portolan-python 0.1.1__tar.gz → 0.1.2__tar.gz

{portolan_python-0.1.1 → portolan_python-0.1.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: portolan-python
-Version: 0.1.1
+Version: 0.1.2
 Summary: A lightweight Python implementation of the Portolan specification.
 Author: Portolan contributors
 License: Apache-2.0

{portolan_python-0.1.1 → portolan_python-0.1.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "portolan-python"
-version = "0.1.1"
+version = "0.1.2"
 description = "A lightweight Python implementation of the Portolan specification."
 readme = "README.md"
 requires-python = ">=3.10"

{portolan_python-0.1.1 → portolan_python-0.1.2}/src/portolan/registry.py

@@ -106,7 +106,12 @@ def _write_catalog_tree(
     root_url: str,
     output_root: Path,
     fetch_json: Callable[[str], JsonObject],
+    visited: set[str] | None = None,
 ) -> None:
+    if visited is None:
+        visited = set()
+    visited.add(document_url)
+
     relative_path = _relative_document_path(root_url, document_url)
     target = output_root / relative_path
     target.parent.mkdir(parents=True, exist_ok=True)
@@ -121,9 +126,11 @@ def _write_catalog_tree(
         if not isinstance(href, str):
             continue
         child_url = urljoin(document_url, href)
+        if child_url in visited:
+            continue
         child = fetch_json(child_url)
         if child.get("type") in {"Catalog", "Collection"}:
-            _write_catalog_tree(child_url, child, root_url, output_root, fetch_json)
+            _write_catalog_tree(child_url, child, root_url, output_root, fetch_json, visited)
 
 
 def _with_absolute_asset_hrefs(document_url: str, collection: JsonObject) -> JsonObject:

{portolan_python-0.1.1 → portolan_python-0.1.2}/tests/test_registry_api.py

@@ -169,6 +169,33 @@ def test_download_registry_catalog_writes_snapshot_with_absolute_asset_hrefs(
     assert collection["assets"]["data"]["href"] == "https://example.test/demo/roads/roads.parquet"
 
 
+def test_download_registry_catalog_skips_child_cycles(tmp_path: Path) -> None:
+    root_url = "https://example.test/demo/catalog.json"
+    child_url = "https://example.test/demo/roads/collection.json"
+    responses = {
+        root_url: {
+            "type": "Catalog",
+            "id": "demo",
+            "links": [{"rel": "child", "href": "./roads/collection.json"}],
+        },
+        child_url: {
+            **_collection("roads", {}),
+            "links": [{"rel": "child", "href": "../catalog.json"}],
+        },
+    }
+    fetched_urls: list[str] = []
+
+    def fetch_json(url: str) -> dict[str, Any]:
+        fetched_urls.append(url)
+        if fetched_urls.count(url) > 1:
+            raise AssertionError(f"Fetched document twice: {url}")
+        return responses[url]
+
+    download_registry_catalog(root_url, tmp_path, fetch_json=fetch_json)
+
+    assert fetched_urls == [root_url, child_url]
+
+
 def test_download_registry_catalog_recurses_nested_catalogs_and_uses_fallback_id(
     tmp_path: Path,
 ) -> None:

{portolan_python-0.1.1 → portolan_python-0.1.2}/uv.lock

@@ -468,7 +468,7 @@ wheels = [
 
 [[package]]
 name = "portolan-python"
-version = "0.1.1"
+version = "0.1.2"
 source = { editable = "." }
 
 [package.optional-dependencies]