two pels. Works
            '''*
            * Each byte read in is 2 columns, so we need to break them out. We also have to deal with the case
            * where the image width is not an integer multiple of 2, in which case we will
            * have one nibble from part of the remaining byte. We then read in the dead bytes so that each
            * scan line is a multiple of 4 bytes. Each color is a nibble (4 bits) which is masked with 0x0F.
            * The screen ordering of the pels is High-Nibble Low-Nibble, so the most significant element is first in the array of pels.
            *'''
            iPelsPerRow   = dibdumper.bmpInfoHeader_biWidth
            iBytesPerRow  = iPelsPerRow // 2
            iTrailingBits = iPelsPerRow % 2  # Will either be 0 or 1

            iDeadBytes = iBytesPerRow
            if iTrailingBits > 0:
                  iDeadBytes += 1
            iDeadBytes = (4 - iDeadBytes % 4) % 4

            for row in range(dibdumper.bmpInfoHeader_biHeight): # read over the rows
               if dibdumper.topDownDIB:
                  i = row
               else:
                  i = dibdumper.bmpInfoHeader_biHeight - 1 - row

               for j in range(iBytesPerRow):
                  iByteVal = struct.unpack(">B", fstream.read(1))[0]

                  for k in range(2): # Two pels per byte
                     iColumn = j * 2 + k          # 1 - k  is needed to have High, Low nibble ordering for the image.
                     pel = colorPallet[(iByteVal >> ((1 - k) * 4)) & 0x0F] # shift 4 bits at a time
                     dibdumper.imageArray[i][iColumn] = pel

               if iTrailingBits > 0: # pick up the trailing nibble for images that are not mod 2 columns wide
                  iByteVal = struct.unpack(">B", fstream.read(1))[0]

                  iColumn = iBytesPerRow * 2
                  pel = colorPallet[(iByteVal >> 4) & 0x0F] # The High nibble is the last remaining pel
                  dibdumper.imageArray[i][iColumn] = pel

               for j in range(iDeadBytes):
                  struct.unpack(">B", fstream.read(1))[0] # Now read in the "dead bytes" to pad to a 4 byte boundary
                  # for (i = bmpInfoHeader_biHeight - 1 i >= 0 --i)
         
         elif dibdumper.bmpInfoHeader_biBitCount == 8: # 1 byte, 1 pel, Works
            '''*
            * Each byte read in is 1 column. We then read in the dead bytes so that each scan line is a multiple of 4 bytes.
            *'''
            iPelsPerRow = dibdumper.bmpInfoHeader_biWidth
            iDeadBytes = (4 - iPelsPerRow % 4) % 4
            for row in range(dibdumper.bmpInfoHeader_biHeight): # read over the rows
               if (dibdumper.topDownDIB):
                  i = row 
               else:
                  i = dibdumper.bmpInfoHeader_biHeight - 1 - row

               for j in range(iPelsPerRow):      # j is now just the column counter
                  iByteVal = struct.unpack(">B", fstream.read(1))[0]
                  pel = colorPallet[iByteVal]
                  dibdumper.imageArray[i][j] = pel

               for j in range (iDeadBytes):
                  struct.unpack(">B", fstream.read(1))[0] # Now read in the "dead bytes" to pad to a 4 byte boundary
         
         elif dibdumper.bmpInfoHeader_biBitCount == 16: # Not likely to work (format is not internally consistent), not tested.
            '''*
            * Each two bytes read in is 1 column. Each color is 5 bits in the 2 byte word value, so we shift 5 bits and then mask them
            * off with 0x1F which is %11111 in binary. We then read in the dead bytes so that each scan line is a multiple of 4 bytes.
            *'''
            iPelsPerRow = dibdumper.bmpInfoHeader_biWidth
            iDeadBytes = (4 - iPelsPerRow % 4) % 4
            for row in range(dibdumper.bmpInfoHeader_biHeight): # read over the rows
               if (dibdumper.topDownDIB):
                  i = row
               else:
                  i = dibdumper.bmpInfoHeader_biHeight - 1 - row

               for j in range(iPelsPerRow):   # j is now just the column counter
                  pel = struct.unpack("<H", fstream.read(2))[0] # Need to deal with little endian values
                  rgbQuad_rgbBlue      =  pel        & 0x1F
                  rgbQuad_rgbGreen     = (pel >> 5)  & 0x1F   
                  rgbQuad_rgbRed       = (pel >> 10) & 0x1F
                  pel = (rgbQuad_rgbRed << 16) | (rgbQuad_rgbGreen << 8) | rgbQuad_rgbBlue
                  dibdumper.imageArray[i][j] = pel

               for j in range(iDeadBytes):
                  struct.unpack(">B", fstream.read(1))[0] # Now read in the "dead bytes" to pad to a 4 byte boundary
               # for (i = bmpInfoHeader_biHeight - 1 i >= 0 --i)

         elif dibdumper.bmpInfoHeader_biBitCount == 24: # Works
            '''*
            * Each three bytes read in is 1 column. Each scan line is padded to by a multiple of 4 bytes. The disk image has only 3 however.
            *'''
            #print("Case 24")
            iPelsPerRow = dibdumper.bmpInfoHeader_biWidth
            iDeadBytes = (4 - (iPelsPerRow * 3) % 4) % 4

            for row in range(dibdumper.bmpInfoHeader_biHeight): # read over the rows
               if (dibdumper.topDownDIB):
                  i = row
               else:
                  i = dibdumper.bmpInfoHeader_biHeight - 1 - row

               for j in range(iPelsPerRow):      # j is now just the column counter
                  rgbQuad_rgbBlue      = struct.unpack(">B", fstream.read(1))[0]
                  rgbQuad_rgbGreen     = struct.unpack(">B", fstream.read(1))[0]
                  rgbQuad_rgbRed       = struct.unpack(">B", fstream.read(1))[0]
                  pel = (rgbQuad_rgbRed << 16) | (rgbQuad_rgbGreen << 8) | rgbQuad_rgbBlue
                  dibdumper.imageArray[i][j] = pel

               for j in range(iDeadBytes):
                  struct.unpack(">B", fstream.read(1))[0] # Now read in the "dead bytes" to pad to a 4 byte boundary
            
         elif dibdumper.bmpInfoHeader_biBitCount == 32: # Works
            '''*
            * Each four bytes read in is 1 column. The number of bytes per line will always be a multiple of 4, so there are no dead bytes.
            *'''
            #print("Case 32")
            iPelsPerRow = dibdumper.bmpInfoHeader_biWidth
            for row in range(dibdumper.bmpInfoHeader_biHeight): # read over the rows
               if (dibdumper.topDownDIB):
                  i = row
               else:
                  i = dibdumper.bmpInfoHeader_biHeight - 1 - row

               for j in range(iPelsPerRow):       # j is now just the column counter
                  rgbQuad_rgbBlue      = struct.unpack(">B", fstream.read(1))[0]
                  #print(rgbQuad_rgbBlue)
                  rgbQuad_rgbGreen     = struct.unpack(">B", fstream.read(1))[0]
                  #print(rgbQuad_rgbGreen)
                  rgbQuad_rgbRed       = struct.unpack(">B", fstream.read(1))[0]
                  #print(rgbQuad_rgbRed)
                  rgbQuad_rgbReserved  = struct.unpack(">B", fstream.read(1))[0]
                  #print(rgbQuad_rgbReserved)
                  pel =  (rgbQuad_rgbReserved << 24) |(rgbQuad_rgbRed << 16) | (rgbQuad_rgbGreen << 8) | rgbQuad_rgbBlue
                  #print(pel)
                  dibdumper.imageArray[i][j] = pel
                  #print("Good")

         else: # Oops
            print("This error should not occur - 1!\n")

            # switch (bmpInfoHeader_biBitCount)
         fstream.close()

      except Exception as err:
         print("File input error" + str(err))

      '''*
      * Console dump of image bytes in HEX if the image is smaller than 33 x 33
      *'''

      if ((dibdumper.bmpInfoHeader_biWidth < 33) and (dibdumper.bmpInfoHeader_biHeight < 33)):
         iBytesPerRow = dibdumper.bmpInfoHeader_biWidth
         for i in range(dibdumper.bmpInfoHeader_biHeight): # read over the rows
            for j in range(iBytesPerRow):         # j is now just the column counter
               print(hex(dibdumper.imageArray[i][j]), end = '\t')
            print("\n", end = '')

      '''*
      * Now write out the true color bitmap to a disk file. This is here mostly to be sure we did it all correctly.
      *
      *'''
      try:
         iDeadBytes = (4 - (dibdumper.bmpInfoHeader_biWidth * 3) % 4) % 4

         dibdumper.bmpInfoHeader_biSizeImage =  (dibdumper.bmpInfoHeader_biWidth * 3 + iDeadBytes) * dibdumper.bmpInfoHeader_biHeight
         dibdumper.bmpFileHeader_bfOffBits = 54        # 54 byte offset for 24 bit images (just open one with this app to get this value)
         dibdumper.bmpFileHeader_bfSize = dibdumper.bmpInfoHeader_biSizeImage + dibdumper.bmpFileHeader_bfOffBits
         dibdumper.bmpInfoHeader_biBitCount = 24       # 24 bit color image
         dibdumper.bmpInfoHeader_biCompression = 0     # BI_RGB (which is a value of zero)
         dibdumper.bmpInfoHeader_biClrUsed = 0         # Zero for true color
         dibdumper.bmpInfoHeader_biClrImportant = 0    # Zero for true color

         fstream = open(outFileName, 'w')
            # BITMAPFILEHEADER
         # there is no color table for this true color image, so write out the pels

         for i in range(dibdumper.bmpInfoHeader_biHeight):    # write over the rows (in the usual inverted format)
            for j in range(dibdumper.bmpInfoHeader_biWidth): # and the columns
               pel = dibdumper.imageArray[i][j]
               rgbQuad_rgbBlue  = pel & 0x00FF
               rgbQuad_rgbGreen = (pel >> 8)  & 0x00FF
               rgbQuad_rgbRed   = (pel >> 16) & 0x00FF
               pel = dibdumper.colorToGrayscale(pel)
               fstream.write(str(pel) + '\n') # lowest byte in the color
               # highest byte in the color
               # Now write out the "dead bytes" to pad to a 4 byte boundary
               
            # for (i = bmpInfoHeader_biHeight - 1 i >= 0 --i)

         fstream.close()

      except Exception as err:
         print("File output error" + str(err))
      # public static void main
   # public class BitmapInput

if (__name__ == "__main__"):
   BitmapInput.main()