Merge pull request #66 from YosefLab/copy-speedup

better performance with many trees

Author: colganwi

CHANGELOG.md

@@ -16,6 +16,19 @@ and this project adheres to [Semantic Versioning][].
 
 ### Fixed
 
+## [0.2.3] - 2025-10-14
+
+### Added
+
+### Changed
+
+- Optimized tree overlap detection to speed up copying and subsetting with many trees (#66)
+- Switch to hatch with v0.0.6 template update (#61)
+
+### Fixed
+
+- Fixed codecov configuration (#64)
+
 ## [0.2.2] - 2025-09-18
 
 ### Added

pyproject.toml

@@ -4,7 +4,7 @@ requires = [ "hatchling" ]
 
 [project]
 name = "treedata"
-version = "0.2.2"
+version = "0.2.3"
 description = "anndata with trees"
 readme = "README.md"
 license = { file = "LICENSE" }

src/treedata/_core/aligned_mapping.py

@@ -81,13 +81,9 @@ def _validate_tree(self, tree: nx.DiGraph, key: str) -> set[str]:
 
     def _check_tree_overlap(self, nodes=()) -> bool:
         """Check if the leaves overlap with other trees."""
-        nodes = set(nodes)
-        for _, tree_nodes in self._tree_to_node.items():
-            if nodes.intersection(tree_nodes):
-                return True
-            else:
-                nodes = nodes.union(tree_nodes)
-        return False
+        if not nodes:
+            return any(len(tree_keys) > 1 for tree_keys in self._node_to_tree.values())
+        return any(node in self._node_to_tree for node in nodes)
 
     def _update_tree_labels(self):
         """Update the tree labels in the parent object."""
@@ -176,6 +172,12 @@ def __setitem__(self, key: str, value: nx.DiGraph):
         self._check_uniqueness()
         nodes = self._validate_tree(value, key)
 
+        if key in self._tree_to_node:
+            for node in self._tree_to_node[key]:
+                self._node_to_tree[node].discard(key)
+                if not self._node_to_tree[node]:
+                    del self._node_to_tree[node]
+
         for node in nodes:
             self._node_to_tree[node].add(key)
         self._tree_to_node[key] = nodes