add: support for db urls in search

Author: amkhrjee

utils/accessionLinks.test.ts

@@ -3,6 +3,7 @@ import {
   getExternalArchiveUrl,
   getInternalUrl,
   parseAccessions,
+  isAccessionUrl,
   startsWithAccession,
 } from "./accessionLinks";
 
@@ -161,3 +162,44 @@ describe("parseAccessions", () => {
     expect(parseAccessions("brain single cell rna")).toEqual([]);
   });
 });
+
+describe("isAccessionUrl", () => {
+  it("recognizes pasted archive URLs and finds the accession inside", () => {
+    const cases: [string, string][] = [
+      [
+        "http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE317357",
+        "/p/GSE317357",
+      ],
+      ["https://www.ncbi.nlm.nih.gov/sra/?term=SRX1234567", "/e/SRX1234567"],
+      ["https://www.ebi.ac.uk/ena/browser/view/SRP123456", "/p/SRP123456"],
+      [
+        "https://www.ebi.ac.uk/biostudies/arrayexpress/studies/E-MTAB-1234",
+        "/p/E-MTAB-1234",
+      ],
+      ["https://ddbj.nig.ac.jp/search/entry/gea/E-GEAD-282", "/p/E-GEAD-282"],
+      ["https://ngdc.cncb.ac.cn/gsa/browse/CRA000004", "/p/CRA000004"],
+      ["www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSM7", "/s/GSM7"],
+    ];
+    for (const [url, expected] of cases) {
+      expect(isAccessionUrl(url)).toBe(true);
+      expect(parseAccessions(url)[0]?.url).toBe(expected);
+    }
+  });
+
+  it("keeps PRJ/submission URLs on their async-resolve path", () => {
+    const prj = parseAccessions(
+      "https://www.ebi.ac.uk/ena/browser/view/PRJEB12345",
+    )[0];
+    expect(isAccessionUrl("https://www.ebi.ac.uk/ena/browser/view/PRJEB12345"))
+      .toBe(true);
+    expect(prj.isPrj).toBe(true);
+  });
+
+  it("ignores non-URLs and URLs with no accession", () => {
+    expect(isAccessionUrl("https://seqout.org/about")).toBe(false);
+    expect(isAccessionUrl("https://github.com/some/repo")).toBe(false);
+    // Not a URL: already handled by startsWithAccession, not this.
+    expect(isAccessionUrl("GSE317357")).toBe(false);
+    expect(isAccessionUrl("cancer single cell")).toBe(false);
+  });
+});

utils/accessionLinks.ts

@@ -109,6 +109,17 @@ export function startsWithAccession(query: string): boolean {
   return ACC_ANCHORED.test(query.trim());
 }
 
+// A pasted archive URL — ".../acc.cgi?acc=GSE317357", ".../ena/browser/view/SRP12",
+// ".../arrayexpress/experiments/E-MTAB-1234" — is an accession the user wants to
+// open, not a phrase to full-text search. The accession sits mid-string, so
+// startsWithAccession can't see it. Any host counts: wherever the link came from,
+// the useful answer is the project page. Takes the first accession in the URL.
+export function isAccessionUrl(query: string): boolean {
+  const trimmed = query.trim();
+  if (!/^(?:https?:\/\/|www\.)/i.test(trimmed)) return false;
+  return parseAccessions(trimmed).length > 0;
+}
+
 export type ExternalArchive = { url: string; archive: string; label: string };
 
 export function getExternalArchiveUrl(

utils/useSearchHistory.ts

@@ -1,6 +1,10 @@
 import { useState } from "react";
 import { SERVER_URL } from "./constants";
-import { parseAccessions, startsWithAccession } from "./accessionLinks";
+import {
+  isAccessionUrl,
+  parseAccessions,
+  startsWithAccession,
+} from "./accessionLinks";
 import { getProjectShortUrl } from "./shortUrl";
 
 const HISTORY_KEY = "searchHistory";
@@ -57,10 +61,10 @@ export function useSearchHistory() {
     );
     saveHistory(newHistory);
 
-    // "<accession> ..." (optionally with a pasted title/notes) → jump to the
-    // first recognized accession instead of full-text search. Any further
-    // accessions in the text are ignored.
-    if (startsWithAccession(trimmed)) {
+    // "<accession> ..." (optionally with a pasted title/notes), or an archive URL
+    // with the accession inside it → jump to the first recognized accession
+    // instead of full-text search. Any further accessions in the text are ignored.
+    if (startsWithAccession(trimmed) || isAccessionUrl(trimmed)) {
       const first = parseAccessions(trimmed)[0];
       if (first) {
         if (first.isPrj) {