vynte-infra/plane
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[SILO-506] fix: notion image parser extension (#4125)

Browse Source 
* fix: notion image parser extension

* fix: formatting for file parser extension
This commit is contained in:
Henit Chobisa
2025-09-04 18:53:22 +05:30
committed by GitHub
parent f0d23f4a38
commit 59205bf37f
1 changed files with 20 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files
apps/silo/src/apps/notion-importer/drivers/common/content-parser/extensions/process-files.ts
+20 -10
View File
@@ -16,7 +16,7 @@ export interface NotionFileParserConfig {
}
export class NotionFileParserExtension implements IParserExtension {
constructor(readonly config: NotionFileParserConfig) {}
constructor(readonly config: NotionFileParserConfig) { }
shouldParse(node: HTMLElement): boolean {
/*
@@ -30,15 +30,9 @@ export class NotionFileParserExtension implements IParserExtension {
*/
// Only process anchor tags that are likely local file links
const hasAnchorTag = node.tagName === "A" || node.rawTagName === "a";
const hasSingleImageChild = node.querySelector("img");
if (hasAnchorTag && hasSingleImageChild) {
return false;
}
const isNotAnHTTPLink = !node.getAttribute("href")?.startsWith("http");
const isNotAnHTMLPage = !node.getAttribute("href")?.endsWith(".html");
return hasAnchorTag && isNotAnHTTPLink && isNotAnHTMLPage;
const isImage = this.isImage(node.getAttribute("href") || "");
const isFile = this.isNonHTMLFile(node.getAttribute("href") || "");
return hasAnchorTag && isFile && !isImage;
}
async mutate(node: HTMLElement): Promise<HTMLElement> {
@@ -70,6 +64,22 @@ export class NotionFileParserExtension implements IParserExtension {
return node;
}
protected isImage(href: string): boolean {
return (
href?.endsWith(".png") ||
href?.endsWith(".jpg") ||
href?.endsWith(".jpeg") ||
href?.endsWith(".gif") ||
href?.endsWith(".svg") ||
href?.endsWith(".webp")
);
}
protected isNonHTMLFile(href: string): boolean {
const parts = href.split(".");
return !href.startsWith("http") && parts.length > 1 && parts[parts.length - 1] !== "html";
}
protected getFileSource(node: HTMLElement): string {
const href = node.getAttribute("href");
return href || "";