Add lean exporter

exporters/lean.py

@@ -0,0 +1,68 @@
+#
+# SPDX-FileCopyrightText: 2026 Stanford University, ETH Zurich, and the project authors (see CONTRIBUTORS.md)
+# SPDX-FileCopyrightText: 2026 This source file is part of the SensorTSLM open-source project.
+#
+# SPDX-License-Identifier: MIT
+#
+from __future__ import annotations
+
+from pathlib import Path
+
+import pyarrow as pa
+import pyarrow.feather as feather
+
+from timef.schema import CaptionResult, Recording
+
+
+SCHEMA_VERSION = "1"
+
+
+def write_caption_result(
+    result: CaptionResult,
+    path: Path,
+    *,
+    rows_per_shard: int = 4096,
+    compression: str | None = "zstd",
+) -> None:
+    path = Path(path)
+    path.mkdir(parents=True, exist_ok=True)
+    rows = result.rows
+    if not rows:
+        return
+    for shard_idx, start in enumerate(range(0, len(rows), rows_per_shard)):
+        shard_rows = rows[start : start + rows_per_shard]
+        dicts = [_recording_to_dict(rec) for rec in shard_rows]
+        table = pa.Table.from_pylist(dicts)
+        table = table.replace_schema_metadata(
+            {"sensortslm.schema_version": SCHEMA_VERSION}
+        )
+        shard_path = path / f"recordings_{shard_idx:04d}.arrow"
+        feather.write_feather(table, shard_path, compression=compression)
+
+
+def _recording_to_dict(rec: Recording) -> dict:
+    return {
+        "row_id": rec.row_id,
+        "user_id": rec.user_id,
+        "date": rec.date,
+        "values": rec.values.tolist(),
+        "channel_names": list(rec.channel_names),
+        "display_names": list(rec.display_names),
+        "units": list(rec.units),
+        "has_any_data": rec.has_any_data.tolist(),
+        "minutes_nonzero_or_nan": rec.minutes_nonzero_or_nan.tolist(),
+        "channel_variance": rec.channel_variance.tolist(),
+        "total_nonwear_minutes": rec.total_nonwear_minutes,
+        "wear_pct": rec.wear_pct,
+        "annotations": [
+            {
+                "caption_type": a.caption_type,
+                "text": a.text,
+                "channel_idxs": list(a.channel_idxs),
+                "window_start": a.window[0] if a.window else None,
+                "window_end": a.window[1] if a.window else None,
+                "label": None if a.label is None else str(a.label),
+            }
+            for a in rec.annotations
+        ],
+    }