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

Removed construct_attributes_dict(attrs_obj) and replaced by {key: utils.to_serializable_dtype(val) for key, val in obj.attrs.items()}

Browse Source
This commit is contained in:
florez_j
2024-10-01 10:42:20 +02:00
parent 4d48e84e50
commit 4420f81642
1 changed files with 23 additions and 59 deletions
Download Patch File Download Diff File Expand all files Collapse all files

82
src/hdf5_ops.py
View File

@ -31,20 +31,16 @@ class HDF5DataOpsManager():
"""
def __init__(self, file_path, mode = 'r+') -> None:
# Class attributes
if mode in ['r','r+']:
self.mode = mode
self.file_path = file_path
self.file_obj = None
#self._open_file()
self.list_of_datasets = []
self.dataset_metadata_df = None
# Define private methods
def _collect_dataset_names(self, name, obj, list_of_datasets):
if isinstance(obj, h5py.Dataset):
list_of_datasets.append(name)
# Define public methods
def open_file(self):
@ -56,16 +52,23 @@ class HDF5DataOpsManager():
self.file_obj.flush() # Ensure all data is written to disk
self.file_obj.close()
self.file_obj = None
def load_dataset_metadata(self):
def retrieve_dataframe_of_dataset_names(self):
def __get_datasets(name, obj, list_of_datasets):
if isinstance(obj,h5py.Dataset):
list_of_datasets.append(name)
#print(f'Adding dataset: {name}') #tail: {head} head: {tail}')
list_of_datasets = []
self.file_obj.visititems(lambda name, obj: self._collect_dataset_names(name, obj, list_of_datasets))
with h5py.File(self.file_path,'r') as file:
list_of_datasets = []
file.visititems(lambda name, obj: __get_datasets(name, obj, list_of_datasets))
dataset_df = pd.DataFrame({'dataset_name': list_of_datasets})
dataset_df['parent_instrument'] = dataset_df['dataset_name'].apply(lambda x: x.split('/')[-3])
dataset_df['parent_file'] = dataset_df['dataset_name'].apply(lambda x: x.split('/')[-2])
dataset_metadata_df = pd.DataFrame({'dataset_name': list_of_datasets})
dataset_metadata_df['parent_instrument'] = dataset_metadata_df['dataset_name'].apply(lambda x: x.split('/')[-3])
dataset_metadata_df['parent_file'] = dataset_metadata_df['dataset_name'].apply(lambda x: x.split('/')[-2])
return dataset_df
self.dataset_metadata_df = dataset_metadata_df
def read_dataset_as_dataframe(self,dataset_name):
"""
@ -371,25 +374,6 @@ def read_dataset_from_hdf5file(hdf5_file_path, dataset_path):
#metadata_df = pd.DataFrame.from_dict(data, orient='columns')
return df
def list_datasets_in_hdf5file(hdf5_file_path):
def get_datasets(name, obj, list_of_datasets):
if isinstance(obj,h5py.Dataset):
list_of_datasets.append(name)
#print(f'Adding dataset: {name}') #tail: {head} head: {tail}')
with h5py.File(hdf5_file_path,'r') as file:
list_of_datasets = []
file.visititems(lambda name, obj: get_datasets(name, obj, list_of_datasets))
dataset_df = pd.DataFrame({'dataset_name':list_of_datasets})
dataset_df['parent_instrument'] = dataset_df['dataset_name'].apply(lambda x: x.split('/')[-3])
dataset_df['parent_file'] = dataset_df['dataset_name'].apply(lambda x: x.split('/')[-2])
return dataset_df
def get_parent_child_relationships(file: h5py.File):
nodes = ['/']
@ -423,29 +407,6 @@ def get_parent_child_relationships(file: h5py.File):
return nodes, parent, values
def construct_attributes_dict(attrs_obj):
attr_dict = {}
for key, value in attrs_obj.items():
attr_dict[key] = {}
if not key in ['file_list','filtered_file_list']:
if utils.is_structured_array(value):
#for subattr in value.dtype.names:
#attr_dict[key][subattr] = make_dtype_yaml_compatible(value[subattr])
attr_dict[key] = utils.to_serializable_dtype(value)
else:
attr_dict[key] = utils.to_serializable_dtype(value) # {"rename_as" : key,
#"value" : utils.to_serializable_dtype(value)
#}
#if isinstance(value,str):
# value.replace('\\','\\\\')
return attr_dict
def __print_metadata__(name, obj, folder_depth, yaml_dict):
# TODO: should we enable deeper folders ?
@ -459,7 +420,8 @@ def __print_metadata__(name, obj, folder_depth, yaml_dict):
#attr_dict = {}
group_dict = {}
attr_dict = construct_attributes_dict(obj.attrs)
# Convert attribute dict to a YAML/JSON serializable dict
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)
@ -482,9 +444,11 @@ def __print_metadata__(name, obj, folder_depth, yaml_dict):
#print(name)
yaml_dict[obj.name] = group_dict
elif isinstance(obj, h5py.Dataset):
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": construct_attributes_dict(obj.attrs)}
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:
@ -522,8 +486,8 @@ 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:
# Construct attributes dictionary and top-level structure
attrs_dict = construct_attributes_dict(f.attrs)
# 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,