Moved is_structured_array() and to_serializable_dtype() to utils, ranamed a few functions and propagated changes to dependent modules.

utils/g5505_utils.py

@@ -109,7 +109,7 @@ def created_at():
     created_at = now_tz_aware.strftime('%Y-%m-%d_%H-%M-%S') + '_UTC-OFST_' + tz
     return created_at
 
-def dataframe_to_np_structured_array(df: pd.DataFrame):
+def convert_dataframe_to_np_structured_array(df: pd.DataFrame):
 
      # Define the dtype for the structured array, ensuring compatibility with h5py
     dtype = []
@@ -153,6 +153,47 @@ def convert_string_to_bytes(input_list: list):
 
     return input_array_bytes
 
+def convert_attrdict_to_np_structured_array(attr_value: dict):
+    """
+    Converts a dictionary of attributes into a numpy structured array for HDF5 
+    compound type compatibility.
+
+    Each dictionary key is mapped to a field in the structured array, with the 
+    data type (S) determined by the longest string representation of the values. 
+    If the dictionary is empty, the function returns 'missing'.
+
+    Parameters
+    ----------
+    attr_value : dict
+        Dictionary containing the attributes to be converted. Example:
+        attr_value = {
+            'name': 'Temperature',
+            'unit': 'Celsius',
+            'value': 23.5,
+            'timestamp': '2023-09-26 10:00'
+        }
+
+    Returns
+    -------
+    new_attr_value : ndarray or str
+        Numpy structured array with UTF-8 encoded fields. Returns 'missing' if 
+        the input dictionary is empty.
+    """
+    dtype = []
+    values_list = []
+    max_length = max(len(str(attr_value[key])) for key in attr_value.keys())
+    for key in attr_value.keys():
+        if key != 'rename_as':
+            dtype.append((key, f'S{max_length}'))
+            values_list.append(attr_value[key])  
+    if values_list:
+        new_attr_value = np.array([tuple(values_list)], dtype=dtype)
+    else:
+        new_attr_value = 'missing'
+
+    return new_attr_value
+
+
 def infer_units(column_name):
     # TODO: complete or remove
 
@@ -165,23 +206,6 @@ def infer_units(column_name):
         
     return match
 
-def parse_attribute(attr_value : dict):
-    "Parse a dictionary attribute into an equivalent numpy structured array, which compatible with compound HDF5 type"
-    dtype = []
-    values_list = []
-    max_length = max(len(str(attr_value[key])) for key in attr_value.keys())
-    for key in attr_value.keys():
-        if (not key=='rename_as'):
-            dtype.append((key,f'S{max_length}'))
-            values_list.append(attr_value[key])  
-
-    if values_list:
-        new_attr_value = np.array([tuple(values_list)],dtype=dtype)
-    else:
-        new_attr_value = 'missing'
-
-    return new_attr_value
-
 def progressBar(count_value, total, suffix=''):
     bar_length = 100
     filled_up_Length = int(round(bar_length* count_value / float(total)))
@@ -270,4 +294,59 @@ def copy_directory_with_contraints(input_dir_path, output_dir_path,
                         except Exception as e:
                             logging.error("Failed to copy %s: %s", src_file_path, e)
 
-    return path_to_files_dict 
+    return path_to_files_dict 
+
+def to_serializable_dtype(value):
+
+    """Transform value's dtype into  YAML/JSON compatible dtype
+
+    Parameters
+    ----------
+    value : _type_
+        _description_
+
+    Returns
+    -------
+    _type_
+        _description_
+    """
+    try:
+        if isinstance(value, np.generic):
+            if np.issubdtype(value.dtype, np.bytes_):
+                value = value.decode('utf-8')
+            elif np.issubdtype(value.dtype, np.unicode_):
+                value = str(value)
+            elif np.issubdtype(value.dtype, np.number):
+                value = float(value)
+            else:
+                print('Yaml-compatible data-type was not found. Value has been set to NaN.')
+                value = np.nan    
+        elif isinstance(value, np.ndarray):  
+            # Handling structured array types (with fields)
+            if value.dtype.names:
+                value = {field: to_serializable_dtype(value[field]) for field in value.dtype.names}
+            else:
+                # Handling regular array NumPy types                         
+                if np.issubdtype(value.dtype, np.bytes_):
+                    value = [item.decode('utf-8') for item in value] if len(value) > 1 else value[0].decode('utf-8')
+                elif np.issubdtype(value.dtype, np.unicode_):
+                    value = [str(item) for item in value] if len(value) > 1 else str(value[0])
+                elif np.issubdtype(value.dtype, np.integer):
+                    value = [int(item) for item in value] if len(value) > 1 else int(value[0])
+                elif np.issubdtype(value.dtype, np.floating):
+                    value = [float(item) for item in value] if len(value) > 1 else float(value[0])
+                else:
+                    print('Yaml-compatible data-type was not found. Value has been set to NaN.')
+                    value = np.nan
+
+    except Exception as e:
+        print(f'Error converting value: {e}. Value has been set to NaN.')
+        value = np.nan
+
+    return value
+
+def is_structured_array(attr_val):
+    if isinstance(attr_val,np.ndarray):
+        return True if attr_val.dtype.names is not None else False
+    else: 
+        return False