Merge pull request #137 from scipp/move-file-registyr

Download files from public dmsc server instead of github

Author: nvaytet

src/tof/facilities/__init__.py

@@ -8,17 +8,34 @@
 _source_library = {}
 _source_library.update(ess.sources)
 
-
-_source_registry = pooch.create(
-    path=pooch.os_cache("tof"),
-    base_url="https://github.com/scipp/tof-sources/raw/refs/heads/main/1/",
-    retry_if_failed=2,
-    registry={f["path"]: f["hash"] for f in _source_library.values()},
-)
+_BASE_URLS = [
+    "https://public.esss.dk/groups/scipp/tof/1/",  # primary: DMSC server
+    "https://github.com/scipp/tof-sources/raw/refs/heads/main/1/",  # fallback: GitHub
+]
+
+# One registry per mirror URL
+_registries = [
+    pooch.create(
+        path=pooch.os_cache("tof"),
+        base_url=base_url,
+        retry_if_failed=2,
+        registry={f["path"]: f["hash"] for f in _source_library.values()},
+    )
+    for base_url in _BASE_URLS
+]
 
 
 def get_source_path(name: str) -> str:
-    return _source_registry.fetch(_source_library[name]["path"])
+    path = _source_library[name]["path"]
+    last_exc = None
+    for registry in _registries:
+        try:
+            return registry.fetch(path)
+        except Exception as exc:
+            last_exc = exc
+    raise RuntimeError(
+        f"Failed to download '{path}' from all mirrors: {_BASE_URLS}"
+    ) from last_exc
 
 
 __all__ = ["ess", "get_source_path"]

tests/facilities/source_library_test.py

@@ -0,0 +1,31 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright (c) 2026 Scipp contributors (https://github.com/scipp)
+
+from pathlib import Path
+
+import pooch
+
+from tof.facilities import _BASE_URLS, _source_library
+
+
+def test_source_library_files_identical_on_public_and_github(tmp_path: Path) -> None:
+    registry = {f["path"]: f["hash"] for f in _source_library.values()}
+
+    public = pooch.create(
+        path=tmp_path / "cache_public", base_url=_BASE_URLS[0], registry=registry
+    )
+
+    gh = pooch.create(
+        path=tmp_path / "cache_github", base_url=_BASE_URLS[1], registry=registry
+    )
+
+    for entry in _source_library.values():
+        rel = entry["path"]
+
+        # Verify that hashes are the same in both registries.
+        assert public.registry[rel] == gh.registry[rel]
+
+        p_gh = Path(gh.fetch(rel))
+        p_public = Path(public.fetch(rel))
+
+        assert p_gh.name == p_public.name