code/scripts/recompute_oven_visibility_pregrasp.py

from pathlib import Path
import argparse
import json
import sys
from typing import Dict, Optional

import pandas as pd


PROJECT_ROOT = Path(__file__).resolve().parents[1]
if str(PROJECT_ROOT) not in sys.path:
    sys.path.insert(0, str(PROJECT_ROOT))

from rr_label_study.oven_study import (
    BimanualTakeTrayOutOfOven,
    ReplayCache,
    _aggregate_summary,
    _annotate_phase_columns,
    _analyze_episode,
    _derive_templates,
    _episode_metrics_from_frames,
    _keyframe_subset,
    _keypoint_discovery,
    _launch_replay_env,
    _load_demo,
    _load_descriptions,
    _pregrasp_progress_and_distance,
    _pregrasp_score_and_success,
    _frame_metrics,
)


def _recompute_columns(
    episode_dir: Path,
    templates,
    checkpoint_stride: int,
    base_df: pd.DataFrame,
) -> pd.DataFrame:
    demo = _load_demo(episode_dir)
    num_frames = min(len(demo), len(base_df))
    frame_df = base_df.iloc[:num_frames].copy()

    env = _launch_replay_env()
    try:
        task = env.get_task(BimanualTakeTrayOutOfOven)
        cache = ReplayCache(task, demo, checkpoint_stride=checkpoint_stride)
        cache.reset()

        for frame_index in range(num_frames):
            cache.step_to(frame_index)
            state = cache.current_state()
            visibility = _frame_metrics(episode_dir, demo, state, templates)
            pregrasp_progress, pregrasp_distance = _pregrasp_progress_and_distance(
                state.left_gripper_pose,
                state.tray_pose,
                templates,
            )
            p_pre, y_pre = _pregrasp_score_and_success(task, templates)

            frame_df.at[frame_index, "frame_index"] = frame_index
            frame_df.at[frame_index, "time_norm"] = frame_index / max(1, num_frames - 1)
            frame_df.at[frame_index, "door_angle"] = state.door_angle
            frame_df.at[frame_index, "right_gripper_open"] = state.right_gripper_open
            frame_df.at[frame_index, "left_gripper_open"] = state.left_gripper_open
            frame_df.at[frame_index, "pregrasp_progress"] = pregrasp_progress
            frame_df.at[frame_index, "pregrasp_distance"] = pregrasp_distance
            frame_df.at[frame_index, "p_pre"] = p_pre
            frame_df.at[frame_index, "y_pre_raw"] = float(bool(y_pre))
            frame_df.at[frame_index, "y_pre"] = float(bool(y_pre))
            for key, value in visibility.items():
                frame_df.at[frame_index, key] = value

            if (frame_index + 1) % 25 == 0 or (frame_index + 1) == num_frames:
                print(
                    f"[{episode_dir.name}] recomputed {frame_index + 1}/{num_frames} dense frames",
                    flush=True,
                )
        return frame_df
    finally:
        env.shutdown()


def main() -> int:
    parser = argparse.ArgumentParser()
    parser.add_argument("--dataset-root", required=True)
    parser.add_argument("--episode-dir", required=True)
    parser.add_argument("--input-dense-csv", required=True)
    parser.add_argument("--output-dir", required=True)
    parser.add_argument("--checkpoint-stride", type=int, default=16)
    parser.add_argument("--template-episode-dir")
    args = parser.parse_args()

    dataset_root = Path(args.dataset_root)
    episode_dir = Path(args.episode_dir)
    output_dir = Path(args.output_dir)
    output_dir.mkdir(parents=True, exist_ok=True)

    base_df = pd.read_csv(args.input_dense_csv)
    demo = _load_demo(episode_dir)
    descriptions = _load_descriptions(episode_dir)

    template_episode_dir = (
        Path(args.template_episode_dir) if args.template_episode_dir else episode_dir
    )
    templates, template_frames = _derive_templates(dataset_root, template_episode_dir)
    with output_dir.joinpath("templates.json").open("w", encoding="utf-8") as handle:
        json.dump(
            {
                "templates": templates.to_json(),
                "template_episode": template_episode_dir.name,
                "template_frames": template_frames,
            },
            handle,
            indent=2,
        )

    frame_df = _recompute_columns(
        episode_dir=episode_dir,
        templates=templates,
        checkpoint_stride=args.checkpoint_stride,
        base_df=base_df,
    )
    frame_df = _annotate_phase_columns(frame_df)

    keyframes = [index for index in _keypoint_discovery(demo) if index < len(frame_df)]
    key_df = _keyframe_subset(frame_df, keyframes)
    metrics = _episode_metrics_from_frames(
        frame_df=frame_df,
        key_df=key_df,
        episode_name=episode_dir.name,
        description=descriptions[0],
        interventions={},
    )

    frame_df.to_csv(output_dir.joinpath(f"{episode_dir.name}.dense.csv"), index=False)
    key_df.to_csv(output_dir.joinpath(f"{episode_dir.name}.keyframes.csv"), index=False)
    with output_dir.joinpath(f"{episode_dir.name}.metrics.json").open("w", encoding="utf-8") as handle:
        json.dump(metrics, handle, indent=2)
    summary = _aggregate_summary([metrics])
    with output_dir.joinpath("summary.json").open("w", encoding="utf-8") as handle:
        json.dump(summary, handle, indent=2)
    print(json.dumps(summary, indent=2))
    return 0


if __name__ == "__main__":
    raise SystemExit(main())