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 13:09:08 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix: turndown rules

Browse Source
This commit is contained in:
yanlong.wang 2024-10-31 17:22:51 +08:00
parent 340fb517d8
commit b10931b8ed
No known key found for this signature in database
GPG Key ID: C0A623C0BADF9F37
1 changed files with 4 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
backend/functions/src/services/snapshot-formatter.ts
View File

@ -552,25 +552,10 @@ ${suffixMixins.length ? `\n${suffixMixins.join('\n\n')}\n` : ''}`;
return delimiter + extraSpace + content + (delimiter === '```' && !content.endsWith(extraSpace) ? extraSpace : '') + delimiter;
}
});
turnDownService.addRule('flattened-tables', {
filter: (node) => {
if (node.tagName !== 'TABLE') {
return false;
}
let parentHasTable = false;
let ptr = node.parentElement;
while (ptr) {
if (ptr.tagName === 'TABLE') {
parentHasTable = true;
break;
}
ptr = ptr.parentElement;
}
return parentHasTable;
},
replacement: (innerText) => {
return innerText.trim();
turnDownService.addRule('picture', {
filter: 'picture',
replacement: (content, _node) => {
return content;
}
});