You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: CHANGELOG.md
+5-5Lines changed: 5 additions & 5 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -1,20 +1,20 @@
1
1
# HDMF Changelog
2
2
3
-
## HDMF 6.0.3 (Upcoming)
4
-
5
-
### Breaking changes
6
-
- Removed the `file` argument from `HERD.add_ref` and `HERD.add_ref_termset`. The file is now always resolved automatically from the container's parent hierarchy, so a reference can only be added to a container that has already been added to a file. This enforces that an external reference cannot be attached to a container that is not yet in a file. Callers passing `file=...` to these methods should remove that argument and ensure the container has been added to its file first. @bendichter
3
+
## HDMF 6.1.0 (June 25, 2026)
7
4
8
5
### Enhancements
9
-
- Accept pandas `Series` and `ExtensionArray` (including `StringArray` and `ArrowStringArray`) as `data` in `Data` and its subclasses (e.g., `VectorData`), normalizing to numpy at construction. This restores compatibility with pandas 3.0, where DataFrame string columns are PyArrow-backed by default and previously failed HDMF's type validation. Inputs containing missing values (`pd.NA`/`NaN`) raise an informative `TypeError` rather than silently failing at HDF5 write time. The `pandas<3` cap has been lifted from the dependency pin. @rly[#1384](https://github.com/hdmf-dev/hdmf/issues/1384)
6
+
- Accept pandas `Series` and `ExtensionArray` (including `StringArray` and `ArrowStringArray`) as `data` in `Data` and its subclasses (e.g., `VectorData`), normalizing to numpy at construction. This restores compatibility with pandas 3.0, where DataFrame string columns are PyArrow-backed by default and previously failed HDMF's type validation. Inputs containing missing values (`pd.NA`/`NaN`) raise an informative `TypeError` rather than silently failing at HDF5 write time. The `pandas<3` cap has been lifted from the dependency pin. @rly[#1469](https://github.com/hdmf-dev/hdmf/pull/1469)
10
7
- Added a `HERD`-specific `__repr__` and `_repr_html_` that surface the references as a flattened table, so a `HERD` (especially one read back from a file) no longer appears empty in its default display. @rly[#1510](https://github.com/hdmf-dev/hdmf/pull/1510)
11
8
-`HERD.add_ref` now defaults `key` to the value of a scalar string `attribute` when `key` is not provided, removing the redundant argument in the common case. @rly[#1511](https://github.com/hdmf-dev/hdmf/pull/1511)
12
9
-`HERD.add_ref` no longer warns when an `entity_uri` is provided for an already-existing `entity_id` and the URI matches the stored one. The entity tables are normalized, so re-passing the same `entity_uri` (common when annotating many objects or files with the same entity) is harmless; a warning is now emitted only when a *different*`entity_uri` is provided, in which case the existing URI is kept. @bendichter[#1513](https://github.com/hdmf-dev/hdmf/pull/1513)
13
10
14
11
### Changed
12
+
- Removed the `file` argument from `HERD.add_ref` and `HERD.add_ref_termset`. The file is now always resolved automatically from the container's parent hierarchy, so a reference can only be added to a container that has already been added to a file. This enforces that an external reference cannot be attached to a container that is not yet in a file. This is technically a breaking change, but the `file` argument was not yet used publicly. @bendichter[#1512](https://github.com/hdmf-dev/hdmf/pull/1512)
15
13
- Refactored `HERD` internals. The object lookup now identifies an object by its file together with its object_id, relative_path, and field, so a shared object_id across files (e.g. a copied, modified file) resolves to the correct object instead of colliding. @rly[#1515](https://github.com/hdmf-dev/hdmf/pull/1515)
16
14
17
15
### Internal improvements
16
+
- Minor `HERD` internal cleanups: removed a dead assignment and hoisted the files dataframe construction. @rly[#1507](https://github.com/hdmf-dev/hdmf/pull/1507)
17
+
- Switched the release pipeline to PyPI OIDC trusted publishing and the `gh` CLI, and now build the GitHub release notes from the changelog. @rly[#1517](https://github.com/hdmf-dev/hdmf/pull/1517)
18
18
- Hardened the GitHub Actions CI: added least-privilege `permissions` blocks, pinned actions to commit SHAs (or immutable release tags), deduplicated the test setup into a composite action, passed untrusted inputs through environment variables, and added a zizmor security audit of the workflows. @rly[#1518](https://github.com/hdmf-dev/hdmf/pull/1518)
0 commit comments