rime-ice/others/script/rime/check.go

320 lines
8.6 KiB
Go
Raw Normal View History

2022-10-30 16:47:40 +01:00
package rime
import (
"bufio"
"fmt"
2023-04-13 11:04:44 +02:00
mapset "github.com/deckarep/golang-set/v2"
2022-10-30 16:47:40 +01:00
"log"
"os"
"path"
"strconv"
"strings"
2023-04-13 11:04:44 +02:00
"sync"
2022-10-30 16:47:40 +01:00
"time"
"unicode"
"unicode/utf8"
)
2023-04-13 11:04:44 +02:00
var (
specialWords = mapset.NewSet[string]() // 特殊词汇列表,不进行任何检查
polyphoneWords = mapset.NewSet[string]() // 需要注音的字词
2022-10-30 16:47:40 +01:00
2023-04-13 11:04:44 +02:00
wrongWords = mapset.NewSet[string]() // 异形词、错别字
wrongWordsFilter = mapset.NewSet[string]() // 过滤一部分,包含错别字但不是错别字,如「作爱」是错的,但「叫作爱」是对的
hanPinyin = make(map[string][]string) // 汉字拼音映射,用于检查注音是否正确
hanPinyinFilter = mapset.NewSet[string]() // 过滤一部分比如「深厉浅揭qi」只在这个词中念qi并不是错误。
)
// 初始化特殊词汇列表、需要注音列表、错别字列表、拼音列表
2022-10-30 16:47:40 +01:00
func init() {
2023-04-13 11:04:44 +02:00
// 特殊词汇列表,不进行任何检查
2022-10-30 16:47:40 +01:00
specialWords.Add("狄尔斯–阿尔德反应")
specialWords.Add("特里斯坦–达库尼亚")
specialWords.Add("特里斯坦–达库尼亚群岛")
specialWords.Add("茱莉亚·路易斯-德瑞弗斯")
specialWords.Add("科科斯(基林)群岛")
specialWords.Add("刚果(金)")
specialWords.Add("刚果(布)")
specialWords.Add("赛博朋克:边缘行者")
specialWords.Add("赛博朋克:边缘跑手")
specialWords.Add("赛博朋克:命运之轮")
2023-01-26 10:02:41 +01:00
specialWords.Add("哈勃–勒梅特定律")
2022-10-30 16:47:40 +01:00
2023-04-13 11:04:44 +02:00
// 需要注音的列表
file1, err := os.Open(需要注音TXT)
2022-10-30 16:47:40 +01:00
if err != nil {
2023-04-13 11:04:44 +02:00
log.Fatalln(err)
2022-10-30 16:47:40 +01:00
}
2023-04-13 11:04:44 +02:00
defer file1.Close()
sc := bufio.NewScanner(file1)
2022-10-30 16:47:40 +01:00
for sc.Scan() {
line := sc.Text()
if strings.HasPrefix(line, "#") {
continue
}
polyphoneWords.Add(line)
}
2023-04-13 11:04:44 +02:00
// 错别字的两个列表: wrongWords wrongWordsFilter
file2, err := os.Open(错别字TXT)
2022-10-30 16:47:40 +01:00
if err != nil {
2023-04-13 11:04:44 +02:00
log.Fatalln(err)
2022-10-30 16:47:40 +01:00
}
2023-04-13 11:04:44 +02:00
defer file2.Close()
sc = bufio.NewScanner(file2)
2022-10-30 16:47:40 +01:00
isMark := false
for sc.Scan() {
line := sc.Text()
2023-04-13 11:04:44 +02:00
if strings.HasPrefix(line, "# -_-") {
2022-10-30 16:47:40 +01:00
isMark = true
continue
}
2023-04-28 17:57:40 +02:00
if strings.HasPrefix(line, "#") || line == "" {
continue
}
2022-10-30 16:47:40 +01:00
if !isMark {
wrongWords.Add(line)
} else {
2023-04-13 11:04:44 +02:00
wrongWordsFilter.Add(line)
2022-10-30 16:47:40 +01:00
}
}
2022-11-03 07:16:48 +01:00
2023-04-13 11:04:44 +02:00
// 汉字拼音映射 hanPinyin hanPinyinFilter
// 将所有读音读入 hanPinyin
file3, err := os.Open(HanziPath)
2022-11-03 07:16:48 +01:00
if err != nil {
2023-04-13 11:04:44 +02:00
log.Fatalln(err)
2022-11-03 07:16:48 +01:00
}
2023-04-13 11:04:44 +02:00
defer file3.Close()
2022-11-03 07:16:48 +01:00
isMark = false
2023-04-13 11:04:44 +02:00
sc = bufio.NewScanner(file3)
2022-11-03 07:16:48 +01:00
for sc.Scan() {
line := sc.Text()
if !isMark {
if strings.Contains(line, mark) {
isMark = true
}
continue
}
parts := strings.Split(line, "\t")
text, code := parts[0], parts[1]
2023-04-13 11:04:44 +02:00
hanPinyin[text] = append(hanPinyin[text], code)
2022-11-03 07:16:48 +01:00
}
2023-04-13 11:04:44 +02:00
// 给 hanPinyin 补充不再字表的读音,和过滤列表 hanPinyinFilter
file4, err := os.Open(汉字拼音映射TXT)
2022-11-03 07:16:48 +01:00
if err != nil {
2023-04-13 11:04:44 +02:00
log.Fatalln(err)
2022-11-03 07:16:48 +01:00
}
2023-04-13 11:04:44 +02:00
defer file4.Close()
sc = bufio.NewScanner(file4)
2022-11-03 07:16:48 +01:00
isMark = false
for sc.Scan() {
line := sc.Text()
2023-04-13 11:04:44 +02:00
if strings.HasPrefix(line, "#") && !strings.HasPrefix(line, "# -_-") || line == "" {
2022-11-03 07:16:48 +01:00
continue
}
if strings.HasPrefix(line, "# -_-") {
isMark = true
continue
}
if !isMark {
parts := strings.Split(line, " ")
key := parts[0]
values := parts[1:]
2023-04-13 11:04:44 +02:00
hanPinyin[key] = append(hanPinyin[key], values...)
2022-11-03 07:16:48 +01:00
} else {
2023-04-13 11:04:44 +02:00
hanPinyinFilter.Add(line)
2022-11-03 07:16:48 +01:00
}
}
2022-10-30 16:47:40 +01:00
}
// Check 对传入的词库文件进行检查
2023-04-13 11:04:44 +02:00
// dictPath: 词库文件路径
// _type: 词库类型 1 只有汉字 2 汉字+注音 3 汉字+注音+权重 4 汉字+权重
func Check(dictPath string, _type int) {
2022-10-30 16:47:40 +01:00
// 控制台输出
defer printlnTimeCost("检查 "+path.Base(dictPath), time.Now())
// 打开文件
file, err := os.Open(dictPath)
if err != nil {
2023-04-13 11:04:44 +02:00
log.Fatalln(err)
2022-10-30 16:47:40 +01:00
}
defer file.Close()
// 开始检查!
lineNumber := 0
isMark := false
sc := bufio.NewScanner(file)
2023-04-13 11:04:44 +02:00
var wg sync.WaitGroup
2022-10-30 16:47:40 +01:00
for sc.Scan() {
lineNumber++
line := sc.Text()
if !isMark {
2023-04-13 11:04:44 +02:00
if strings.HasPrefix(line, mark) {
2022-10-30 16:47:40 +01:00
isMark = true
}
continue
}
2023-04-13 11:04:44 +02:00
wg.Add(1)
go checkLine(dictPath, _type, line, lineNumber, &wg)
}
2022-10-30 16:47:40 +01:00
2023-04-13 11:04:44 +02:00
if err := sc.Err(); err != nil {
log.Fatalln(err)
}
wg.Wait()
}
// 检查一行
func checkLine(dictPath string, _type int, line string, lineNumber int, wg *sync.WaitGroup) {
defer wg.Done()
2022-10-30 16:47:40 +01:00
2023-04-13 11:04:44 +02:00
// 忽略注释base 中有很多被注视了词汇,暂时没有删除
if strings.HasPrefix(line, "#") {
// 注释以 '#' 开头,但不是以 '# '开头(强迫症晚期)
if !strings.HasPrefix(line, "# ") {
2022-10-30 16:47:40 +01:00
fmt.Println("has # but not #␣", line)
}
2023-04-13 11:04:44 +02:00
return
}
2022-10-30 16:47:40 +01:00
2023-04-13 11:04:44 +02:00
// 有空行
if strings.TrimSpace(line) == "" {
fmt.Println("empty line", line)
}
2022-10-30 16:47:40 +01:00
2023-04-13 11:04:44 +02:00
// 开头结尾有空格
if strings.HasPrefix(line, " ") || strings.HasSuffix(line, " ") {
fmt.Println()
}
// +---------------------------------------------------------------
// | 开始检查分割后的内容,分割为: 词汇text 编码code 权重weight
// +---------------------------------------------------------------
parts := strings.Split(line, "\t")
var text, code, weight string
switch {
case _type == 1 && len(parts) == 1: // 一列,【汉字】
text = parts[0]
case _type == 2 && len(parts) == 2: // 两列,【汉字+注音】
text, code = parts[0], parts[1]
case _type == 3 && len(parts) == 3: // 散列,【汉字+注音+权重】
text, code, weight = parts[0], parts[1], parts[2]
case _type == 4 && len(parts) == 2: // 两列,【汉字+权重】
text, weight = parts[0], parts[1]
default:
2023-04-17 14:46:35 +02:00
log.Fatalln("❌ 分割错误:", lineNumber, line)
2023-04-13 11:04:44 +02:00
}
2022-10-30 16:47:40 +01:00
2023-04-13 11:04:44 +02:00
// weight 应该是纯数字
if weight != "" {
_, err := strconv.Atoi(weight)
if err != nil {
fmt.Println("❌ weight 非数字:", line)
2022-10-30 16:47:40 +01:00
}
2023-04-13 11:04:44 +02:00
}
2022-10-30 16:47:40 +01:00
2023-04-13 11:04:44 +02:00
// text 和 weight 不应该含有空格
if strings.Contains(text, " ") || strings.Contains(weight, " ") {
fmt.Println("❌ text 和 weight 含有空格:", line)
}
2022-10-30 16:47:40 +01:00
2023-04-13 11:04:44 +02:00
// code 前后不应该有空格
if strings.HasPrefix(code, " ") || strings.HasSuffix(code, " ") {
fmt.Println("❌ code 前后有空格:", line)
}
2022-10-30 16:47:40 +01:00
2023-04-13 11:04:44 +02:00
// code 不应该有非小写字母
for _, r := range code {
if string(r) != " " && !unicode.IsLower(r) {
fmt.Println("❌ code 含有非小写字母:", line)
break
2022-10-30 16:47:40 +01:00
}
2023-04-13 11:04:44 +02:00
}
2022-10-30 16:47:40 +01:00
2023-04-13 11:04:44 +02:00
// 过滤特殊词条
if specialWords.Contains(text) {
return
}
2022-10-30 16:47:40 +01:00
2023-04-13 11:04:44 +02:00
// text 不应该有非汉字内容,除了间隔号 ·
for _, c := range text {
if string(c) != "·" && !unicode.Is(unicode.Han, c) {
fmt.Println("❌ text 含有非汉字内容:", line)
break
2022-10-30 16:47:40 +01:00
}
2023-04-13 11:04:44 +02:00
}
2022-10-30 16:47:40 +01:00
2023-04-13 11:04:44 +02:00
// 除了字表,其他词库不应该含有单个的汉字
if dictPath != HanziPath && utf8.RuneCountInString(text) == 1 {
fmt.Println("❌ 意外的单个汉字:", line)
}
2022-10-30 16:47:40 +01:00
2023-04-13 11:04:44 +02:00
// 除了 base其他词库不应该含有两个字的词汇
if dictPath != BasePath && utf8.RuneCountInString(text) == 2 {
fmt.Println("❌ 意外的两字词:", line)
}
2022-10-30 16:47:40 +01:00
2023-04-13 11:04:44 +02:00
// 汉字个数应该和拼音个数相等
if code != "" {
codeCount := len(strings.Split(code, " "))
textCount := utf8.RuneCountInString(text)
if strings.Contains(text, "·") {
textCount -= strings.Count(text, "·")
2022-10-30 16:47:40 +01:00
}
2023-04-13 11:04:44 +02:00
if strings.HasPrefix(text, "# ") {
textCount -= 2
2022-10-30 16:47:40 +01:00
}
2023-04-13 11:04:44 +02:00
if textCount != codeCount {
fmt.Println("❌ 汉字个数 != 拼音个数:", line)
}
}
2022-10-30 16:47:40 +01:00
2023-04-13 11:04:44 +02:00
// +---------------------------------------------------------------
// | 其他检查
// +---------------------------------------------------------------
// 需要注音但没有注音的字
if dictPath == ExtPath || dictPath == TencentPath {
for _, word := range polyphoneWords.ToSlice() {
if strings.Contains(text, word) {
fmt.Println("❌ 需要注音:", line)
2022-11-03 07:16:48 +01:00
}
2023-04-13 11:04:44 +02:00
}
}
2022-11-03 07:16:48 +01:00
2023-04-13 11:04:44 +02:00
// 检查拼写错误如「赞zan」写成了zna顺便检查是否存在字表中没有注音的字
if dictPath != HanziPath && (_type == 2 || _type == 3) && !hanPinyinFilter.Contains(text) {
// 把汉字和拼音弄成一一对应关系,「拼音:pin yin」→「拼:pin」「音:yin」
textWithoutDian := strings.ReplaceAll(text, "·", "") // 去掉间隔号
pinyins := strings.Split(code, " ")
i := 0
for _, zi := range textWithoutDian {
if !contains(hanPinyin[string(zi)], pinyins[i]) {
2023-05-03 12:20:13 +02:00
fmt.Printf("❌ 注音错误 or 字表未包含的汉字及注音: %s - %s.+%s\n", line, string(zi), pinyins[i])
2022-10-30 16:47:40 +01:00
}
2023-04-13 11:04:44 +02:00
i++
}
}
2022-10-30 16:47:40 +01:00
2023-04-13 11:04:44 +02:00
// 错别字检查,最耗时的检查
if dictPath != HanziPath && !wrongWordsFilter.Contains(text) {
wrongWords.Each(func(wrongWord string) bool {
2023-05-03 12:20:13 +02:00
if strings.HasPrefix(wrongWord, "=") {
wrongWord = strings.TrimLeft(wrongWord, "= ")
if text == wrongWord {
fmt.Printf("❌ 错别字: %s = %s\n", text, wrongWord)
return true
}
} else if strings.Contains(text, wrongWord) {
2023-04-13 11:04:44 +02:00
fmt.Printf("❌ 错别字: %s - %s\n", text, wrongWord)
return true
2022-10-30 16:47:40 +01:00
}
2023-04-13 11:04:44 +02:00
return false
})
2022-10-30 16:47:40 +01:00
}
}