5505-public/dima
2
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Commented out metadata info about group members for a given group. This is to simplify yaml or json representation of the metadata.

Browse Source
This commit is contained in:
florez_j
2024-11-24 15:57:54 +01:00
parent d7f6e52068
commit 1174ffc8b8
1 changed files with 43 additions and 41 deletions
Download Patch File Download Diff File Expand all files Collapse all files

84
src/hdf5_ops.py
View File

@ -459,50 +459,51 @@ def get_parent_child_relationships(file: h5py.File):
def __print_metadata__(name, obj, folder_depth, yaml_dict):
# TODO: should we enable deeper folders ?
if len(obj.name.split('/')) <= folder_depth:
"""
Extracts metadata from HDF5 groups and datasets and organizes them into a dictionary with compact representation.
Parameters:
-----------
name (str): Name of the HDF5 object being inspected.
obj (h5py.Group or h5py.Dataset): The HDF5 object (Group or Dataset).
folder_depth (int): Maximum depth of folders to explore.
yaml_dict (dict): Dictionary to populate with metadata.
"""
# Process only objects within the specified folder depth
if len(obj.name.split('/')) <= folder_depth: # and ".h5" not in obj.name:
name_to_list = obj.name.split('/')
name_head = name_to_list[-1]
name_head = name_to_list[-1] if not name_to_list[-1]=='' else obj.name
if isinstance(obj,h5py.Group):
#print('name:', obj.name)
#print('attributes:', dict(obj.attrs))
#attr_dict = {}
group_dict = {}
# Convert attribute dict to a YAML/JSON serializable dict
if isinstance(obj, h5py.Group): # Handle groups
# Convert attributes to a YAML/JSON serializable format
attr_dict = {key: utils.to_serializable_dtype(val) for key, val in obj.attrs.items()}
#for key, value in obj.attrs.items():
#print (key, value.dtype)
# if key == 'Layout':
# print(value)
# Initialize the group dictionary
group_dict = {"name": name_head, "attributes": attr_dict}
# if not key in ['file_list','filtered_file_list']:
# Handle group members compactly
#subgroups = [member_name for member_name in obj if isinstance(obj[member_name], h5py.Group)]
#datasets = [member_name for member_name in obj if isinstance(obj[member_name], h5py.Dataset)]
# value = make_dtype_yaml_compatible(value)
# attr_dict[key] = {'rename_as' : key,
# 'value' : value
# }
#group_dict[obj.name] = {'name': obj.name, 'attributes': attr_dict}
group_dict = {"name": name_head, "attributes": attr_dict, "datasets":{}}
#group_dict[obj.name]["name"] = obj.name
#group_dict[obj.name]["attributes"] = attr_dict
#group_dict[obj.name]["datasets"] = {}
#print(name)
# Summarize groups and datasets
#group_dict["content_summary"] = {
# "group_count": len(subgroups),
# "group_preview": subgroups[:3] + (["..."] if len(subgroups) > 3 else []),
# "dataset_count": len(datasets),
# "dataset_preview": datasets[:3] + (["..."] if len(datasets) > 3 else [])
#}
yaml_dict[obj.name] = group_dict
elif isinstance(obj, h5py.Dataset):
# Convert attribute dict to a YAML/JSON serializable dict
attr_dict = {key: utils.to_serializable_dtype(val) for key, val in obj.attrs.items()}
parent_name = '/'.join(name_to_list[:-1])
yaml_dict[parent_name]["datasets"][name_head] = {"rename_as": name_head ,"attributes": attr_dict}
#print(yaml.dump(group_dict,sort_keys=False))
#elif len(obj.name.split('/')) == 3:
# print(yaml.dump())
elif isinstance(obj, h5py.Dataset): # Handle datasets
# Convert attributes to a YAML/JSON serializable format
attr_dict = {key: utils.to_serializable_dtype(val) for key, val in obj.attrs.items()}
dataset_dict = {"name": name_head, "attributes": attr_dict}
yaml_dict[obj.name] = dataset_dict
def serialize_metadata(input_filename_path, folder_depth: int = 4, output_format: str = 'yaml') -> str:
"""
@ -537,12 +538,13 @@ def serialize_metadata(input_filename_path, folder_depth: int = 4, output_format
# Open the HDF5 file and extract metadata
with h5py.File(input_filename_path, 'r') as f:
# Convert attribute dict to a YAML/JSON serializable dict
attrs_dict = {key: utils.to_serializable_dtype(val) for key, val in f.attrs.items()}
yaml_dict[f.name] = {
"name": f.name,
"attributes": attrs_dict,
"datasets": {}
}
#attrs_dict = {key: utils.to_serializable_dtype(val) for key, val in f.attrs.items()}
#yaml_dict[f.name] = {
# "name": f.name,
# "attributes": attrs_dict,
# "datasets": {}
#}
__print_metadata__(f.name, f, folder_depth, yaml_dict)
# Traverse HDF5 file hierarchy and add datasets
f.visititems(lambda name, obj: __print_metadata__(name, obj, folder_depth, yaml_dict))