Files
liberama/server/core/Reader/BookConverter/ConvertPdf.js
2020-12-14 02:22:38 +07:00

376 lines
12 KiB
JavaScript

//const _ = require('lodash');
const fs = require('fs-extra');
const path = require('path');
const sax = require('../../sax');
const utils = require('../../utils');
const ConvertHtml = require('./ConvertHtml');
class ConvertPdf extends ConvertHtml {
check(data, opts) {
const {inputFiles} = opts;
return this.config.useExternalBookConverter &&
inputFiles.sourceFileType && inputFiles.sourceFileType.ext == 'pdf';
}
async run(notUsed, opts) {
if (!this.check(notUsed, opts))
return false;
await this.checkExternalConverterPresent();
const {inputFiles, callback, abort, uploadFileName} = opts;
const inpFile = inputFiles.sourceFile;
const outBasename = `${inputFiles.filesDir}/${utils.randomHexString(10)}`;
const outFile = `${outBasename}.xml`;
const pdftohtmlPath = '/usr/bin/pdftohtml';
if (!await fs.pathExists(pdftohtmlPath))
throw new Error('Внешний конвертер pdftohtml не найден');
//конвертируем в xml
let perc = 0;
await this.execConverter(pdftohtmlPath, ['-nodrm', '-c', '-s', '-xml', inpFile, outFile], () => {
perc = (perc < 80 ? perc + 10 : 40);
callback(perc);
}, abort);
callback(80);
const data = await fs.readFile(outFile);
callback(90);
await utils.sleep(100);
//парсим xml
let lines = [];
let pagelines = [];
let line = {text: ''};
let page = {};
let fonts = {};
let sectionTitleFound = false;
let images = [];
let loading = [];
let inText = false;
let bold = false;
let italic = false;
let i = -1;
const loadImage = async(image) => {
const src = path.parse(image.src);
let type = 'unknown';
switch (src.ext) {
case '.jpg': type = 'image/jpeg'; break;
case '.png': type = 'image/png'; break;
}
if (type != 'unknown') {
image.data = (await fs.readFile(image.src)).toString('base64');
image.type = type;
image.name = src.base;
}
};
const putImage = (curTop) => {
if (!isNaN(curTop) && images.length) {
while (images.length && images[0].top < curTop) {
i++;
lines[i] = images[0];
images.shift();
}
}
};
const isTextBold = (text) => {
const m = text.trim().match(/^<b>(.*)<\/b>$/);
return m && !m[1].match(/<b>|<\/b>|<i>|<\/i>/g);
};
const isTextEmpty = (text) => {
return text.replace(/<b>|<\/b>|<i>|<\/i>/g, '').trim() == '';
};
const putPageLines = () => {
pagelines.sort((a, b) => (Math.abs(a.top - b.top) > 3 ? a.top - b.top : 0)*10000 + (a.left - b.left))
//объединяем в одну строку равные по высоте
const pl = [];
let pt = 0;
let j = -1;
pagelines.forEach(line => {
if (isTextEmpty(line.text))
return;
//проверим, возможно это заголовок
if (line.fontId && line.pageWidth) {
const centerLeft = (line.pageWidth - line.width)/2;
if (isTextBold(line.text) && Math.abs(centerLeft - line.left) < 10) {
if (!sectionTitleFound) {
line.isSectionTitle = true;
sectionTitleFound = true;
} else {
line.isSubtitle = true;
}
}
}
//объединяем
if (pt == 0 || Math.abs(pt - line.top) > 3) {
j++;
pl[j] = line;
} else {
pl[j].text += ` ${line.text}`;
}
pt = line.top;
});
//заполняем lines
const lastIndex = i;
pl.forEach(line => {
putImage(line.top);
//добавим пустую строку, если надо
const prevLine = (i > lastIndex ? lines[i] : {fonts: [], top: 0});
if (prevLine && !prevLine.isImage) {
const f = (prevLine.fontId ? fonts[prevLine.fontId] : (line.fontId ? fonts[line.fontId] : null));
if (f && f.fontSize && !line.isImage && line.top - prevLine.top > f.fontSize * 1.8) {
i++;
lines[i] = {text: '<br>'};
}
}
i++;
lines[i] = line;
});
pagelines = [];
putImage(100000);
};
const onTextNode = (text, cutCounter, cutTag) => {// eslint-disable-line no-unused-vars
if (!cutCounter && inText) {
let tOpen = (bold ? '<b>' : '');
tOpen += (italic ? '<i>' : '');
let tClose = (italic ? '</i>' : '');
tClose += (bold ? '</b>' : '');
line.text += ` ${tOpen}${text}${tClose}`;
}
};
const onStartNode = (tag, tail, singleTag, cutCounter, cutTag) => {// eslint-disable-line no-unused-vars
if (inText) {
switch (tag) {
case 'i':
italic = true;
break;
case 'b':
bold = true;
break;
}
}
if (tag == 'page') {
const attrs = sax.getAttrsSync(tail);
page = {
width: parseInt((attrs.width && attrs.width.value ? attrs.width.value : null), 10),
};
putPageLines();
}
if (tag == 'fontspec') {
const attrs = sax.getAttrsSync(tail);
const fontId = (attrs.id && attrs.id.value ? attrs.id.value : '');
const fontSize = (attrs.size && attrs.size.value ? attrs.size.value : '');
if (fontId) {
fonts[fontId] = {fontSize};
}
}
if (tag == 'text' && !inText) {
const attrs = sax.getAttrsSync(tail);
line = {
text: '',
top: parseInt((attrs.top && attrs.top.value ? attrs.top.value : null), 10),
left: parseInt((attrs.left && attrs.left.value ? attrs.left.value : null), 10),
width: parseInt((attrs.width && attrs.width.value ? attrs.width.value : null), 10),
height: parseInt((attrs.height && attrs.height.value ? attrs.height.value : null), 10),
isSectionTitle: false,
isSubtitle: false,
pageWidth: page.width,
fontId: (attrs.font && attrs.font.value ? attrs.font.value : ''),
};
if (line.width != 0 || line.height != 0) {
inText = true;
pagelines.push(line);
}
}
if (tag == 'image') {
const attrs = sax.getAttrsSync(tail);
let src = (attrs.src && attrs.src.value ? attrs.src.value : '');
if (src) {
const image = {
isImage: true,
src,
data: '',
type: '',
top: parseInt((attrs.top && attrs.top.value ? attrs.top.value : null), 10) || 0,
left: parseInt((attrs.left && attrs.left.value ? attrs.left.value : null), 10) || 0,
width: parseInt((attrs.width && attrs.width.value ? attrs.width.value : null), 10) || 0,
height: parseInt((attrs.height && attrs.height.value ? attrs.height.value : null), 10) || 0,
};
loading.push(loadImage(image));
images.push(image);
images.sort((a, b) => (a.top - b.top)*10000 + (a.left - b.left));
}
}
};
const onEndNode = (tag, tail, singleTag, cutCounter, cutTag) => {// eslint-disable-line no-unused-vars
if (inText) {
switch (tag) {
case 'i':
italic = false;
break;
case 'b':
bold = false;
break;
}
}
if (tag == 'text')
inText = false;
};
let buf = this.decode(data).toString();
sax.parseSync(buf, {
onStartNode, onEndNode, onTextNode
});
putPageLines();
await Promise.all(loading);
await utils.sleep(100);
//найдем параграфы и отступы
const indents = [];
for (const line of lines) {
if (line.isImage)
continue;
if (!isNaN(line.left)) {
indents[line.left] = 1;
}
}
let j = 0;
for (let i = 0; i < indents.length; i++) {
if (indents[i]) {
j++;
indents[i] = j;
}
}
indents[0] = 0;
//author & title
let {author, title} = await this.getPdfTitleAndAuthor(inpFile);
if (!title && uploadFileName)
title = uploadFileName;
//console.log(JSON.stringify(lines, null, 2));
//формируем текст
const limitSize = 2*this.config.maxUploadFileSize;
let text = '';
if (title)
text += `<fb2-title>${title}</fb2-title>`;
if (author)
text += `<fb2-author>${author}</fb2-author>`;
let concat = '';
let sp = '';
let firstLine = true;
for (const line of lines) {
if (text.length > limitSize) {
throw new Error(`Файл для конвертирования слишком большой|FORLOG| text.length: ${text.length} > ${limitSize}`);
}
if (line.isImage) {
text += `<fb2-image type="${line.type}" name="${line.name}">${line.data}</fb2-image>`;
continue;
}
if (line.isSectionTitle) {
if (firstLine)
text += `<fb2-section-title>${line.text.trim()}</fb2-section-title>`;
else
text += `<fb2-subtitle>${line.text.trim()}</fb2-subtitle>`;
continue;
}
firstLine = false;
if (line.isSubtitle) {
text += `<br><fb2-subtitle>${line.text.trim()}</fb2-subtitle>`;
continue;
}
if (concat == '') {
const left = line.left || 0;
sp = ' '.repeat(indents[left]);
}
let t = line.text.trim();
if (t.substr(-1) == '-') {
t = t.substr(0, t.length - 1);
concat += t;
} else {
text += sp + concat + t + "\n";
concat = '';
}
}
if (concat)
text += sp + concat + "\n";
//console.log(text);
await utils.sleep(100);
return await super.run(Buffer.from(text), {skipCheck: true, isText: true});
}
async getPdfTitleAndAuthor(pdfFile) {
const result = {author: '', title: ''};
const pdfinfoPath = '/usr/bin/pdfinfo';
if (!await fs.pathExists(pdfinfoPath))
throw new Error('Внешний конвертер pdfinfo не найден');
const execResult = await this.execConverter(pdfinfoPath, [pdfFile]);
const titlePrefix = 'Title:';
const authorPrefix = 'Author:';
const stdout = execResult.stdout.split("\n");
stdout.forEach(line => {
if (line.indexOf(titlePrefix) == 0)
result.title = line.substring(titlePrefix.length).trim();
if (line.indexOf(authorPrefix) == 0)
result.author = line.substring(authorPrefix.length).trim();
});
return result;
}
}
module.exports = ConvertPdf;