Улучшил определение codePage

This commit is contained in:
Book Pauk
2019-02-09 11:33:04 +07:00
parent 3095231f92
commit 1f6dd9a00f
2 changed files with 77 additions and 2 deletions

View File

@@ -4,6 +4,7 @@ const iconv = require('iconv-lite');
const chardet = require('chardet');
const _ = require('lodash');
const sax = require('./sax');
const getEncoding = require('./getEncoding');
const FileDetector = require('../FileDetector');
@@ -43,9 +44,9 @@ class BookConverter {
}
decode(data) {
const charsetAll = chardet.detectAll(data.slice(0, 10000));
const charsetAll = chardet.detectAll(data.slice(0, 20000));
let selected = 'ISO-8859-1';
let selected = 'ISO-8859-5';
for (const charset of charsetAll) {
if (charset.name.indexOf('ISO-8859') < 0) {
selected = charset.name;
@@ -53,6 +54,10 @@ class BookConverter {
}
}
if (selected == 'ISO-8859-5') {
selected = getEncoding(data);
}
return iconv.decode(data, selected);
}