72 lines
3.2 KiB
JavaScript
72 lines
3.2 KiB
JavaScript
import {awaitWithAbort} from './abortable.js';
|
||
import { readPdfPages } from './pdfDocumentReader.js';
|
||
|
||
async function openPdf(file,signal) {
|
||
signal?.throwIfAborted();
|
||
const pdfjs = await import('pdfjs-dist');
|
||
pdfjs.GlobalWorkerOptions.workerSrc = (await import('pdfjs-dist/build/pdf.worker.min.mjs?url')).default;
|
||
let task,destruction;
|
||
const destroy=()=>destruction ||= task?.destroy();
|
||
try {
|
||
const data=await awaitWithAbort(file.arrayBuffer(),signal);
|
||
signal?.throwIfAborted();
|
||
task = pdfjs.getDocument({data});
|
||
return {doc:await awaitWithAbort(task.promise,signal,destroy),task:{destroy}};
|
||
}
|
||
catch(error) {
|
||
await destroy();
|
||
signal?.throwIfAborted();
|
||
if(error.name === 'InvalidPDFException') throw new Error('PDF dosyası bozuk veya geçerli PDF biçiminde değil.');
|
||
if(error.name === 'PasswordException') throw new Error('PDF parola korumalı; okunabilir bir kopyasını yükleyin.');
|
||
throw new Error('PDF okunamadı. Dosyayı kontrol edip yeniden deneyin.');
|
||
}
|
||
}
|
||
|
||
async function readPdf(file,{ocr=false,forceOcr=false,maxPages=Infinity,onProgress,signal}={}) {
|
||
const {doc,task}=await openPdf(file,signal);
|
||
let worker,termination,render;
|
||
const stopWorker=()=>termination ||= worker?.terminate();
|
||
const cancel=()=>{render?.cancel();void Promise.resolve(stopWorker()).catch(()=>{});void Promise.resolve(task.destroy()).catch(()=>{});};
|
||
signal?.addEventListener('abort',cancel,{once:true});
|
||
try {
|
||
signal?.throwIfAborted();
|
||
const result=await readPdfPages(doc,{maxPages,forceOcr,onProgress,signal,ocrPage:ocr ? async(page)=>{
|
||
if(!worker) {
|
||
const {createWorker}=await import('tesseract.js');
|
||
signal?.throwIfAborted();
|
||
worker=await awaitWithAbort(createWorker('tur').then(created=>{
|
||
if(signal?.aborted){void created.terminate().catch(()=>{});signal.throwIfAborted();}
|
||
return created;
|
||
}),signal);
|
||
}
|
||
const viewport=page.getViewport({scale:2});
|
||
const canvas=document.createElement('canvas');
|
||
canvas.width=viewport.width;canvas.height=viewport.height;
|
||
try {
|
||
signal?.throwIfAborted();
|
||
render=page.render({canvasContext:canvas.getContext('2d'),viewport});
|
||
await awaitWithAbort(render.promise,signal,()=>render.cancel());
|
||
signal?.throwIfAborted();
|
||
return (await awaitWithAbort(worker.recognize(canvas),signal,stopWorker)).data.text;
|
||
} finally { render=null;canvas.width=0;canvas.height=0; }
|
||
} : undefined});
|
||
return result;
|
||
} finally {
|
||
signal?.removeEventListener('abort',cancel);
|
||
try { await stopWorker(); }
|
||
finally { await task.destroy(); }
|
||
}
|
||
}
|
||
|
||
export const extractPdfText=(file,options={})=>readPdf(file,{...options,ocr:false});
|
||
export const ocrPdfText=(file,options={})=>readPdf(file,{...options,ocr:true,forceOcr:true});
|
||
export async function extractPlainText(file,{signal}={}) {
|
||
signal?.throwIfAborted();
|
||
return {text:(await awaitWithAbort(file.text(),signal)).trim(),pageCount:1,usedPages:1,complete:true};
|
||
}
|
||
export async function extractText(file,{ocr=true,onProgress,signal}={}) {
|
||
if((file.name || '').toLowerCase().endsWith('.pdf') || file.type==='application/pdf')
|
||
return readPdf(file,{ocr,onProgress,signal});
|
||
return extractPlainText(file,{signal});
|
||
}
|