Javascript 从PDF中提取字体名称
我正在使用pdf.js从pdf中提取文本,但字体名称显示为g_d0_f6等。我需要字体名称以使用适当的表转换为Unicode。以下是从pdf2svg.js示例派生的代码:-Javascript 从PDF中提取字体名称,javascript,pdf,pdf.js,Javascript,Pdf,Pdf.js,我正在使用pdf.js从pdf中提取文本,但字体名称显示为g_d0_f6等。我需要字体名称以使用适当的表转换为Unicode。以下是从pdf2svg.js示例派生的代码:- var fs = require('fs'); var util = require('util'); var path = require('path'); var stream = require('stream'); // HACK few hacks to let PDF.js be loaded not as a
var fs = require('fs');
var util = require('util');
var path = require('path');
var stream = require('stream');
// HACK few hacks to let PDF.js be loaded not as a module in global space.
require('./domstubs.js').setStubs(global);
var pdfjsLib = require('pdfjs-dist');
var pdfPath = process.argv[2] || '../../web/compressed.tracemonkey-pldi-09.pdf';
var data = new Uint8Array(fs.readFileSync(pdfPath));
var loadingTask = pdfjsLib.getDocument({
data: data,
nativeImageDecoderSupport: pdfjsLib.NativeImageDecoding.DISPLAY,
});
loadingTask.promise.then(function(doc) {
var lastPromise = Promise.resolve(); // will be used to chain promises
var loadPage = function (pageNum) {
return doc.getPage(pageNum).then(function (page) {
return page.getTextContent().then(function (textContent) {
console.log(textContent);
});
});
};
for (var i = 1; i <= doc.numPages; i++) {
lastPromise = lastPromise.then(loadPage.bind(null, i));
}
return lastPromise;
}).then(function () {
console.log('# End of Document');
}, function (err) {
console.error('Error: ' + err);
});
样本输出:-
{ items:
[ { str: 'bl fp=k osQ ckjs esa cPpksa ls ckrphr djsa & ;g LowQy esa fdl le; dk n`\'; gS\\ cPps',
dir: 'ltr',
width: 396.2250000000001,
height: 15,
transform: [Array],
fontName: 'g_d0_f1' },
{ str: 'D;k dj jgs gSa\\ cPps dkSu&dkSu ls [ksy] [ksy j',
dir: 'ltr',
width: 216.1650000000001,
height: 15,
transform: [Array],
fontName: 'g_d0_f1' },
{ str: 'g',
dir: 'ltr',
width: 6.42,
height: 15,
transform: [Array],
fontName: 'g_d0_f1' },
{ str: 's gSa\\ fp=k esa fdrus cPps gSa vkSj fdrus',
dir: 'ltr',
width: 173.865,
height: 15,
transform: [Array],
fontName: 'g_d0_f1' },
{ str: 'cM+s gSa\\ vkil esa dkSu D;k ckr dj jgk gksxk\\ cPpksa ls fp=k esa lcosQ fy, uke lkspus',
dir: 'ltr',
width: 396.54000000000013,
height: 15,
transform: [Array],
fontName: 'g_d0_f1' },
{ str: 'dks dgasaA',
dir: 'ltr',
width: 40.74,
height: 15,
transform: [Array],
fontName: 'g_d0_f1' },
{ str: 'csVh cpkvks',
dir: 'ltr',
width: 66.725,
height: 17,
transform: [Array],
fontName: 'g_d0_f2' },
{ str: 'csVh i<+kvksA',
dir: 'ltr',
width: 66.75899999999999,
height: 17,
transform: [Array],
fontName: 'g_d0_f2' },
{ str: '2018-19',
dir: 'ltr',
width: 36.690000000000005,
height: 10,
transform: [Array],
fontName: 'g_d0_f3' } ],
styles:
{ g_d0_f1:
{ fontFamily: 'sans-serif',
ascent: 0.837,
descent: -0.216,
vertical: false },
g_d0_f2:
{ fontFamily: 'sans-serif',
ascent: 0.786,
descent: -0.181,
vertical: false },
g_d0_f3:
{ fontFamily: 'sans-serif',
ascent: 0.9052734375,
descent: -0.2119140625,
vertical: false } } }
以下是使用嵌入式字体的pdf:
这里有一个相关的问题,但建议的commonObjs是空的:我想你的思路是对的:page.commonObjs是找到实际字体名称的地方。但是,page.commonObjs仅在访问页面的文本/运算符时才会填充,因此如果在访问之前查看,您会发现它是空的 注意:下面的答案与pdf.js没有任何关系,但是它回答了这个问题,从pdf中提取字体名称 我还没有找到解决方案,所以我继续抓取,它有下面的命令来获取每页的字体信息
mutool info -F input.pdf 0-2147483647
然后我抓取了spawn函数,通过一些正则表达式和模式匹配对输出进行黑客攻击,以返回数据
const extractFontData = async str => {
const getMatches = str => {
const regex = /Page (\d+):\nFonts \((\d+)\):/;
const match = str.match(regex);
if (match) {
return { page: match[1], fonts: match[2] };
}
return {};
};
const singleFont = fontData => {
const match = fontData.match(/\+([a-zA-Z0-9_-]+[.,]?[a-zA-Z0-9_-]+)/);
return match && match[1];
};
return str
.split("Page ")
.map(singlePageData => {
const { page, fonts } = getMatches(`Page ` + singlePageData);
if (fonts) {
const split = singlePageData.split("\n").filter(e => e.length);
const fontList = split.slice(2).map(singleFont);
return { page, fonts, fontList };
}
})
.filter(e => e);
};
// Taken and adjusted from: https://stackoverflow.com/a/52611536/6161265
function run(...cmd) {
return new Promise((resolve, reject) => {
var { spawn } = require("child_process");
var command = spawn(...cmd);
var result = "";
command.stdout.on("data", function(data) {
result += data.toString();
});
command.on("close", function(code) {
resolve(result);
});
command.on("error", function(err) {
reject(err);
});
});
}
async function wrapper(filePath) {
const data = await run("mutool", ["info", "-F", filePath, "0-2147483647"]);
return extractFontData(data);
}
示例用法:
wrapper("ahhn101.pdf").then(data => console.log(data));
结果:
我需要字体名称来使用适当的表来转换为Unicode-一般来说,人们根本无法从字体名称中确定。我不高兴我现在没有足够的时间注释掉代码。唯一的建议可能是,如果您不是作为主机服务的一部分运行此服务,请使用spawnSync,因为它比span和收集I/O更容易。此外,可能希望收集所有数据作为缓冲区,附加到末尾,然后将整个结果串起来。感谢您的建议,我计划重写、注释和优化代码,这只是一个概念证明谢谢我一直在寻找任何给定文本的字体名称,但这是一个很好的起点,并回答了问题。如何访问文本/运算符?调用page.getOperatorList并等待解决,然后可以使用page.commonObjs。