Implemented hdf5_vis.py to gather functions that display or represent properties of an hdf5 file in a human readable file format like yalm or html files that enble interative visualizations on the browser.

src/hdf5_vis.py

@@ -0,0 +1,152 @@
+import h5py
+import yaml
+import os
+import numpy as np
+import numpy as pd
+
+import config_file
+import hdf5_lib
+import g5505_utils as utils
+
+import matplotlib.pyplot as plt
+import plotly.express as px
+import plotly.graph_objects as go
+from plotly.subplots import make_subplots
+
+import subprocess
+
+#output = subprocess.run("git status",capture_output=True)
+#output.stdout()
+
+def display_group_hierarchy_on_a_treemap(filename: str):
+
+    """
+    filename (str): hdf5 file's filename"""
+
+    with h5py.File(filename,'r') as file:
+        nodes, parents, values = hdf5_lib.get_parent_child_relationships(file)
+
+        metadata_list = []
+        metadata_dict={}
+        for key in file.attrs.keys():            
+            if 'metadata' in key:
+                metadata_dict[key[key.find('_')+1::]]= file.attrs[key]
+                metadata_list.append(key[key.find('_')+1::]+':'+file.attrs[key])
+        metadata = '<br>'.join(['<br>'] + metadata_list)
+
+    customdata_series = pd.Series(nodes)
+    customdata_series[0] = metadata
+
+    fig = make_subplots(1, 1, specs=[[{"type": "domain"}]],)
+    fig.add_trace(go.Treemap(
+        labels=nodes, #formating_df['formated_names'][nodes],
+        parents=parents,#formating_df['formated_names'][parents],
+        values=values,
+        branchvalues='remainder',
+        customdata= customdata_series,
+        #marker=dict(
+        #    colors=df_all_trees['color'],
+        #    colorscale='RdBu',
+        #    cmid=average_score),
+        #hovertemplate='<b>%{label} </b> <br> Number of files: %{value}<br> Success rate: %{color:.2f}',
+        hovertemplate='<b>%{label} </b> <br> Count: %{value} <br> Path: %{customdata}',
+        name='',
+        root_color="lightgrey"
+        ))
+    fig.update_layout(width = 800, height= 600, margin = dict(t=50, l=25, r=25, b=25))
+    fig.show()
+    file_name, file_ext = os.path.splitext(filename)
+    fig.write_html(file_name + ".html")
+
+#
+    
+def make_dtype_yaml_compatible(value):
+    try:
+        if isinstance(value, np.generic):
+            if np.issubdtype(value.dtype, np.string_) or np.issubdtype(value.dtype, np.unicode_):
+                #value = value.astype(str)
+                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):                        
+            if np.issubdtype(value.dtype, np.string_):
+                value = value.astype(str).tolist()
+            elif np.issubdtype(value.dtype, np.integer):
+                value = value.astype(int).tolist()
+            elif np.issubdtype(value.dtype, np.floating):
+                value = value.astype(float).tolist()
+            elif np.issubdtype(value.dtype, np.generic): 
+                value = value.astype(str).tolist()   
+        #elif isinstance(value,(int,float,str)):
+    
+            #print('Yaml-compatible data-type was not found. Value has been set to Nan.')
+            #value = np.nan    
+            #print('leave value as is')
+    except:
+        print('Yaml-compatible data-type was not found. Value has been set to Nan.')
+        value = np.nan
+
+    return value
+
+def print_metadata(name, obj, yaml_dict): 
+
+    # TODO: should we enable deeper folders ?
+    if len(obj.name.split('/')) <= 4:
+        name_to_list = obj.name.split('/')
+        name_head = name_to_list[-1]
+
+        if isinstance(obj,h5py.Group):
+            #print('name:', obj.name)
+            #print('attributes:', dict(obj.attrs))
+            attr_dict = {}
+            group_dict = {}
+            for key, value in obj.attrs.items():
+                #print (key, value.dtype)
+                if key == 'Layout':
+                    print(value)
+
+                if not key in ['file_list','filtered_file_list']:
+
+                    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)
+
+            yaml_dict[obj.name] = group_dict
+        elif isinstance(obj, h5py.Dataset):            
+            parent_name = '/'.join(name_to_list[:-1])
+            yaml_dict[parent_name]["datasets"][name_head] = {'rename_as': name_head ,'attributes':dict(obj.attrs)}
+            #print(yaml.dump(group_dict,sort_keys=False))
+
+        #elif len(obj.name.split('/')) == 3:
+        #    print(yaml.dump())
+
+
+
+
+def take_yml_snapshot_of_hdf5_file(input_filename_path):
+
+    yaml_dict = {}
+
+    output_filename_tail, ext = os.path.splitext(input_filename_path)
+
+    with h5py.File(input_filename_path,'r') as f:
+        f.visititems(lambda name, obj: print_metadata(name,obj,yaml_dict))
+
+    with open(output_filename_tail+".yaml","w") as yaml_file:
+        yaml.dump(yaml_dict,yaml_file,sort_keys=False)
+
+    return output_filename_tail+".yaml"
+
+