Added function to validate review yaml file, and updated update_hdf5_with_review function

pipelines/metadata_revision.py

@@ -189,7 +189,65 @@ def load_yaml(yaml_review_file):
             print(exc)
             return None
 
-#def update_hdf5_attributes(input_hdf5_file, yaml_dict):
+def validate_yaml_dict(input_hdf5_file, yaml_dict):
+    errors = []
+    notes = []
+
+    with h5py.File(input_hdf5_file, 'r') as hdf5_file:
+        # 1. Check for valid object names
+        for key in yaml_dict:
+            if key not in hdf5_file:
+                error_msg = f"Error: {key} is not a valid object's name in the HDF5 file."
+                print(error_msg)
+                errors.append(error_msg)
+
+        # 2. Confirm metadata dict for each object is a dictionary
+        for key, meta_dict in yaml_dict.items():
+            if not isinstance(meta_dict, dict):
+                error_msg = f"Error: Metadata for {key} should be a dictionary."
+                print(error_msg)
+                errors.append(error_msg)
+            else:
+                if 'attributes' not in meta_dict:
+                    warning_msg = f"Warning: No 'attributes' in metadata dict for {key}."
+                    print(warning_msg)
+                    notes.append(warning_msg)
+
+        # 3. Verify update, append, and delete operations are well specified            
+        for key, meta_dict in yaml_dict.items():
+            attributes = meta_dict.get("attributes", {})
+
+            for attr_name, attr_value in attributes.items():
+                # Ensure the object exists before accessing attributes
+                if key in hdf5_file:
+                    hdf5_obj_attrs = hdf5_file[key].attrs  # Access object-specific attributes
+
+                    if attr_name in hdf5_obj_attrs:
+                        # Attribute exists: it can be updated or deleted
+                        if isinstance(attr_value, dict) and "delete" in attr_value:
+                            note_msg = f"Note: '{attr_name}' in {key} may be deleted if 'delete' is set as true."
+                            print(note_msg)
+                            notes.append(note_msg)
+                        else:
+                            note_msg = f"Note: '{attr_name}' in {key} will be updated."
+                            print(note_msg)
+                            notes.append(note_msg)
+                    else:
+                        # Attribute does not exist: it can be appended or flagged as an invalid delete
+                        if isinstance(attr_value, dict) and "delete" in attr_value:
+                            error_msg = f"Error: Cannot delete non-existent attribute '{attr_name}' in {key}."
+                            print(error_msg)
+                            errors.append(error_msg)
+                        else:
+                            note_msg = f"Note: '{attr_name}' in {key} will be appended."
+                            print(note_msg)
+                            notes.append(note_msg)
+                else:
+                    error_msg = f"Error: '{key}' is not a valid object in the HDF5 file."
+                    print(error_msg)
+                    errors.append(error_msg)
+
+    return len(errors) == 0, errors, notes
 
 
 def update_hdf5_file_with_review(input_hdf5_file, yaml_review_file):
@@ -215,6 +273,10 @@ def update_hdf5_file_with_review(input_hdf5_file, yaml_review_file):
     """
     yaml_dict = load_yaml(yaml_review_file)
 
+    success, errors, notes = validate_yaml_dict(input_hdf5_file,yaml_dict)
+    if not success:
+        raise ValueError(f"Review yaml file {yaml_review_file} is invalid. Validation errors: {errors}")
+
     # Initialize HDF5 operations manager
     DataOpsAPI = hdf5_ops.HDF5DataOpsManager(input_hdf5_file)
     DataOpsAPI.load_file_obj()
@@ -413,6 +475,9 @@ def main():
      #   third_update_hdf5_file_with_review(output_filename_path, os.path.join(os.path.join(os.path.abspath(os.curdir),"review"),filename))
     #fourth_complete_metadata_review()    
 
-#if __name__ == '__main__':
+if __name__ == "__main__":
+    if len(sys.argv) != 3:
+        print("Usage: python pipeline.py run  <path/to/descriptor.json>")
+        sys.exit(1)
 
-#    main()
+    #run(sys.argv[2])