#!/usr/bin/python3

import math

from lib import *


MAP = {
	'ALetter':            'LE',
	'CR':                 'CR',
	'Double_Quote':       'DQ',
	'E_Base':             'EB',
	'E_Base_GAZ':         'EBG',
	'E_Modifier':         'EM',
	'Extend':             'EXTEND',
	'ExtendNumLet':       'EX',
	'Format':             'FO',
	'Glue_After_Zwj':     'GAZ',
	'Hebrew_Letter':      'HL',
	'Katakana':           'KA',
	'LF':                 'LF',
	'MidLetter':          'ML',
	'MidNumLet':          'MB',
	'MidNum':             'MN',
	'Newline':            'NL',
	'Numeric':            'NU',
	'Other':              'XX',
	'Regional_Indicator': 'RI',
	'Single_Quote':       'SQ',
	'WSegSpace':          'WSEGSPACE',
	'ZWJ':                'ZWJ',
}

longest = 0

def parse(file: str) -> list[bool]:
	global longest

	xs = ['WB_XX'] * 0x110000
	with open(file, 'r') as f:
		for line in f.readlines():
			if len(line.strip()) == 0 or line[0] == '#':
				continue

			parts = line.split(';')
			ranges = [int(x, 16) for x in parts[0].strip().split('..')]
			prop = 'WB_' + MAP[parts[1].split('#')[0].strip()]
			longest = max(longest, len(prop))

			for i in range(ranges[0], ranges[len(ranges) - 1] + 1):
				xs[i] = prop
	return xs

def genfile(cs: list[tuple[bool, ...]], blksize: int) -> None:
	Cs = cs
	cs = list(dict.fromkeys(Cs))

	print('''\
/* This file is autogenerated by gen/prop/wb; DO NOT EDIT. */

#include <stdint.h>

#include "unicode/prop.h"
''')

	print(f'static constexpr {typename(len(cs) - 1)} stage1[] = {{')
	for i, c in enumerate(Cs):
		print(f'%c%{len(str(len(cs) - 1))}d,' % ('\t' if i % 16 == 0 else ' ', cs.index(c)), end='')
		if i % 16 == 15:
			print()
	print('};')

	print()

	ppc = columns(blksize, longest + 1)
	print(f'static constexpr enum uprop_wb stage2[][{blksize}] = {{')
	for c in cs:
		for i in range(blksize // ppc):
			print('\t{' if i == 0 else '\t ', end='')
			for j in range(ppc):
				print(c[i*ppc + j], end='')
				if i < blksize // ppc - 1 or j < ppc - 1:
					print(',', end='')
				if j < ppc - 1:
					print(' ' * (longest + 1 - len(c[i*ppc + j])), end='')
			if i < blksize // ppc - 1:
				print()
		print('},')
	print('};')

	print()

	print(f'''\
enum uprop_wb
uprop_get_wb(rune ch)
{{
	return stage2[stage1[ch / {blksize}]][ch % {blksize}];
}}''')

def main() -> None:
	cwd_init()
	xs = parse('data/WordBreakProperty')

	blksize = -1
	smallest = math.inf

	for bs in powers_of_2():
		if bs > len(xs):
			break
		Cs = [tuple(x) for x in chunks(xs, bs)]
		cs = set(Cs)

		sz_s1 = len(Cs) * isize(len(cs) - 1)
		sz_s2 = len(cs) * bs
		sz = sz_s1 + sz_s2

		if sz < smallest:
			smallest = sz
			blksize = bs

	Cs = [tuple(x) for x in chunks(xs, blksize)]
	with open('lib/unicode/prop/uprop_get_wb.c', 'w') as f:
		sys.stdout = f
		genfile(Cs, blksize)

	report_size(len(xs), smallest)

if __name__ == '__main__':
	main()