ayrisdev/lagos-back
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: resolve UDF extraction false positive rejection on real UDF content.xml files

Browse Source
This commit is contained in:
Mustafa Yildiz
2026-08-15 11:06:41 +03:00
parent 389bbe33ef
commit cab55c4fae
1 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/controllers/document.controller.ts
+3 -3
View File
@@ -490,9 +490,9 @@ export async function extractTextFromUploadedFile(file: { buffer: Buffer; mimety
.trim(); .trim();
const cleaned = sanitizePostgresText(plainText); const cleaned = sanitizePostgresText(plainText);
// Akıllı filtreleme: content.xml, content.xmlPK gibi ZIP başlık isimlerini metin sanma // Sadece tek başına "content.xml" veya "content.xmlPK" olan kısa çöp dizgileri süz (Gerçek metinleri kesinlikle engelleme)
const isZipNoise = /^content\.xml/i.test(cleaned) || cleaned.includes('content.xmlPK') || cleaned.includes('documentproperties.xml'); const isExactZipHeader = /^content\.xml(PK)?$/i.test(cleaned.trim()) && cleaned.length < 35;
if (cleaned.length > 5 && !cleaned.startsWith('PK') && !isZipNoise) { if (cleaned.length > 5 && !cleaned.startsWith('PK') && !isExactZipHeader) {
return cleaned; return cleaned;
} }
} }