#!/usr/bin/env python3

import argparse
import os, re, sys, struct
import textwrap
import traceback


def sanitize_name(name):
    # replace special characters with underscores
    name = re.sub('[^a-zA-Z0-9_]', '_', name)
    # clean up
    name = name.strip('_')

    # not needed: 'DATA_' prefix
    #if name[0] in "0123456789": name = '_' + name

    return name


def main():
    parser = argparse.ArgumentParser(
        description="Convert data files into byte arrays in a C header")

    parser.add_argument('input', nargs='*', action='append',
                        help="Input data file to convert to a header file. If "
                            +"none are specified, data is read from standard input.")
    parser.add_argument('--output', type=str, default=None,
                        help="Output header file path, standard output if none specified")

    args = parser.parse_args()

    #print(repr(args), file=sys.stderr)

    inputfiles = args.input[0]
    datafiles = {}

    if len(inputfiles) == 0:
        # need to read from stdio
        datafiles['data'] = sys.stdin.buffer.read()
    else:
        for path in inputfiles:
            # cannot use os.access, POSIX-only
            if not os.path.exists(path) or os.path.isdir(path):
                print("ERROR: input path '%s' does not exist, or is not a file" % path)
                sys.exit(1)

            # get base name of the path (no folder names, no extension)
            name = os.path.splitext(os.path.split(path)[1])[0]
            # make it usable for C code names
            name = sanitize_name(name)

            # now read the data
            data = None
            try:
                with open(path, 'rb') as f:
                    data = f.read()
            except IOError:
                print("ERROR: input file '%s' not readable" % path)
                traceback.print_exc()
                sys.exit(1)

            data = struct.unpack('<'+('H'*(len(data)//2)), data)

            # add it to the list
            datafiles[name] = data

    #print(repr(datafiles), file=sys.stderr)

    outname = "hardcoded_data"
    if args.output is not None:
        outname = os.path.splitext(os.path.split(args.output)[1])[0]
        outname = sanitize_name(outname)

    # generate the C source code: one uint8_t array for each data file
    # variable name is "DATA_<datafile>"
    # (width=84 is the equivalent of 16 bytes/line)
    headersrc = '\n\n'.join(
        "static uint16_t DATA_%s[%d] = {\n%s\n};" % (name, len(data), \
            textwrap.fill(','.join("0x%04x" % b for b in data), width=81,
                          initial_indent="    ", subsequent_indent="    ")
        ) for name, data in datafiles.items()
    )

    # a C header needs include guards
    headersrc = """
#ifndef AUTOGEN_{0}_H_
#define AUTOGEN_{0}_H_

{1}

#endif
""".format(outname.upper(), headersrc)

    #print(headersrc, file=sys.stderr)

    if args.output is None:
        # no output specified -> output to stdout
        sys.stdout.write(headersrc)
    else:
        try:
            with open(args.output, 'w') as f:
                f.write(headersrc)
        except IOError:
            print("ERROR: output file '%s' not writable" % args.output)
            traceback.print_exc()
            sys.exit(1)


if __name__ == '__main__':
    main()