Merge pull request #13 from VisLab/main

Problem with proposed validation strategy proposed in #10

pyproject.toml

@@ -35,9 +35,9 @@ keywords = [
     'ndx-extension',
 ]
 dependencies = [
-    "pynwb>=2.5.0",
-    "hdmf>=3.10.0",
-    "hedtools>=0.4.0"
+    "pynwb>=2.8.0",
+    "hdmf>=3.14.1",
+    "hedtools>=0.5.0"
 ]
 
 # TODO: add URLs before release
@@ -77,7 +77,7 @@ include = [
 exclude = []
 
 [tool.pytest.ini_options]
-addopts = "--cov --cov-report html"
+# addopts = "--cov --cov-report html"
 
 [tool.codespell]
 skip = "htmlcov,.git,.mypy_cache,.pytest_cache,.coverage,*.pdf,*.svg,venvs,.tox,hdmf-common-schema,./docs/_build/*,*.ipynb"

requirements-dev.txt

@@ -1,16 +1,16 @@
 # pinned dependencies to reproduce an entire development environment to
 # run tests, check code style, and generate documentation
-black==23.9.1
-codespell==2.2.6
-coverage==7.3.2
-hdmf==3.10.0
-hdmf-docutils==0.4.6
-hedtools==0.4.0
+black==24.4.2
+codespell==2.3.0
+coverage==7.5.3
+hdmf==3.14.1
+hdmf-docutils==0.4.7
+hedtools==0.5.0
 pre-commit==3.4.0
-pynwb==2.5.0
-pytest==7.4.2
+pynwb==2.8.0
+pytest==8.2.2
 pytest-cov==4.1.0
 python-dateutil==2.8.2
 pytest-subtests==0.6.0
-ruff==0.0.292
-tox==4.11.3
+ruff==0.4.10
+tox==4.15.1

requirements-min.txt

@@ -1,6 +1,6 @@
 # minimum versions of package dependencies for installation
 # these should match the minimum versions specified in pyproject.toml
 # NOTE: it may be possible to relax these minimum requirements
-pynwb==2.5.0
-hdmf==3.10.0
-hedtools==0.4.0
+pynwb==2.8.0
+hdmf==3.14.1
+hedtools==0.5.0

spec/ndx-hed.extensions.yaml

@@ -1,31 +1,12 @@
 datasets:
 - neurodata_type_def: HedTags
   neurodata_type_inc: VectorData
-  name: HED
   dtype: text
   doc: An extension of VectorData for Hierarchical Event Descriptor (HED) tags. If
     HED tags are used, the HED schema version must be specified in the NWB file using
     the HedMetadata type.
-groups:
-- neurodata_type_def: HedVersionAttr
-  neurodata_type_inc: LabMetaData
-  name: hed_version
-  doc: An extension of LabMetaData to store the Hierarchical Event Descriptor (HED)
-    schema version. TODO When merged with core, this will no longer inherit from LabMetaData
-    but from NWBContainer and be placed optionally in /general.
   attributes:
-  - name: version
+  - name: hed_version
     dtype: text
-    doc: The version of the HED schema used to validate the HED tags, e.g., '8.2.0'.
-      Required if HED tags are used in the NWB file.
-- neurodata_type_def: HedVersion
-  neurodata_type_inc: LabMetaData
-  name: hed_version
-  doc: An extension of LabMetaData to store the Hierarchical Event Descriptor (HED)
-    schema version. TODO When merged with core, this will no longer inherit from LabMetaData
-    but from NWBContainer and be placed optionally in /general.
-  datasets:
-  - name: version
-    dtype: text
-    doc: HED schema version to use for this dataset
-    quantity: '?'
+    doc: "The version of the HED schema used to validate the HED tags, e.g., '8.2.0'.
+      (Required). "

spec/ndx-hed.namespace.yaml

@@ -14,7 +14,6 @@ namespaces:
   schema:
   - namespace: core
     neurodata_types:
-    - LabMetaData
     - VectorData
   - source: ndx-hed.extensions.yaml
   version: 0.1.0

src/pynwb/README.md

@@ -1 +1,40 @@
-## ndx_hed NWB extension
+## ndx_hed Extension for NWB
+
+This is an NWB extension for adding Hierarchical Event Descriptor (HED) tags to NWB data.
+HED is a system for
+
+The latest version is 0.1.0. This is the first release of ndx-hed.
+
+
+This extension was developed by Kay Robbins, Ryan Ly, Oliver Rübel, and the HED Working Group.
+The documentation for this extension can be found at 
+
+## Installation
+Python:
+```bash
+pip install -U ndx-hed
+```
+
+## Developer installation
+In a Python 3.8-3.12 environment:
+```bash
+pip install -r requirements-dev.txt
+pip install -e .
+```
+
+Run tests:
+```bash
+pytest
+```
+
+Install pre-commit hooks:
+```bash
+pre-commit install
+```
+
+Style and other checks:
+```bash
+black .
+ruff .
+codespell .
+```

src/pynwb/ndx_hed/__init__.py

@@ -53,6 +53,4 @@
 # # Remove these functions from the package
 del load_namespaces, get_class
 
-from .hed_version_attr import HedVersionAttr
 from .hed_tags import HedTags
-from .hed_version import HedVersion

src/pynwb/ndx_hed/hed_tags.py

@@ -6,7 +6,6 @@
 from hed.models import HedString
 from pynwb import register_class
 from pynwb.file import LabMetaData, NWBFile
-from ndx_hed import HedVersionAttr
 
 
 @register_class('HedTags', 'ndx-hed')
@@ -18,81 +17,47 @@ class HedTags(VectorData):
 
     """
 
-    __nwbfields__ = ('_hed_schema')
+    __nwbfields__ = ('_hed_schema', 'hed_version')
 
-    @docval({'name': 'name', 'type': 'str', 'doc': 'Must be HED', 'default': 'HED'},
-            {'name': 'description', 'type': 'str', 'doc': 'Description of the HED annotations',
-             'default': 'Hierarchical Event Descriptors (HED) annotations'},
-             *get_docval(VectorData.__init__, 'data'))
+    @docval({'name': 'hed_version', 'type': 'str', 'doc': 'The version of HED used by this data.'},
+            {'name': 'name', 'type': str, 'doc': 'The name of this VectorData', 'default': 'HED'},
+            {'name': 'description', 'type': str, 'doc': 'a description for this column',
+             'default': "Column that stores HED tags as text annotating their respective row."},
+            *get_docval(VectorData.__init__, 'data'))
     def __init__(self, **kwargs):
-        kwargs['name'] = 'HED'
+        hed_version = popargs('hed_version', kwargs)
         super().__init__(**kwargs)
+        self.hed_version = hed_version
         self._init_internal()
 
     def _init_internal(self):
         """
-        This finds the HED schema object of use in this NWBFile.
-
-        TODO: How should errors be handled if this file doesn't have a HedVersion object in the LabMetaData?
-
+        This loads (a private pointer to) the HED schema and validates the HED data.
         """
-        self._hed_schema = None
+        self._hed_schema = load_schema_version(self.hed_version)
+        issues = []
+        for index in range(len(self.data)):
+            hed_obj = HedString(self.get(index), self._hed_schema)
+            these_issues = hed_obj.validate()
+            if these_issues:
+                issues.append(f"line {str(index)}: {get_printable_issue_string(these_issues)}")
+        if issues:
+            raise HedFileError("InvalidHEDData", "\n".join(issues), "")
 
     @docval({'name': 'val', 'type': str,
              'doc': 'the value to add to this column. Should be a valid HED string -- just forces string.'})
     def add_row(self, **kwargs):
         """Append a data value to this column."""
         val = getargs('val', kwargs)
+        hed_obj = HedString(val, self._hed_schema)
+        these_issues = hed_obj.validate()
+        if these_issues:
+            raise HedFileError("InvalidHEDValue",
+                               f"{str(val)} issues: {get_printable_issue_string(these_issues)}", "")
         super().append(val)
 
-    # @docval({'name': 'schema', 'type': (HedSchema, None), 'doc': 'HedSchema to use to validate.', 'default': None},
-    #         {'name': 'return', 'type': 'list', 'doc': 'list of issues or none'})
-    def validate(self, schema):
-        """Validate this VectorData. This is assuming a list --- where is the general iterator."""
-        if not schema:
-            raise HedFileError('HedSchemaMissing', "Must provide a valid HedSchema", "")
-        issues = []
-        for index in range(len(self.data)):
-            hed_obj = HedString(self.get(index), schema)
-            these_issues = hed_obj.validate()
-            if these_issues:
-                issues.append(f"line {str(index)}: {get_printable_issue_string(these_issues)}")
-        return "\n".join(issues)
+    def get_hed_version(self):
+        return self.hed_version
 
     def get_hed_schema(self):
-        if not self._hed_schema:
-            root = self._get_root()
-            if isinstance(root, NWBFile):
-                self._hed_schema = root.get_lab_meta_data("hed_version").get_schema()
         return self._hed_schema
-
-    def _get_root(self):
-        root = self
-        while hasattr(root, 'parent') and root.parent:
-            root = root.parent
-        if root == self:
-            return None
-        return root
-
-
-        #     root = parent
-        #     parent = root.parent
-        # if parent:
-        #     hed_version = parent.get_lab_meta_data("HedVersion")
-        # else:
-        #     hed_version = None
-        # if hed_version:
-        #     self.hed_schema = hed_version.get_schema()
-
-    # root = self
-    # parent = root.parent
-    # while parent is not None:
-    #     root = parent
-    #     parent = root.parent
-    # if parent:
-    #     hed_version = parent.get_lab_meta_data("HedVersion")
-    # else:
-    #     hed_version = None
-    # if hed_version:
-    #     self.hed_schema = hed_version.get_schema()
-