GitHub-Proxy/jina_reader
1
0
Fork 0
mirror of https://git.mirrors.martin98.com/https://github.com/jina-ai/reader.git synced 2025-08-19 16:19:06 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix: root content

Browse Source
This commit is contained in:
yanlong.wang 2024-04-25 15:43:17 +08:00
parent f1016649ac
commit 39a446f5e7
No known key found for this signature in database
GPG Key ID: C0A623C0BADF9F37
2 changed files with 16 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
backend/functions/src/cloud-functions/crawler.ts
View File

@ -263,6 +263,7 @@ ${this.content}
tags: ['Crawler'], tags: ['Crawler'],
httpMethod: ['get', 'post'], httpMethod: ['get', 'post'],
returnType: [String, OutputServerEventStream], returnType: [String, OutputServerEventStream],
exposeRoot: true,
}) })
async crawl( async crawl(
@RPCReflect() rpcReflect: RPCReflection, @RPCReflect() rpcReflect: RPCReflection,
@ -272,21 +273,29 @@ ${this.content}
}, },
) { ) {
const noSlashURL = ctx.req.url.slice(1); const noSlashURL = ctx.req.url.slice(1);
if (!noSlashURL) {
return assignTransferProtocolMeta(`[Usage] https://r.jina.ai/YOUR_URL
[Homepage] https://jina.ai/reader
[Source code] https://github.com/jina-ai/reader
`,
{ contentType: 'text/plain', envelope: null }
);
}
let urlToCrawl; let urlToCrawl;
try { try {
urlToCrawl = new URL(normalizeUrl(noSlashURL.trim(), { stripWWW: false, removeTrailingSlash: false, removeSingleSlash: false })); urlToCrawl = new URL(normalizeUrl(noSlashURL.trim(), { stripWWW: false, removeTrailingSlash: false, removeSingleSlash: false }));
if (urlToCrawl.protocol !== 'http:' && urlToCrawl.protocol !== 'https:') {
throw new ParamValidationError({
message: `Invalid protocol ${urlToCrawl.protocol}`,
path: 'url'
});
}
} catch (err) { } catch (err) {
throw new ParamValidationError({ throw new ParamValidationError({
message: `${err}`, message: `${err}`,
path: 'url' path: 'url'
}); });
} }
if (urlToCrawl.protocol !== 'http:' && urlToCrawl.protocol !== 'https:') {
throw new ParamValidationError({
message: `Invalid protocol ${urlToCrawl.protocol}`,
path: 'url'
});
}
const customMode = ctx.req.get('x-respond-with') || 'markdown'; const customMode = ctx.req.get('x-respond-with') || 'markdown';
const noCache = Boolean(ctx.req.get('x-no-cache')); const noCache = Boolean(ctx.req.get('x-no-cache'));

2
thinapps-shared

@ -1 +1 @@
Subproject commit 64157bc57ef9ce2cec69f37b5f55fccb71742b6f Subproject commit 1346b514f3a8d183922bf9257a7f52446a581ac0