Fix BatchStore metadata preservation paths

Preserve user vlmeta when BatchStore recreates its empty backing
SChunk during initial layout inference, avoid persisting empty
batch_lengths metadata that breaks vlmeta.getall() on empty stores,
and keep user meta/vlmeta when copy(storage=...) is used.

Add BatchStore regression tests covering:
- vlmeta preservation during inferred layout initialization
- clear()/delete-last on empty stores
- metadata preservation on copy(storage=...)

Author: FrancescAlted

src/blosc2/batch_store.py

@@ -370,6 +370,10 @@ def _load_batch_lengths(self) -> list[int] | None:
     def _persist_batch_lengths(self) -> None:
         if self._batch_lengths is None:
             return
+        if len(self._batch_lengths) == 0:
+            if _BATCHSTORE_VLMETA_KEY in self.vlmeta:
+                del self.vlmeta[_BATCHSTORE_VLMETA_KEY]
+            return
         self.schunk.vlmeta[_BATCHSTORE_VLMETA_KEY] = {"batch_lengths": list(self._batch_lengths)}
 
     def _get_batch_lengths(self) -> list[int] | None:
@@ -542,6 +546,7 @@ def _persist_layout_metadata(self) -> None:
         if len(self) > 0:
             return
         batch_lengths = None if self._batch_lengths is None else list(self._batch_lengths)
+        user_vlmeta = self._user_vlmeta_items() if len(self.vlmeta) > 0 else {}
         storage = self._make_storage()
         fixed_meta = dict(storage.meta or {})
         fixed_meta["batchstore"] = {
@@ -559,6 +564,8 @@ def _persist_layout_metadata(self) -> None:
             storage=storage,
         )
         self._attach_schunk(schunk)
+        for key, value in user_vlmeta.items():
+            self.vlmeta[key] = value
         if batch_lengths is not None and self._batch_lengths is None:
             self._batch_lengths = batch_lengths
 
@@ -894,17 +901,24 @@ def copy(self, **kwargs: Any) -> BatchStore:
         kwargs["dparams"] = kwargs.get("dparams", copy.deepcopy(self.dparams))
         kwargs["max_blocksize"] = kwargs.get("max_blocksize", self.max_blocksize)
         kwargs["serializer"] = kwargs.get("serializer", self.serializer)
-
-        if "storage" not in kwargs:
+        user_vlmeta = self._user_vlmeta_items() if len(self.vlmeta) > 0 else {}
+
+        if "storage" in kwargs:
+            storage = self._coerce_storage(kwargs["storage"], {})
+            fixed_meta = self._copy_meta()
+            if storage.meta is not None:
+                fixed_meta.update(storage.meta)
+            storage.meta = fixed_meta
+            kwargs["storage"] = storage
+        else:
             kwargs["meta"] = self._copy_meta()
             kwargs["contiguous"] = kwargs.get("contiguous", self.schunk.contiguous)
             if "urlpath" in kwargs and "mode" not in kwargs:
                 kwargs["mode"] = "w"
 
         out = BatchStore(**kwargs)
-        if "storage" not in kwargs and len(self.vlmeta) > 0:
-            for key, value in self._user_vlmeta_items().items():
-                out.vlmeta[key] = value
+        for key, value in user_vlmeta.items():
+            out.vlmeta[key] = value
         out.extend(self)
         return out
 

tests/test_batch_store.py

@@ -143,6 +143,26 @@ def test_batchstore_arrow_ipc_roundtrip():
     blosc2.remove_urlpath(urlpath)
 
 
+def test_batchstore_inferred_layout_preserves_user_vlmeta():
+    barray = blosc2.BatchStore()
+    barray.vlmeta["user"] = {"x": 1}
+
+    barray.append([1, 2, 3])
+
+    assert barray.vlmeta["user"] == {"x": 1}
+
+
+def test_batchstore_arrow_layout_persistence_preserves_user_vlmeta():
+    pa = pytest.importorskip("pyarrow")
+
+    barray = blosc2.BatchStore(serializer="arrow")
+    barray.vlmeta["user"] = {"x": 1}
+
+    barray.append(pa.array([[1], [2, 3]]))
+
+    assert barray.vlmeta["user"] == {"x": 1}
+
+
 def test_batchstore_from_cframe():
     barray = blosc2.BatchStore()
     barray.extend(BATCHES)
@@ -233,6 +253,38 @@ def test_batchstore_pop_keeps_batch_lengths_metadata_in_sync():
     assert items["nbatches"].startswith("2 (items per batch: mean=2.00")
 
 
+def test_batchstore_clear_keeps_empty_store_vlmeta_readable():
+    urlpath = "test_batchstore_clear_empty_vlmeta.b2b"
+    blosc2.remove_urlpath(urlpath)
+
+    barray = blosc2.BatchStore(urlpath=urlpath, mode="w", contiguous=True)
+    barray.append([1, 2, 3])
+    barray.clear()
+
+    assert barray.vlmeta.getall() == {}
+
+    reopened = blosc2.open(urlpath, mode="r")
+    assert reopened.vlmeta.getall() == {}
+
+    blosc2.remove_urlpath(urlpath)
+
+
+def test_batchstore_delete_last_keeps_empty_store_vlmeta_readable():
+    urlpath = "test_batchstore_delete_last_empty_vlmeta.b2b"
+    blosc2.remove_urlpath(urlpath)
+
+    barray = blosc2.BatchStore(urlpath=urlpath, mode="w", contiguous=True)
+    barray.append([1, 2, 3])
+    barray.delete(0)
+
+    assert barray.vlmeta.getall() == {}
+
+    reopened = blosc2.open(urlpath, mode="r")
+    assert reopened.vlmeta.getall() == {}
+
+    blosc2.remove_urlpath(urlpath)
+
+
 def test_batchstore_zstd_does_not_use_dict_by_default():
     barray = blosc2.BatchStore()
     assert barray.cparams.codec == blosc2.Codec.ZSTD
@@ -573,6 +625,26 @@ def test_batchstore_copy():
     blosc2.remove_urlpath(copy_path)
 
 
+def test_batchstore_copy_with_storage_preserves_user_metadata():
+    urlpath = "test_batchstore_copy_storage.b2b"
+    copy_path = "test_batchstore_copy_storage_out.b2b"
+    blosc2.remove_urlpath(urlpath)
+    blosc2.remove_urlpath(copy_path)
+
+    original = blosc2.BatchStore(urlpath=urlpath, mode="w", contiguous=True, meta={"user_meta": {"a": 1}})
+    original.vlmeta["user_vlmeta"] = {"b": 2}
+    original.extend(BATCHES)
+
+    copied = original.copy(storage=blosc2.Storage(contiguous=False, urlpath=copy_path, mode="w"))
+
+    assert [batch[:] for batch in copied] == [batch[:] for batch in original]
+    assert copied.meta["user_meta"] == {"a": 1}
+    assert copied.vlmeta["user_vlmeta"] == {"b": 2}
+
+    blosc2.remove_urlpath(urlpath)
+    blosc2.remove_urlpath(copy_path)
+
+
 @pytest.mark.parametrize(("contiguous", "nthreads"), [(False, 2), (True, 4)])
 def test_batchstore_multithreaded_inner_vl(contiguous, nthreads):
     batches = []