GitHub-Proxy/jina_reader
1
0
Fork 0
mirror of https://git.mirrors.martin98.com/https://github.com/jina-ai/reader.git synced 2025-08-20 16:39:10 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix: pdf mode and google web cache

Browse Source
This commit is contained in:
yanlong.wang 2024-10-09 17:47:23 +08:00
parent f6bbddcb48
commit e9258af742
No known key found for this signature in database
GPG Key ID: C0A623C0BADF9F37
1 changed files with 3 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
backend/functions/src/services/snapshot-formatter.ts
View File

@ -136,7 +136,8 @@ export class SnapshotFormatter extends AsyncService {
} }
let pdfMode = false; let pdfMode = false;
if (snapshot.pdfs?.length && !snapshot.title) { // in case of Google Web Cache content
if (snapshot.pdfs?.length && (!snapshot.title || snapshot.title.startsWith('cache:'))) {
const pdf = await this.pdfExtractor.cachedExtract(snapshot.pdfs[0], const pdf = await this.pdfExtractor.cachedExtract(snapshot.pdfs[0],
this.threadLocal.get('cacheTolerance') this.threadLocal.get('cacheTolerance')
); );
@ -330,7 +331,7 @@ export class SnapshotFormatter extends AsyncService {
const n = code - 200; const n = code - 200;
if (n < 0 || n >= 200) { if (n < 0 || n >= 200) {
const text = snapshot.statusText || STATUS_CODES[code]; const text = snapshot.statusText || STATUS_CODES[code];
formatted.warning = `Target URL returned error ${code}${text? `: ${text}` : ''}`; formatted.warning = `Target URL returned error ${code}${text ? `: ${text}` : ''}`;
} }
} }