fix: add readability fallback extraction
parent
0840029982
commit
1d862cf5c2
|
|
@ -81,6 +81,14 @@ export async function extractReadableContent(params: {
|
|||
url: string;
|
||||
extractMode: ExtractMode;
|
||||
}): Promise<{ text: string; title?: string } | null> {
|
||||
const fallback = (): { text: string; title?: string } => {
|
||||
const rendered = htmlToMarkdown(params.html);
|
||||
if (params.extractMode === "text") {
|
||||
const text = markdownToText(rendered.text) || normalizeWhitespace(stripTags(params.html));
|
||||
return { text, title: rendered.title };
|
||||
}
|
||||
return rendered;
|
||||
};
|
||||
try {
|
||||
const [{ Readability }, { parseHTML }] = await Promise.all([
|
||||
import("@mozilla/readability"),
|
||||
|
|
@ -94,15 +102,15 @@ export async function extractReadableContent(params: {
|
|||
}
|
||||
const reader = new Readability(document, { charThreshold: 0 });
|
||||
const parsed = reader.parse();
|
||||
if (!parsed?.content) return null;
|
||||
if (!parsed?.content) return fallback();
|
||||
const title = parsed.title || undefined;
|
||||
if (params.extractMode === "text") {
|
||||
const text = normalizeWhitespace(parsed.textContent ?? "");
|
||||
return { text, title };
|
||||
return text ? { text, title } : fallback();
|
||||
}
|
||||
const rendered = htmlToMarkdown(parsed.content);
|
||||
return { text: rendered.text, title: title ?? rendered.title };
|
||||
} catch {
|
||||
return null;
|
||||
return fallback();
|
||||
}
|
||||
}
|
||||
|
|
|
|||
Loading…
Reference in New Issue