Clean up

Author: rly

.gitignore

@@ -131,4 +131,4 @@ dmypy.json
 .DS_Store
 /out/*
 /my_container.h5
-/my_container.zarr
+/my_container.zarr

src/linkml_arrays/dumpers/hdf5_dumper.py

@@ -15,7 +15,7 @@ def _iterate_element(
 ):
     """Recursively iterate through the elements of a LinkML model and save them.
 
-    Writes Pydantic BaseModel objects as groups, slots that implement "linkml:elements"
+    Write Pydantic BaseModel objects as groups, slots with the "array" element
     as datasets, and other slots as attributes.
     """
     # get the type of the element

src/linkml_arrays/dumpers/yaml_array_file_dumper.py

@@ -1,4 +1,4 @@
-"""Base class for dumpling a LinkML model to a YAML file with paths to NumPy files."""
+"""Base class for dumping a LinkML model to a YAML file with paths to files containing individual arrays."""
 
 import os
 from abc import ABCMeta, abstractmethod
@@ -24,10 +24,9 @@ def _iterate_element(
 ):
     """Recursively iterate through the elements of a LinkML model and save them.
 
-    Returns a dictionary with the same structure as the input element, but with the slots
-    that implement "linkml:elements" (arrays) are written to HDF5 files and the paths to these
-    files are returned in the dictionary. Each array is written to an HDF5 dataset at path
-    "/data" in a new HDF5 file.
+    Return a dictionary with the same structure as the input element, but where the slots
+    with the "array" element are written to an array file and the paths to these
+    files are returned in the dictionary. The paths are relative to the output directory.
 
     Raises:
         ValueError: If the class requires an identifier and it is not provided.

src/linkml_arrays/dumpers/yaml_dumper.py

@@ -1,4 +1,4 @@
-"""Class for dumpling a LinkML model to a YAML file."""
+"""Class for dumping a LinkML model to a YAML file."""
 
 from typing import Union
 
@@ -14,8 +14,8 @@ def _iterate_element(
 ):
     """Recursively iterate through the elements of a LinkML model and save them.
 
-    Returns a dictionary with the same structure as the input element, but with the slots
-    that implement "linkml:elements" (arrays) are written as lists or lists of lists.
+    Returns a dictionary with the same structure as the input element, but where the slots
+    with the "array" element are written as lists of lists in YAML.
 
     Raises:
         ValueError: If the class requires an identifier and it is not provided.
@@ -35,10 +35,11 @@ def _iterate_element(
     ret_dict = dict()
     for k, v in vars(element).items():
         found_slot = schemaview.induced_slot(k, element_type)
-        if "linkml:elements" in found_slot.implements:
+        if found_slot.array:
             if id_slot is None and parent_identifier is None:
                 raise ValueError("The class requires an identifier.")
-            ret_dict[k] = v.tolist()
+            assert isinstance(v, list)
+            ret_dict[k] = v
         else:
             if isinstance(v, BaseModel):
                 v2 = _iterate_element(v, schemaview, id_value)

src/linkml_arrays/dumpers/yaml_hdf5_dumper.py

@@ -10,15 +10,18 @@
 
 
 class YamlHdf5Dumper(YamlArrayFileDumper):
-    """Dumper class for LinkML models to YAML files with paths to NumPy files."""
+    """Dumper class for LinkML models to YAML files with paths to HDF5 files, one for each array.
+
+    Each array is written to an HDF5 dataset at path "/data" in a new HDF5 file.
+    """
 
     FILE_SUFFIX = ".h5"  # used in parent class
 
     @classmethod
     def write_array(
         cls, array: Union[List, np.ndarray], output_file_path_no_suffix: Union[str, Path]
     ):
-        """Write an array to a file."""
+        """Write an array to an HDF5 file."""
         # TODO do not assume that there is only one by this name
         # add suffix to the file name
         output_file_path = output_file_path_no_suffix.parent / (

src/linkml_arrays/dumpers/yaml_numpy_dumper.py

@@ -1,4 +1,4 @@
-"""Class for dumpling a LinkML model to a YAML file with paths to NumPy files."""
+"""Class for dumping a LinkML model to a YAML file with paths to NumPy files."""
 
 from pathlib import Path
 from typing import List, Union
@@ -9,15 +9,18 @@
 
 
 class YamlNumpyDumper(YamlArrayFileDumper):
-    """Dumper class for LinkML models to YAML files with paths to NumPy files."""
+    """Dumper class for LinkML models to YAML files with paths to NumPy .npy files, one for each array.
+
+    Each array is written to an HDF5 dataset at path "/data" in a new HDF5 file.
+    """
 
     FILE_SUFFIX = ".npy"  # used in parent class
 
     @classmethod
     def write_array(
         cls, array: Union[List, np.ndarray], output_file_path_no_suffix: Union[str, Path]
     ):
-        """Write an array to a file."""
+        """Write an array to a NumPy file."""
         # TODO do not assume that there is only one by this name
         # add suffix to the file name
         output_file_path = output_file_path_no_suffix.parent / (

src/linkml_arrays/dumpers/zarr_directory_store_dumper.py

@@ -15,16 +15,16 @@ def _iterate_element(
 ):
     """Recursively iterate through the elements of a LinkML model and save them.
 
-    Writes Pydantic BaseModel objects as groups, slots that implement "linkml:elements"
-    as datasets, and other slots as attributes.
+    Write Pydantic BaseModel objects as groups, slots with the "array" element
+    as arrays, and other slots as attributes.
     """
     # get the type of the element
     element_type = type(element).__name__
 
     for k, v in vars(element).items():
         found_slot = schemaview.induced_slot(k, element_type)
         if found_slot.array:
-            # save the numpy array to an hdf5 dataset
+            # save the numpy array to a zarr array
             group.create_dataset(found_slot.name, data=v)
         else:
             if isinstance(v, BaseModel):

src/linkml_arrays/loaders/yaml_loader.py

@@ -14,10 +14,7 @@
 def _iterate_element(
     input_dict: dict, element_type: ClassDefinition, schemaview: SchemaView
 ) -> dict:
-    """Recursively iterate through the elements of a LinkML model and load them into a dict.
-
-    Datasets are loaded into NumPy arrays.
-    """
+    """Recursively iterate through the elements of a LinkML model and load them into a dict."""
     ret_dict = dict()
     for k, v in input_dict.items():
         found_slot = schemaview.induced_slot(k, element_type.name)