update to use date_index

src/constantinople_lab_to_nwb/mah_2024/mah_2024_convert_all_sessions.py

@@ -2,8 +2,9 @@
 from datetime import datetime
 from pathlib import Path
 from typing import Union, List
+from warnings import warn
 
-from nwbinspector import inspect_all, format_messages, save_report
+import pandas as pd
 from pymatreader import read_mat
 from tqdm import tqdm
 
@@ -71,6 +72,50 @@ def _get_sessions_to_convert_from_mat(
     return bpod_files_to_convert
 
 
+def _get_date_index(bpod_file_path: Union[str, Path], a_struct_file_path: Union[str, Path]) -> Union[int, None]:
+    """
+    Figure out the date index for the processed behavior file.
+
+    Parameters
+    ----------
+    bpod_file_path: Union[str, Path]
+        Path to the raw Bpod output (.mat file).
+    a_struct_file_path: Union[str, Path]
+        Path to the processed behavior data (.mat file).
+
+    Returns
+    -------
+    int
+        The date index for the processed behavior file.
+    """
+    bpod_data = read_mat(str(bpod_file_path))
+    try:
+        bpod_session_data = bpod_data["SessionData"]
+    except KeyError:
+        warn(
+            f"'SessionData' key not found in '{bpod_file_path}'. The date index could not be determined from the file."
+        )
+        return None
+
+    num_trials = bpod_session_data["nTrials"]
+    date = bpod_session_data["Info"]["SessionDate"]
+
+    a_struct_data = read_mat(str(a_struct_file_path))
+    dates = a_struct_data["A"]["date"]
+    num_trials_per_day = a_struct_data["A"]["ntrials"]
+
+    dates_and_trials = pd.DataFrame(dict(date=dates, num_trials=num_trials_per_day))
+    filtered_dates_and_trials = dates_and_trials[
+        (dates_and_trials["date"] == date) & (dates_and_trials["num_trials"] == num_trials)
+    ]
+
+    if filtered_dates_and_trials.empty:
+        warn(f"Date index for '{date}' not found in '{a_struct_file_path}'.")
+        return None
+
+    return filtered_dates_and_trials.index[0]
+
+
 def sessions_to_nwb(
     raw_behavior_folder_path: Union[str, Path],
     processed_behavior_folder_path: Union[str, Path],
@@ -108,7 +153,7 @@ def sessions_to_nwb(
     processed_mat_files = list(processed_behavior_folder_path.glob("ratTrial*.mat"))
     subject_ids = [
         processed_behavior_file_path.stem.split("_")[-1] for processed_behavior_file_path in processed_mat_files
-    ][:10]
+    ]
     sessions_to_convert_per_subject = {
         subject_id: _get_sessions_to_convert_from_mat(
             file_path=processed_behavior_file_path, bpod_folder_path=raw_behavior_folder_path
@@ -128,14 +173,22 @@ def sessions_to_nwb(
         )
 
         for raw_behavior_file_path in progress_bar:
-            session_id = raw_behavior_file_path.stem.split("_", maxsplit=1)[1].replace("_", "-")
+            session_id = Path(raw_behavior_file_path).stem.split("_", maxsplit=1)[1].replace("_", "-")
             subject_nwb_folder_path = nwbfile_folder_path / f"sub-{subject_id}"
             if not subject_nwb_folder_path.exists():
                 os.makedirs(subject_nwb_folder_path, exist_ok=True)
             nwbfile_path = subject_nwb_folder_path / f"sub-{subject_id}_ses-{session_id}.nwb"
 
             if nwbfile_path.exists() and not overwrite:
-                print(f"Skipping existing NWB file: {nwbfile_path}")
+                continue
+
+            date_index = _get_date_index(
+                bpod_file_path=raw_behavior_file_path, a_struct_file_path=processed_behavior_file_path
+            )
+            if date_index is None:
+                print(
+                    f"Skipping '{subject_id}' session '{session_id}', session not found in the processed behavior file."
+                )
                 continue
 
             date_from_mat = session_id.split("-")[1]
@@ -149,6 +202,7 @@ def sessions_to_nwb(
             session_to_nwb(
                 raw_behavior_file_path=raw_behavior_file_path,
                 processed_behavior_file_path=processed_behavior_file_path,
+                date_index=date_index,
                 nwbfile_path=nwbfile_path,
                 column_name_mapping=column_name_mapping,
                 column_descriptions=column_descriptions,