Merge pull request #1402 from guangyang1206/fix/extract-domain-helper-1368

Fix/extract domain helper 1368

surfsense_web/components/assistant-ui/assistant-message.tsx

@@ -24,6 +24,8 @@ import dynamic from "next/dynamic";
 import type { FC } from "react";
 import { useEffect, useMemo, useRef, useState } from "react";
 import { commentsEnabledAtom, targetCommentIdAtom } from "@/atoms/chat/current-thread.atom";
+import { tryGetHostname } from "@/lib/url";
+
 import {
 	globalNewLLMConfigsAtom,
 	newLLMConfigsAtom,
@@ -99,20 +101,12 @@ const GenerateImageToolUI = dynamic(
 		import("@/components/tool-ui/generate-image").then((m) => ({ default: m.GenerateImageToolUI })),
 	{ ssr: false }
 );
-function extractDomain(url: string): string | undefined {
-	try {
-		return new URL(url).hostname.replace(/^www\./, "");
-	} catch {
-		return undefined;
-	}
-}
-
 function useCitationsFromMetadata(): SerializableCitation[] {
 	const allCitations = useAllCitationMetadata();
 	return useMemo(() => {
 		const result: SerializableCitation[] = [];
 		for (const [url, meta] of allCitations) {
-			const domain = extractDomain(url);
+			const domain = tryGetHostname(url);
 			result.push({
 				id: `url-cite-${url}`,
 				href: url,

surfsense_web/components/assistant-ui/inline-citation.tsx

@@ -193,14 +193,7 @@ const SurfsenseDocCitation: FC<{ chunkId: number }> = ({ chunkId }) => {
 	);
 };
 
-function extractDomain(url: string): string {
+import { tryGetHostname } from "@/lib/url";
-	try {
-		const hostname = new URL(url).hostname;
-		return hostname.replace(/^www\./, "");
-	} catch {
-		return url;
-	}
-}
 
 interface UrlCitationProps {
 	url: string;
@@ -212,7 +205,7 @@ interface UrlCitationProps {
  * page title and snippet (extracted deterministically from web_search tool results).
  */
 export const UrlCitation: FC<UrlCitationProps> = ({ url }) => {
-	const domain = extractDomain(url);
+	const domain = tryGetHostname(url) ?? url;
 	const meta = useCitationMetadata(url);
 
 	return (

surfsense_web/components/assistant-ui/markdown-text.tsx

@@ -23,6 +23,8 @@ import "katex/dist/katex.min.css";
 import { toast } from "sonner";
 import { processChildrenWithCitations } from "@/components/citations/citation-renderer";
 import { Skeleton } from "@/components/ui/skeleton";
+import { tryGetHostname } from "@/lib/url";
+
 import {
 	Table,
 	TableBody,
@@ -139,15 +141,6 @@ const MarkdownTextImpl = () => {
 
 export const MarkdownText = memo(MarkdownTextImpl);
 
-function extractDomain(url: string): string {
-	try {
-		const parsed = new URL(url);
-		return parsed.hostname.replace(/^www\./, "");
-	} catch {
-		return "";
-	}
-}
-
 // Canonical local-file virtual paths are mount-prefixed: /<mount>/<relative/path>
 const LOCAL_FILE_PATH_REGEX = /^\/[a-z0-9_-]+\/[^\s`]+(?:\/[^\s`]+)*$/;
 
@@ -288,7 +281,7 @@ function FilePathLink({ path, className }: { path: string; className?: string })
 function MarkdownImage({ src, alt }: { src?: string; alt?: string }) {
 	if (!src) return null;
 
-	const domain = extractDomain(src);
+	const domain = tryGetHostname(src) ?? "";
 
 	return (
 		<div className="my-4 w-fit max-w-lg overflow-hidden rounded-2xl border bg-muted/30 select-none">

surfsense_web/components/tool-ui/citation/citation.tsx

@@ -6,19 +6,11 @@ import * as React from "react";
 import { openSafeNavigationHref, sanitizeHref } from "../shared/media";
 import { cn, Popover, PopoverContent, PopoverTrigger } from "./_adapter";
 import type { CitationVariant, SerializableCitation } from "./schema";
+import { tryGetHostname } from "@/lib/url";
 import { TYPE_ICONS } from "./type-icons";
 
 const FALLBACK_LOCALE = "en-US";
 
-function extractDomain(url: string): string | undefined {
-	try {
-		const urlObj = new URL(url);
-		return urlObj.hostname.replace(/^www\./, "");
-	} catch {
-		return undefined;
-	}
-}
-
 function formatDate(isoString: string, locale: string): string {
 	try {
 		const date = new Date(isoString);
@@ -78,7 +70,7 @@ export function Citation(props: CitationProps) {
 
 	const locale = providedLocale ?? FALLBACK_LOCALE;
 	const sanitizedHref = sanitizeHref(rawHref);
-	const domain = providedDomain ?? extractDomain(rawHref);
+	const domain = providedDomain ?? tryGetHostname(rawHref);
 
 	const citationData: SerializableCitation = {
 		...serializable,

surfsense_web/lib/url.ts

@@ -0,0 +1,14 @@
+/**
+ * Extract a normalized hostname from a URL. Strips a leading `www.`.
+ * Returns `undefined` if the input is not a parseable URL.
+ *
+ * This is the canonical replacement for the four previously-duplicated
+ * `extractDomain` helpers that had subtly different error fallbacks.
+ */
+export function tryGetHostname(url: string): string | undefined {
+    try {
+        return new URL(url).hostname.replace(/^www\./, "");
+    } catch {
+        return undefined;
+    }
+}