create physioevents.tsv.gz file

.gitignore

@@ -4,6 +4,7 @@ eye2bids/_version.py
 *events.json
 *eyetrack.json
 *eyetrack.tsv
+*.ipynb
 2eyes.ipynb
 tmp
 

eye2bids/edf2bids.py

@@ -4,6 +4,7 @@
 
 import gzip
 import json
+import re
 import subprocess
 from pathlib import Path
 
@@ -323,6 +324,111 @@ def _load_asc_file_as_reduced_df(events_asc_file: str | Path) -> pd.DataFrame:
     return pd.DataFrame(df_ms.iloc[0:, 2:])
 
 
+def _df_events_after_start(events: list) -> pd.DataFrame:
+
+    start_index = next(
+        i for i, line in enumerate(events) if re.match(r"START\s+.*", line)
+    )
+    end_index = next(
+        i for i in range(len(events) - 1, -1, -1) if re.match(r"END\s+.*", events[i])
+    )
+
+    if end_index > start_index:
+        data_lines = events[start_index + 1 : end_index]
+        return pd.DataFrame([line.strip().split("\t") for line in data_lines])
+    else:
+        return print("No 'END' found after the selected 'START'.")
+
+
+def _df_physioevents(events_after_start: pd.DataFrame) -> pd.DataFrame:
+    events_after_start["Event_Letters"] = (
+        events_after_start[0].str.extractall(r"([A-Za-z]+)").groupby(level=0).agg("".join)
+    )
+    events_after_start["Event_Numbers"] = events_after_start[0].str.extract(r"(\d+)")
+    events_after_start[["msg_timestamp", "message"]] = events_after_start[1].str.split(
+        n=1, expand=True
+    )
+    events_after_start["message"] = events_after_start["message"].astype(str)
+
+    msg_mask = events_after_start["Event_Letters"] == "MSG"
+    events_after_start.loc[msg_mask, "Event_Numbers"] = events_after_start.loc[
+        msg_mask, "msg_timestamp"
+    ]
+    physioevents_reordered = (
+        pd.concat(
+            [
+                events_after_start["Event_Numbers"],
+                events_after_start[2],
+                events_after_start["Event_Letters"],
+                events_after_start["message"],
+            ],
+            axis=1,
+            ignore_index=True,
+        )
+        .replace({None: np.nan, "None": np.nan})
+        .rename(columns={0: "timestamp", 1: "duration", 2: "trial_type", 3: "message"})
+    )
+    return physioevents_reordered
+
+
+def _physioevents_eye1(physioevents_reordered: pd.DataFrame) -> pd.DataFrame:
+    physioevents_eye1_list = ["MSG", "EFIXL", "ESACCL", "EBLINKL"]
+    physioevents_eye1 = physioevents_reordered[
+        physioevents_reordered["trial_type"].isin(physioevents_eye1_list)
+    ]
+    physioevents_eye1 = physioevents_eye1.replace(
+        {"EFIXL": "fixation", "ESACCL": "saccade", "MSG": np.nan, None: np.nan}
+    )
+
+    physioevents_eye1["blink"] = 0
+    last_non_na_trial_type = None
+
+    for i in range(len(physioevents_eye1)):
+        current_trial_type = physioevents_eye1.iloc[i]["trial_type"]
+        if pd.notna(current_trial_type):
+            if current_trial_type == "saccade" and last_non_na_trial_type == "EBLINKL":
+                physioevents_eye1.iloc[i, physioevents_eye1.columns.get_loc("blink")] = 1
+            last_non_na_trial_type = current_trial_type
+
+    physioevents_eye1.loc[physioevents_eye1["trial_type"].isna(), "blink"] = np.nan
+    physioevents_eye1["blink"] = physioevents_eye1["blink"].astype("object")
+    physioevents_eye1 = physioevents_eye1[physioevents_eye1.trial_type != "EBLINKL"]
+
+    physioevents_eye1 = physioevents_eye1[
+        ["timestamp", "duration", "trial_type", "blink", "message"]
+    ]
+    return physioevents_eye1
+
+
+def _physioevents_eye2(physioevents_reordered: pd.DataFrame) -> pd.DataFrame:
+    physioevents_eye2_list = ["MSG", "EFIXR", "ESACCR", "EBLINKR"]
+    physioevents_eye2 = physioevents_reordered[
+        physioevents_reordered["trial_type"].isin(physioevents_eye2_list)
+    ]
+    physioevents_eye2 = physioevents_eye2.replace(
+        {"EFIXR": "fixation", "ESACCR": "saccade", "MSG": np.nan, None: np.nan}
+    )
+
+    physioevents_eye2["blink"] = 0
+    last_non_na_trial_type = None
+
+    for i in range(len(physioevents_eye2)):
+        current_trial_type = physioevents_eye2.iloc[i]["trial_type"]
+        if pd.notna(current_trial_type):
+            if current_trial_type == "saccade" and last_non_na_trial_type == "EBLINKR":
+                physioevents_eye2.iloc[i, physioevents_eye2.columns.get_loc("blink")] = 1
+            last_non_na_trial_type = current_trial_type
+
+    physioevents_eye2.loc[physioevents_eye2["trial_type"].isna(), "blink"] = np.nan
+    physioevents_eye2["blink"] = physioevents_eye2["blink"].astype("object")
+    physioevents_eye2 = physioevents_eye2[physioevents_eye2.trial_type != "EBLINKR"]
+
+    physioevents_eye2 = physioevents_eye2[
+        ["timestamp", "duration", "trial_type", "blink", "message"]
+    ]
+    return physioevents_eye2
+
+
 def edf2bids(
     input_file: str | Path | None = None,
     metadata_file: str | Path | None = None,
@@ -358,10 +464,8 @@ def edf2bids(
 
     # eye-physio.json Metadata
     base_json = {
+        "PhysioType": "eyetrack",
         "Columns": ["x_coordinate", "y_coordinate", "pupil_size", "timestamp"],
-        "timestamp": {
-            "Description": "Timestamp issued by the eye-tracker indexing the continuous recordings corresponding to the sampled eye."
-        },
         "x_coordinate": {
             "Description": "Gaze position x-coordinate of the recorded eye, in the coordinate units specified in the corresponding metadata sidecar.",
             "Units": "a.u.",
@@ -374,11 +478,14 @@ def edf2bids(
             "Description": "Pupil area of the recorded eye as calculated by the eye-tracker in arbitrary units (see EyeLink's documentation for conversion).",
             "Units": "a.u.",
         },
+        "timestamp": {
+            "Description": "Timestamp issued by the eye-tracker indexing the continuous recordings corresponding to the sampled eye."
+        },
         "Manufacturer": "SR-Research",
         "ManufacturersModelName": _extract_ManufacturersModelName(events),
         "DeviceSerialNumber": _extract_DeviceSerialNumber(events),
         "EnvironmentCoordinates": metadata.get("EnvironmentCoordinates"),
-        "SoftwareVersion": metadata.get("SoftwareVersion"),
+        "SoftwareVersions": metadata.get("SoftwareVersion"),
         "EyeCameraSettings": metadata.get("EyeCameraSettings"),
         "EyeTrackerDistance": metadata.get("EyeTrackerDistance"),
         "FeatureDetectionSettings": metadata.get("FeatureDetectionSettings"),
@@ -503,13 +610,13 @@ def edf2bids(
 
     samples = pd.read_csv(samples_asc_file, sep="\t", header=None)
     samples_eye1 = (
-        pd.DataFrame(samples.iloc[:, 0:4])
+        pd.DataFrame(samples.iloc[:, [2, 1, 3, 0]])
         .map(lambda x: x.strip() if isinstance(x, str) else x)
         .replace(".", np.nan, regex=False)
     )
 
     if _2eyesmode(df_ms_reduced) == True:
-        samples_eye2 = pd.DataFrame(samples.iloc[:, [0, 4, 5, 6]])
+        samples_eye2 = pd.DataFrame(samples.iloc[:, [4, 5, 6, 0]])
 
     # Samples to eye_physio.tsv.gz
 
@@ -539,7 +646,63 @@ def edf2bids(
 
         e2b_log.info(f"file generated: {output_filename_eye2}")
 
-    # Messages and events to physioevents.tsv.gz - tbc
+    # Messages and events to dataframes
+
+    events_after_start = _df_events_after_start(events)
+    physioevents_reordered = _df_physioevents(events_after_start)
+    physioevents_eye1 = _physioevents_eye1(physioevents_reordered)
+    physioevents_eye2 = _physioevents_eye2(physioevents_reordered)
+
+    # Messages and events to physioevents.tsv.gz
+
+    if _2eyesmode(df_ms_reduced) == False:
+        output_eventsfilename_eye1 = generate_output_filename(
+            output_dir=output_dir,
+            input_file=input_file,
+            suffix="_recording-eye1_physioevents",
+            extension="tsv.gz",
+        )
+        if _extract_RecordedEye(df_ms_reduced) == "Left":
+            content = physioevents_eye1.to_csv(
+                sep="\t", index=False, na_rep="n/a", header=None
+            )
+        elif _extract_RecordedEye(df_ms_reduced) == "Right":
+            content = physioevents_eye2.to_csv(
+                sep="\t", index=False, na_rep="n/a", header=None
+            )
+        with gzip.open(output_eventsfilename_eye1, "wb") as f:
+            f.write(content.encode())
+
+        e2b_log.info(f"file generated: {output_eventsfilename_eye1}")
+
+    else:
+        output_eventsfilename_eye1 = generate_output_filename(
+            output_dir=output_dir,
+            input_file=input_file,
+            suffix="_recording-eye1_physioevents",
+            extension="tsv.gz",
+        )
+        content = physioevents_eye1.to_csv(
+            sep="\t", index=False, na_rep="n/a", header=None
+        )
+        with gzip.open(output_eventsfilename_eye1, "wb") as f:
+            f.write(content.encode())
+
+        e2b_log.info(f"file generated: {output_eventsfilename_eye1}")
+
+        output_eventsfilename_eye2 = generate_output_filename(
+            output_dir=output_dir,
+            input_file=input_file,
+            suffix="_recording-eye2_physioevents",
+            extension="tsv.gz",
+        )
+        content = physioevents_eye2.to_csv(
+            sep="\t", index=False, na_rep="n/a", header=None
+        )
+        with gzip.open(output_eventsfilename_eye2, "wb") as f:
+            f.write(content.encode())
+
+        e2b_log.info(f"file generated: {output_eventsfilename_eye2}")
 
 
 def generate_output_filename(