feat(format): follow soft links; explicit errors for external links and external raw data

- Path resolution follows soft links in both old-style (symbol table, cache
  type 2) and new-style (compact and dense Link message) groups: absolute and
  relative targets, links to groups, links through links, with a depth limit
  so a link cycle is NestingDepthExceeded rather than a hang. A dangling link
  reports the target it could not find. Previously every soft link was
  PathNotFound.
- An external link is FormatError::ExternalLinkUnsupported { filename,
  object_path } instead of a misleading PathNotFound.
- Message 0x0007 (External Data Files) is now a known MessageType, and a
  dataset carrying it is FormatError::ExternalDataFilesUnsupported. Such a
  dataset has no data address in this file, so it would otherwise be read as
  "never written" and answered with fill values — wrong data, no error.
- Dense link iteration is shared between hard-link listing and the new
  symbolic-link lookup; entry listing behaviour is unchanged.
- h5py interop test for both libver settings.

Co-Authored-By: Claude Fable 5.1 <[email protected]>

crates/clawhdf5/tests/h5py_interop_tests.rs

@@ -699,3 +699,88 @@ with h5py.File("{path_str}", "r") as f:
         );
     }
 }
+
+// ---------------------------------------------------------------------------
+// h5py writes soft / external links and external raw data -> clawhdf5
+// ---------------------------------------------------------------------------
+
+/// Soft links are followed (absolute, relative, through groups, with a cycle
+/// guard). Things this reader does not follow — external links, and datasets
+/// whose raw data lives in another file — are explicit errors. They used to
+/// surface as a misleading `PathNotFound`, and external raw data could read
+/// back as fill values.
+#[test]
+fn h5py_links_clawhdf5_resolves_or_refuses() {
+    use clawhdf5::Error;
+    use clawhdf5_format::error::FormatError;
+
+    skip_if_no_python!();
+    let dir = tempfile::tempdir().unwrap();
+    let dir_str = dir.path().display().to_string();
+    for (tag, kwargs) in [("default", ""), ("latest", ", libver='latest'")] {
+        let script = format!(
+            r#"
+import h5py, numpy as np, os
+os.chdir("{dir_str}")
+with h5py.File("other_{tag}.h5", "w"{kwargs}) as o:
+    o.create_dataset("remote", data=np.arange(3, dtype="<i4"))
+with h5py.File("links_{tag}.h5", "w"{kwargs}) as f:
+    f.create_dataset("real", data=np.arange(4, dtype="<i4"))
+    g = f.create_group("grp")
+    g.create_dataset("inner", data=np.arange(2, dtype="<i4"))
+    g["rel"] = h5py.SoftLink("inner")
+    f["soft"] = h5py.SoftLink("/real")
+    f["soft_grp"] = h5py.SoftLink("/grp")
+    f["dangling"] = h5py.SoftLink("/nope")
+    f["loop_a"] = h5py.SoftLink("/loop_b")
+    f["loop_b"] = h5py.SoftLink("/loop_a")
+    f["ext"] = h5py.ExternalLink("other_{tag}.h5", "/remote")
+    f.create_dataset("extdata", shape=(4,), dtype="<i4", external=[("raw_{tag}.bin", 0, 16)])
+    f["extdata"][...] = np.array([11, 22, 33, 44], dtype="<i4")
+"#
+        );
+        run_python(&script);
+
+        let file = File::open(dir.path().join(format!("links_{tag}.h5"))).unwrap();
+        let read = |path: &str| file.dataset(path).and_then(|d| d.read_i32());
+        assert_eq!(read("soft").unwrap(), vec![0, 1, 2, 3], "{tag}");
+        assert_eq!(read("soft_grp/inner").unwrap(), vec![0, 1], "{tag}");
+        assert_eq!(
+            read("grp/rel").unwrap(),
+            vec![0, 1],
+            "{tag}: relative target"
+        );
+        assert_eq!(
+            read("soft_grp/rel").unwrap(),
+            vec![0, 1],
+            "{tag}: link via link"
+        );
+
+        assert!(
+            matches!(read("dangling"), Err(Error::Format(FormatError::PathNotFound(p))) if p == "nope"),
+            "{tag}: dangling link names its missing target"
+        );
+        assert!(
+            matches!(
+                read("loop_a"),
+                Err(Error::Format(FormatError::NestingDepthExceeded))
+            ),
+            "{tag}: link cycle"
+        );
+        assert!(
+            matches!(
+                read("ext"),
+                Err(Error::Format(FormatError::ExternalLinkUnsupported { ref object_path, .. }))
+                    if object_path == "/remote"
+            ),
+            "{tag}: external link"
+        );
+        assert!(
+            matches!(
+                read("extdata"),
+                Err(Error::Format(FormatError::ExternalDataFilesUnsupported))
+            ),
+            "{tag}: external raw data must not read as fill values"
+        );
+    }
+}