package main

import (
	"unicode"
	"unicode/utf8"

	"github.com/leo-liu/zhmakeindex/CJK"
)

// 汉字按拼音排序，按拼音首字母与英文一起分组
type ReadingIndexCollator struct{}

func (_ ReadingIndexCollator) InitGroups(style *OutputStyle) []IndexGroup {
	// 分组：符号、数字、字母 A..Z
	groups := make([]IndexGroup, 2+26)
	if style.headings_flag > 0 {
		groups[0].name = style.symhead_positive
		groups[1].name = style.numhead_positive
		for alph, i := 'A', 2; alph <= 'Z'; alph++ {
			groups[i].name = string(alph)
			i++
		}
	} else if style.headings_flag < 0 {
		groups[0].name = style.symhead_negative
		groups[1].name = style.numhead_negative
		for alph, i := 'a', 2; alph <= 'z'; alph++ {
			groups[i].name = string(alph)
			i++
		}
	}
	return groups
}

// 取得分组
func (_ ReadingIndexCollator) Group(entry *IndexEntry) int {
	first, _ := utf8.DecodeRuneInString(entry.level[0].key)
	first = unicode.ToLower(first)
	switch {
	case IsNumString(entry.level[0].key):
		return 1
	case 'a' <= first && first <= 'z':
		return 2 + int(first) - 'a'
	case CJK.Readings[first] != "":
		// 拼音首字母
		reading_first := int(CJK.Readings[first][0])
		return 2 + reading_first - 'a'
	default:
		// 符号组
		return 0
	}
}

// 按汉字读音比较两个字符，读音相同的，内码序
func (_ ReadingIndexCollator) RuneCmp(a, b rune) int {
	a_reading, b_reading := CJK.Readings[a], CJK.Readings[b]
	switch {
	case a_reading == "" && b_reading == "":
		return RuneCmpIgnoreCases(a, b)
	case a_reading == "" && b_reading != "":
		return -1
	case a_reading != "" && b_reading == "":
		return 1
	case a_reading < b_reading:
		return -1
	case a_reading > b_reading:
		return 1
	default:
		return int(a - b)
	}
}

// 判断是否字母或汉字
func (_ ReadingIndexCollator) IsLetter(r rune) bool {
	r = unicode.ToLower(r)
	switch {
	case 'a' <= r && r <= 'z':
		return true
	case CJK.Readings[r] != "":
		return true
	default:
		return false
	}
}