Dealing with derivatives

bids2openminds/converter.py

@@ -8,16 +8,11 @@
 
 
 def convert(input_path, output_path=None, multiple_files=False, include_empty_properties=False):
-    if not (os.path.isdir(input_path)):
-        raise NotADirectoryError(
-            f"The input directory is not valid, you have specified {input_path} which is not a directory."
-        )
-    # if not(BIDSValidator().is_bids(input_path)):
-    #  raise NotADirectoryError(f"The input directory is not valid, you have specified {input_path} which is not a BIDS directory.")
+
     collection = Collection()
-    bids_layout = BIDSLayout(input_path)
 
-    layout_df = bids_layout.to_df()
+    [bids_layout, layout_df, dataset_description] = main.read_bids_directory(
+        input_path)
 
     subjects_id = bids_layout.get_subjects()
 

bids2openminds/main.py

@@ -2,18 +2,70 @@
 import os
 import pathlib
 from warnings import warn
-
+from bids import BIDSLayout
 import pandas as pd
 from nameparser import HumanName
 
 import openminds.latest.core as omcore
 import openminds.latest.controlled_terms as controlled_terms
 from openminds import IRI
 
-from .utility import table_filter, pd_table_value, file_hash, file_storage_size
+from .utility import table_filter, pd_table_value, file_hash, file_storage_size, read_json
 from .mapping import bids2openminds_instance
 
 
+def read_bids_directory(input_path):
+    """
+    Reads the content of the dataset_description.json file and determines whether the dataset is raw, derivative, or contains derivatives. Then, it calls the appropriate BIDSLayout function.
+    Returns three Python objects.
+
+    Parameters:
+    - input_path: The path to the BIDS folder.
+
+    Returns:
+    - bids_layout: An instance of the Layout class from pybids for this dataset.
+    - layout_df: A pandas DataFrame containing all the files and corresponding information.
+    - dataset_description: A Python dictionary containing the content of the dataset_description JSON file.
+    """
+
+    if not os.path.isdir(input_path):
+        raise NotADirectoryError(
+            f"The input path is not valid. You specified '{input_path}', which is not a directory.")
+
+    # if not(BIDSValidator().is_bids(input_path)):
+    #  raise NotADirectoryError(f"The input directory is not valid, you have specified {input_path} which is not a BIDS directory.")
+    dataset_description_path = os.path.join(
+        input_path, "dataset_description.json")
+    dataset_description = read_json(dataset_description_path)
+
+    if not dataset_description:
+        bids_layout = BIDSLayout(input_path)
+        layout_df = bids_layout.to_df()
+        dataset_description_path = table_filter(layout_df, "description")
+        dataset_description = read_json(dataset_description_path.iat[0, 0])
+        if not dataset_description:
+            raise FileNotFoundError(
+                "There was no dataset_description.json file. Every dataset MUST include this file."
+            )
+        return bids_layout, layout_df, dataset_description
+
+    if ("DatasetType" in dataset_description) and (dataset_description["DatasetType"] == "derivative"):
+        bids_layout = BIDSLayout(input_path, is_derivative=True)
+        layout_df = bids_layout.to_df()
+        return bids_layout, layout_df, dataset_description
+
+    derivatives_path = os.path.join(
+        input_path, "derivatives")
+    if os.path.isdir(derivatives_path):
+        bids_layout = BIDSLayout(input_path, derivatives=True)
+        layout_df = bids_layout.to_df()
+        return bids_layout, layout_df, dataset_description
+
+    bids_layout = BIDSLayout(input_path)
+    layout_df = bids_layout.to_df()
+    return bids_layout, layout_df, dataset_description
+
+
 def create_openminds_person(full_name):
     # Regex for detecting any unwanted characters.
     name_regex = re.compile(

test/test_bids_examples.py

@@ -5,7 +5,8 @@
 import bids2openminds.converter
 
 example_dataset = [("ds003", 13), ("ds000247", 6),
-                   ("eeg_cbm", 20), ("asl001", 1), ("eeg_rest_fmri", 3)]
+                   ("eeg_cbm", 20), ("asl001", 1),
+                   ("eeg_rest_fmri", 3), ("ds000001-fmriprep", 4)]
 
 
 @pytest.mark.parametrize("dataset_label,dataset_number_subject", example_dataset)