#!/usr/bin/env python

import sys
import StringIO
import string
import exceptions

def parse_Unihan(input_file, dictionary):
	entries = [line.strip().split("\t") for line in input_file.readlines() if line.strip() != "" and line.strip()[0] != "#"]
	for entry in entries:
		assert(entry[0].startswith("U+"))
		code = int(entry[0][2 : ], 16)
		key = entry[1]
		value = entry[2]
		if code not in dictionary:
			dictionary[code] = []
		dictionary[code].append((key, value))

def get_value(entry, key):
	for x_key, x_value in entry:
		if x_key == key:
			return x_value
	return None

def escape(text):
	if text:
		return text.replace("<", "&lt;").replace(">", "&gt;")
	else:
		return ""

vowels = ["A", "I", "U", "E", "O"]
unromaji_table = { # TODO JI => ZI; or DI; 3063 small tsu. TSU. small YA=3084, SMALL YU=3085, SMALL YO=3087
	# TODO "N"=3093; VU=3094; TI=CHI, DI=JI
	# TODO difference HU/FU
	None: [u"\u3042", u"\u3044", u"\u3046", u"\u3048", u"\u304A", ],
	"W":  [u"\u308F", None,      None,      None,      u"\u3092", ],
	"R":  [u"\u3089", u"\u308A", u"\u308B", u"\u308C", u"\u308D", ],
	"Y":  [u"\u3084", None,      u"\u3086", None,      u"\u3088", ],
	"M":  [u"\u307E", u"\u307F", u"\u3080", u"\u3081", u"\u3082", ],
	"H":  [u"\u306F", u"\u3072", u"\u3075", u"\u3078", u"\u307B", ],
	"F":  [None,      None,      u"\u3075", None,      None       ],
	"N":  [u"\u306A", u"\u306B", u"\u306C", u"\u306D", u"\u306E", ],
	"T":  [u"\u305F", u"\u3061", u"\u3064", u"\u3066", u"\u3068", ],
	"S":  [u"\u3055", u"\u3057", u"\u3059", u"\u305B", u"\u305D", ],
	"K":  [u"\u304B", u"\u304D", u"\u304F", u"\u3051", u"\u3053", ],

	"B":  [u"\u306F", u"\u3073", u"\u3076", u"\u3079", u"\u307C", ],
	"D":  [u"\u3060", u"\u3062", u"\u3065", u"\u3067", u"\u3069", ],
	"Z":  [u"\u3056", u"\u3058", u"\u305A", u"\u305C", u"\u305E", ],
	"G":  [u"\u304C", u"\u304E", u"\u3050", u"\u3052", u"\u3054", ],
	"P":  [u"\u3071", u"\u3074", u"\u3077", u"\u307A", u"\u307D", ],

	"y":  [u"\u3084", None,      u"\u3085", None,      u"\u3087", ],
	"tiny":[u"\u3041", u"\u3043", u"\u3045", u"\u3047", u"\u3049", ],
}

# KYA = KI & YA
# NYUUMU

def unromaji_2(text):
	# FIXME handle N/NI/NYA
	if text is None:
		return text
	result = StringIO.StringIO()
	for c in text:
		assert(c in (string.letters + " "))
	count = 0
	half = None
	previous_c = None
	for c in text:
		if c == " ":
			result.write(" ") # TODO remove?
			previous_c = c
			continue
		#print >>sys.stderr, "CC", count, c
		if count == 0:
			#if c == "N":
			#	result.write("&#x%x;" % 0x3093)
			if c in vowels:
				i = vowels.index(c)
				assert(i != -1)
				half = unromaji_table[None]
				code = ord(half[i])
				result.write("&#x%x;" % code)
			else:
				half_code = c
				if c == "J":
					c = "Z"
				elif c == "C":
					c = "T"
				half = unromaji_table[c]
				count = count + 1
		elif count == 1:
			if c == previous_c: # PPA
				# TODO limit
				result.write("&#x%x;" % 0x3063)
				count = 0
				half_code = c
				half = unromaji_table[c if c != "J" else "Z"]
				count = count + 1
				previous_c = c
				continue

			if c == "S" and half_code == "T": # FIXME not nice.
				previous_c = c
				continue
			if c in ["H", "Y"]:
				i = vowels.index("I")
				assert(i != -1)
				code = ord(half[i])
				result.write("&#x%x;" % code)
				count = count + 1
				previous_c = c
				continue
			try:
				i = vowels.index(c)
			except exceptions.ValueError:
				if half_code == "N":
					result.write("&#x%x;" % 0x3093)
					count = 0
					half_code = c
					if c == "J":
						c = "Z"
					elif c == "C":
						c = "T"
					half = unromaji_table[c]
					count = count + 1
					previous_c = c
					continue
				else:
					raise

			assert(i != -1)
			code = half[i]
			#print >>sys.stderr, "CODE", half, half_code
			if code is None and half_code == "F": # foreign
				result.write("F") # TODO katakana?
				result.write(c)
			else:
				code = ord(code)
				result.write("&#x%x;" % code)
			count = 0
		elif count == 2:
			if c != "I":
				i = vowels.index(c)
				code = unromaji_table["y"][i]
				result.write("&#x%x;" % ord(code))
			count = 0
		previous_c = c

	# handle trailing "N":
	if count == 1:
		result.write("&#x%x;" % ord(half_code))

	return result.getvalue()

def unspace(text):
	if text is not None:
		return text.replace(" ", "|")
	else:
		return text

def unromaji(text):
	try:
		return unromaji_2(text)
	except:
		print >>sys.stderr, "broken:", text
		return text

def get_Japanese_entries(dictionary):
	for code in range(0x4e00, 0x9faf + 1):
		if code >= 0x9fa6 and code <= 0x9faf: # FIXME ???
			continue
		dictionary_entry = dictionary[code]
		JapaneseKun = get_value(dictionary_entry, "kJapaneseKun")
		JapaneseOn = get_value(dictionary_entry, "kJapaneseOn")
		Definition = get_value(dictionary_entry, "kDefinition")
		if JapaneseKun or JapaneseOn:
			yield dictionary_entry, """
<tr><td class="kanji" title="U+%X" rowspan="4">&#x%x;</td><td class="meaning"
>%s</td></tr>
<tr><td class="kun_reading" title="%s">%s</td></tr>
<tr><td class="on_reading">%s</td></tr>
<tr><td class="extra">%%s</td></tr>""" % (code, code, escape(Definition), escape(unspace(JapaneseKun)), unspace(unromaji(JapaneseKun)) or "", escape(unspace(JapaneseOn)))

def group_by_Grade(entries):
	entries = [entry for entry in entries]
	grades = set([int(get_value(dictionary_entry, "kGradeLevel") or 99) for dictionary_entry, HTML in entries])
	result = []
	for grade in sorted([grade for grade in grades]):
		g_entries = [(dictionary_entry, HTML) for dictionary_entry, HTML in entries if int(get_value(dictionary_entry, "kGradeLevel") or 99) == grade]
		result.append((grade, g_entries))
	return sorted(result)

dictionary = {}

parse_Unihan(file("Unihan/Unihan_DictionaryLikeData.txt", "r"), dictionary)
parse_Unihan(file("Unihan/Unihan_Readings.txt", "r"), dictionary)

sys.stdout.write(file("top.txt", "r").read())

for grade, entries in group_by_Grade(get_Japanese_entries(dictionary)):
	print "</table>"
	print "<h2>Grade %s</h2>" % grade
	print "<table border=\"1\">"
	for dictionary_entry, HTML in entries:
		TotalStrokes = get_value(dictionary_entry, "kTotalStrokes")
		GradeLevel = get_value(dictionary_entry, "kGradeLevel")
		extra = ",%s strokes" % TotalStrokes
		print HTML % extra

sys.stdout.write(file("bottom.txt", "r").read())
# FIXME order by frequency desc.