CVBora / utils /fileParser.js
ismizo's picture
Upload 11 files
5fd404e verified
Raw History Blame Contribute Delete
2.22 kB
// Ensure the require statement for pdf-parse is correct
const {PDFParse} = require('pdf-parse');
const mammoth = require('mammoth');
const fs = require('fs');
const path = require('path');
const os = require('os');
/**
* Extracts text from a file buffer based on mimetype
* @param {Buffer} buffer - The file data in memory
* @param {String} mimetype - The type of file (pdf/docx)
* @returns {Promise<String>} - The extracted text
*/
const extractText = async (buffer, mimetype) => {
try {
if (mimetype === 'application/pdf') {
// Create a path in the system's temp directory
const tempFilePath = path.join(os.tmpdir(), `upload_${Date.now()}.pdf`);
// Write the buffer to the temp file
fs.writeFileSync(tempFilePath, buffer);
// Convert file path to a file:// URL
const url = `file://${tempFilePath}`;
const parser = new PDFParse({url});
//const data = await parser.getText();
const allImages = await parser.getScreenshot({ scale: 1.69 });; // Get images to trigger parsing and text extraction 9
fs.unlinkSync(tempFilePath); // Clean up temp file after processing
const data = { init: true, pageImages:[]};
for (const page of allImages.pages) {
//const buffer = Buffer.from(page.dataUrl, 'base64');
data.pageImages.push(page.dataUrl); // Store image data for each page
}
return data;
}
else if (mimetype === 'application/vnd.openxmlformats-officedocument.wordprocessingml.document') { // .docx
const result = await mammoth.extractRawText({ buffer: buffer });
return result.value;
}
else if (mimetype === 'text/plain') {
return buffer.toString('utf8');
}
else {
throw new Error('Unsupported file type');
}
} catch (error) {
console.error("Text Extraction Error:", error);
throw new Error('Failed to parse file content');
}
};
module.exports = { extractText };