Merge pull request 'fix(sources): resolve arXiv IDs via arXiv DOI (OpenAlex /works/arxiv: 404s)' (#5) from fix/openalex-arxiv-doi into main

fix(sources): resolve arXiv IDs via arXiv DOI (OpenAlex /works/arxiv: 404s)

tests/sources/test_openalex.py

@@ -45,7 +45,18 @@ def test_resolve_doi() -> None:
 
 
 def test_resolve_arxiv() -> None:
-    assert _resolve_id("2301.07041") == "arxiv:2301.07041"
+    # arXiv IDs resolve to their arXiv DOI (the arxiv: path 404s on OpenAlex).
+    assert _resolve_id("2301.07041") == "doi:10.48550/arXiv.2301.07041"
+
+
+def test_resolve_arxiv_legacy() -> None:
+    # Legacy category/number IDs (pre-2007) resolve the same way.
+    assert _resolve_id("math/0603097") == "doi:10.48550/arXiv.math/0603097"
+
+
+def test_resolve_arxiv_prefix_stripped() -> None:
+    # An explicit arxiv: prefix is stripped before building the arXiv DOI.
+    assert _resolve_id("arxiv:2301.07041") == "doi:10.48550/arXiv.2301.07041"
 
 
 def test_resolve_w_id_unchanged() -> None:
@@ -54,7 +65,6 @@ def test_resolve_w_id_unchanged() -> None:
 
 def test_resolve_prefixed_unchanged() -> None:
     assert _resolve_id("doi:10.1145/x") == "doi:10.1145/x"
-    assert _resolve_id("arxiv:2301.07041") == "arxiv:2301.07041"
 
 
 # ---------------------------------------------------------------------------
@@ -83,7 +93,7 @@ def test_fetch_paper_success(monkeypatch: pytest.MonkeyPatch) -> None:
 
 
 def test_fetch_paper_arxiv_id_prefixed(monkeypatch: pytest.MonkeyPatch) -> None:
-    """Bare arXiv IDs are prefixed with 'arxiv:' in the URL."""
+    """Bare arXiv IDs are resolved to their arXiv DOI in the URL."""
     called_url: list[str] = []
 
     def mock_get(url: str, params: dict[str, str] | None = None) -> httpx.Response:
@@ -93,7 +103,7 @@ def test_fetch_paper_arxiv_id_prefixed(monkeypatch: pytest.MonkeyPatch) -> None:
     monkeypatch.setattr(openalex, "_get", mock_get)
     openalex.fetch_paper("2301.07041")
 
-    assert called_url and "arxiv:2301.07041" in called_url[0]
+    assert called_url and "doi:10.48550/arXiv.2301.07041" in called_url[0]
 
 
 # ---------------------------------------------------------------------------