fix(io): preserve acq_time timezone in scans.tsv repair

datetime.fromisoformat() parses RFC 3339 'Z' suffix as tz-aware UTC,
but the strftime template used to renormalize the cell had no %z
placeholder — so every load silently rewrote "...Z" to "..." and
persisted the tz-stripped form back to *_scans.tsv. The
"if normalized != ts" guard could not suppress this because strftime
output structurally cannot contain 'Z', so every UTC-suffixed row
took the "differ" branch.

Split the normalization on dt.tzinfo: tz-aware inputs round-trip
through "%Y-%m-%dT%H:%M:%S.%f%z" with the offset preserved (UTC
restored to literal 'Z'; other offsets get ISO 8601 extended-form
colon). Naive inputs and the unparseable -> "n/a" path are unchanged.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Author: AmanJaiswal1503

eegdash/dataset/io.py

@@ -1046,7 +1046,20 @@ def _repair_scans_tsv_timestamps(data_dir: Path) -> bool:
                             # and 6-digit microseconds — mne-bids uses
                             # strptime('%Y-%m-%dT%H:%M:%S.%f') which
                             # requires 'T' and fractional seconds.
-                            normalized = dt.strftime("%Y-%m-%dT%H:%M:%S.%f")
+                            # For tz-aware inputs, preserve the tz: a
+                            # plain "%Y-%m-%dT%H:%M:%S.%f" silently
+                            # strips trailing 'Z' / '+HH:MM' offsets.
+                            if dt.tzinfo is not None:
+                                normalized = dt.strftime("%Y-%m-%dT%H:%M:%S.%f%z")
+                                # %z emits +HHMM; restore RFC 3339 'Z'
+                                # for UTC, otherwise insert the ISO
+                                # 8601 extended-form colon.
+                                if normalized.endswith("+0000"):
+                                    normalized = normalized[:-5] + "Z"
+                                else:
+                                    normalized = normalized[:-2] + ":" + normalized[-2:]
+                            else:
+                                normalized = dt.strftime("%Y-%m-%dT%H:%M:%S.%f")
                             if normalized != ts:
                                 cols[acq_idx] = normalized
                                 changed = True

tests/unit_tests/dataset/test_io.py

@@ -385,8 +385,30 @@ def test_read_bids_channels_tsv_parametrized(tmp_path, rows, expected):
         ("2024-01-02 03:04:05", "2024-01-02T03:04:05.000000", True),
         ("2024-01-02T03:04:99", "n/a", True),
         ("2024-01-02T03:04:05.000000", "2024-01-02T03:04:05.000000", False),
+        (
+            "2024-01-02T03:04:05.000000Z",
+            "2024-01-02T03:04:05.000000Z",
+            False,
+        ),
+        (
+            "2024-01-02T03:04:05.000000+00:00",
+            "2024-01-02T03:04:05.000000Z",
+            True,
+        ),
+        (
+            "2024-01-02T03:04:05.000000-05:00",
+            "2024-01-02T03:04:05.000000-05:00",
+            False,
+        ),
+    ],
+    ids=[
+        "normalize_space_separator",
+        "invalid_timestamp_to_na",
+        "already_normalized",
+        "preserve_z_suffix_utc",
+        "normalize_plus_zero_offset_to_z",
+        "preserve_negative_offset",
     ],
-    ids=["normalize_space_separator", "invalid_timestamp_to_na", "already_normalized"],
 )
 def test_repair_scans_tsv_timestamps_parametrized(
     tmp_path, acq_time, expected_time, expected_repaired