improvements of Genfi

clinica/iotools/converters/genfi_to_bids/genfi_to_bids.py

@@ -9,6 +9,7 @@ def convert_images(
     bids_dir: PathLike,
     path_to_clinical: Optional[PathLike],
     gif: bool,
+    path_to_clinical_tsv: Optional[PathLike],
 ) -> None:
     """Convert the entire dataset to BIDS.
 
@@ -30,6 +31,10 @@ def convert_images(
 
     gif: bool
         If True, indicates the user wants to have the values of the gif parcellation
+
+    path_to_clinical_tsv: PathLike, optional
+        Path to a tsv containing the data the additionnal data the user wants to have in the BIDS
+        If None, no additionnal data will be added.
     """
     import os
 
@@ -43,29 +48,33 @@ def convert_images(
         merge_imaging_data,
         read_imaging_data,
         write_bids,
+        check_clinical_path,
     )
 
+    #check that if a clinical tsv is given, a path to the clinical data is given as well
+    if path_to_clinical_tsv:
+        check_clinical_path(path_to_clinical)
     # read the clinical data files
     if path_to_clinical:
-        df_demographics, df_imaging, df_clinical = find_clinical_data(path_to_clinical)
+        df_demographics, df_imaging, df_clinical, df_biosamples, df_neuropsych = find_clinical_data(path_to_clinical)
+
     # makes a df of the imaging data
     imaging_data = read_imaging_data(path_to_dataset)
-
+    
     # complete the data extracted
     imaging_data = merge_imaging_data(imaging_data)
     # complete clinical data
     if path_to_clinical:
         df_clinical_complete = complete_clinical_data(
-            df_demographics, df_imaging, df_clinical
+            df_demographics, df_imaging, df_clinical, df_biosamples, df_neuropsych
         )
-
     # intersect the data
     if path_to_clinical:
         df_complete = intersect_data(imaging_data, df_clinical_complete)
     else:
         df_complete = imaging_data
     # build the tsv
-    results = dataset_to_bids(df_complete, gif)
+    results = dataset_to_bids(df_complete, gif, path_to_clinical_tsv)
     write_bids(
         to=bids_dir,
         participants=results["participants"],

clinica/iotools/converters/genfi_to_bids/genfi_to_bids_cli.py

@@ -14,17 +14,25 @@
 )
 
 gif = click.option("-gif", is_flag=True, help="Add values from gif to session.tsv")
-
+clinical_data_tsv = click.option(
+    "-cdt",
+    "--clinical-data-tsv",
+    "clinical_data_tsv",
+    type=click.Path(exists=True, file_okay=True, resolve_path=True),
+    help="Path to a tsv containing additionnal clinical data you want to have in the BIDS",
+)
 
 @click.command(name="genfi-to-bids")
 @cli_param.dataset_directory
 @cli_param.bids_directory
 @clinical_data_directory
 @gif
+@clinical_data_tsv
 def cli(
     dataset_directory: PathLike,
     bids_directory: PathLike,
     clinical_data_directory: Optional[PathLike] = None,
+    clinical_data_tsv: Optional[PathLike] = None,
     gif: bool = False,
 ) -> None:
     """GENFI to BIDS converter.
@@ -39,7 +47,7 @@ def cli(
 
     check_dcm2niix()
 
-    convert_images(dataset_directory, bids_directory, clinical_data_directory, gif)
+    convert_images(dataset_directory, bids_directory, clinical_data_directory, gif, clinical_data_tsv)
     _write_bidsignore(str(bids_directory))
 
     cprint("Conversion to BIDS succeeded.")

clinica/iotools/converters/genfi_to_bids/genfi_to_bids_utils.py

@@ -7,7 +7,12 @@
 import pydicom as pdcm
 from pandas import DataFrame
 
-
+def check_clinical_path(path_to_clinical_data: PathLike) -> None:
+    if path_to_clinical_data:
+        return
+    else:
+        raise ValueError("Missing a clinical_data_path.")
+
 def find_dicoms(path_to_source_data: PathLike) -> Iterable[Tuple[PathLike, PathLike]]:
     """Find the dicoms in the given directory.
 
@@ -150,6 +155,8 @@ def find_clinical_data(
             "FINAL*DEMOGRAPHICS*.xlsx",
             "FINAL*IMAGING*.xlsx",
             "FINAL*CLINICAL*.xlsx",
+            "FINAL*BIOSAMPLES*.xlsx",
+            "FINAL*NEUROPSYCH*.xlsx",
         )
     )
 
@@ -168,7 +175,7 @@ def _read_file(data_file: PathLike) -> pd.DataFrame:
 
 
 def complete_clinical_data(
-    df_demographics: DataFrame, df_imaging: DataFrame, df_clinical: DataFrame
+    df_demographics: DataFrame, df_imaging: DataFrame, df_clinical: DataFrame, df_biosamples: DataFrame, df_neuropsych: DataFrame
 ) -> DataFrame:
     """Merges the different clincal dataframes into one.
 
@@ -192,24 +199,13 @@ def complete_clinical_data(
     df_clinical_complete = df_imaging.merge(
         df_demographics, how="inner", on=merge_key
     ).drop(columns="diagnosis")
+    df_clinical_complete = df_clinical_complete.merge(df_biosamples, how="inner", on=merge_key)
+    df_clinical_complete = df_clinical_complete.merge(df_neuropsych, how="inner", on=merge_key)
     df_clinical = df_clinical.dropna(subset=merge_key)
-    return df_clinical_complete.merge(
-        df_clinical[
-            [
-                "blinded_code",
-                "blinded_site",
-                "visit",
-                "diagnosis",
-                "ftld-cdr-global",
-                "cdr-sob",
-            ]
-        ],
-        how="inner",
-        on=merge_key,
-    )
+    return df_clinical_complete.merge(df_clinical, how="inner", on=merge_key)
 
 
-def dataset_to_bids(complete_data_df: DataFrame, gif: bool) -> Dict[str, DataFrame]:
+def dataset_to_bids(complete_data_df: DataFrame, gif: bool, path_to_clinical_tsv: PathLike) -> Dict[str, DataFrame]:
     """Selects the data needed to write the participants, sessions, and scans tsvs.
 
     Parameters
@@ -241,11 +237,23 @@ def dataset_to_bids(complete_data_df: DataFrame, gif: bool) -> Dict[str, DataFra
         "genfi_ref.csv",
     )
     df_ref = pd.read_csv(path_to_ref_csv, sep=";")
-
+    #add additionnal data through csv
+    additionnal_data_df = pd.read_csv(path_to_clinical_tsv, sep="\t")
+
+    #hard written path soon to be changed
+    map_to_level_df = pd.read_csv("/Users/matthieu.joulot/Desktop/clinical_data_dest.tsv", sep="\t")
+    pre_addi_df = map_to_level_df.merge(additionnal_data_df, how="inner", on="data")
+    session_addi_list = pre_addi_df["data"][pre_addi_df["dest"]=="sessions"].values.tolist()
+    participants_addi_list = pre_addi_df["data"][pre_addi_df["dest"]=="participants"].values.tolist()
+    scan_addi_list = pre_addi_df["data"][pre_addi_df["dest"]=="scans"].values.tolist()
+
+    addi_df=pd.DataFrame([participants_addi_list,session_addi_list, scan_addi_list]).transpose()
+    addi_df.columns = ["participants", "sessions", "scans"]
     if not gif:
         df_ref = df_ref.head(8)
+    df_to_write = pd.concat([df_ref, addi_df])
     return {
-        col: complete_data_df.filter(items=list(df_ref[col]))
+        col: complete_data_df.filter(items=list(df_to_write[col]))
         for col in ["participants", "sessions", "scans"]
     }