File size: 6,071 Bytes
e4eb88a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
#!/usr/bin/env python3
"""
Reprocess LIBERO LeRobot dataset to mesh-only tracks using simulator vertices.

This script keeps core fields (image, wrist_image, state, actions, task) and writes
only mesh-point tracks for both views, using per-task BDDL dynamic gripper selection.
"""

from __future__ import annotations

import openpi.shared.local_cache_bootstrap  # noqa: F401

import argparse
import os
import shutil
import sys
from pathlib import Path

import numpy as np

from lerobot.common.datasets.lerobot_dataset import LeRobotDataset
import lerobot_preprocess_cotracker as cot


def _build_features() -> dict[str, dict]:
    return {
        "image": {"dtype": "image", "shape": (256, 256, 3), "names": ["height", "width", "channel"]},
        "wrist_image": {"dtype": "image", "shape": (256, 256, 3), "names": ["height", "width", "channel"]},
        "state": {"dtype": "float32", "shape": (8,), "names": ["state"]},
        "actions": {"dtype": "float32", "shape": (7,), "names": ["actions"]},
        "agentview_tracks": {"dtype": "float32", "shape": (7, 2), "names": ["points", "xy"]},
        "agentview_vis": {"dtype": "float32", "shape": (7,), "names": ["points"]},
        "wrist_tracks": {"dtype": "float32", "shape": (7, 2), "names": ["points", "xy"]},
        "wrist_vis": {"dtype": "float32", "shape": (7,), "names": ["points"]},
        "agentview_mesh_vertices_2d": {"dtype": "float32", "shape": (7, 2), "names": ["points", "xy"]},
        "wrist_mesh_vertices_2d": {"dtype": "float32", "shape": (7, 2), "names": ["points", "xy"]},
        "has_track_mesh": {"dtype": "float32", "shape": (1,), "names": ["flag"]},
    }


def process_episode(ds: LeRobotDataset, ep_idx: int):
    bnds = cot._episode_bounds(ds, ep_idx)
    scene = cot._scene_from_task(bnds.task)
    frames = []
    for i in range(bnds.start, bnds.end):
        row = ds[i]
        frames.append(
            (
                cot._to_hwc_uint8(np.asarray(row["image"])),
                cot._to_hwc_uint8(np.asarray(row["wrist_image"])),
                np.asarray(row["state"], dtype=np.float32),
                np.asarray(row["actions"], dtype=np.float32),
                row["task"],
            )
        )

    images = np.stack([f[0] for f in frames], axis=0)
    wrist_images = np.stack([f[1] for f in frames], axis=0)
    states = np.stack([f[2] for f in frames], axis=0)
    actions = np.stack([f[3] for f in frames], axis=0)
    task = frames[0][4]

    agent_mesh_seq, wrist_mesh_seq = cot._get_mesh_sequence_from_sim(
        scene,
        states[0],
        actions,
        task_name=task,
        img_hw=(images.shape[1], images.shape[2]),
    )

    T = min(images.shape[0], agent_mesh_seq.shape[0], wrist_mesh_seq.shape[0])
    for t in range(T):
        yield {
            "image": images[t],
            "wrist_image": wrist_images[t],
            "state": states[t],
            "actions": actions[t],
            "task": task,
            "agentview_tracks": agent_mesh_seq[t],
            "agentview_vis": np.ones((7,), dtype=np.float32),
            "wrist_tracks": wrist_mesh_seq[t],
            "wrist_vis": np.ones((7,), dtype=np.float32),
            "agentview_mesh_vertices_2d": agent_mesh_seq[t],
            "wrist_mesh_vertices_2d": wrist_mesh_seq[t],
            "has_track_mesh": np.asarray([1.0], dtype=np.float32),
        }


def main():
    p = argparse.ArgumentParser(description="Reprocess LIBERO to mesh-only (both views, dynamic BDDL).")
    p.add_argument("--source-repo-id", default="/mnt/kevin/data/physical-intelligence/libero")
    p.add_argument("--target-repo-id", default="/mnt/kevin/data/physical-intelligence/libero_mesh_only_dynamic")
    p.add_argument("--overwrite", action="store_true")
    p.add_argument("--max-episodes", type=int, default=None)
    p.add_argument("--start-episode", type=int, default=0)
    p.add_argument("--end-episode", type=int, default=None)
    p.add_argument("--libero-root", type=str, default=None)
    p.add_argument("--libero-data", type=str, default=None)
    p.add_argument(
        "--extra-libero-path",
        type=str,
        default="/mnt/kevin/code/wmrl/howard-branch/openpi/third_party/libero/libero",
    )
    args = p.parse_args()

    cot.EXTRA_LIBERO_PATH = args.extra_libero_path
    if args.libero_root and (Path(args.libero_root) / "libero" / "envs" / "mesh_vertex_wrapper.py").exists():
        cot.EXTRA_LIBERO_PATH = args.libero_root

    for pth in [args.libero_root, args.extra_libero_path]:
        if not pth:
            continue
        candidates = [Path(pth), Path(pth) / "libero"]
        for cand in candidates:
            s = str(cand)
            if cand.exists() and s not in sys.path:
                sys.path.insert(0, s)

    if args.libero_data:
        os.environ.setdefault("LIBERO_PATH", args.libero_data)
    os.environ.setdefault("MUJOCO_GL", "egl")
    os.environ.setdefault("PYOPENGL_PLATFORM", "egl")

    src = LeRobotDataset(args.source_repo_id)
    total_eps = len(src.meta.episodes)
    start_ep = max(0, int(args.start_episode))
    end_ep = total_eps if args.end_episode is None else min(int(args.end_episode), total_eps)
    if args.max_episodes is not None:
        end_ep = min(end_ep, start_ep + int(args.max_episodes))
    if end_ep <= start_ep:
        raise ValueError(f"Invalid episode range [{start_ep}, {end_ep})")

    target_path = Path(args.target_repo_id)
    if target_path.exists() and args.overwrite:
        shutil.rmtree(target_path)

    dst = LeRobotDataset.create(
        repo_id=str(target_path),
        robot_type="panda",
        fps=src.fps,
        features=_build_features(),
        image_writer_threads=10,
        image_writer_processes=5,
    )

    for ep_idx in range(start_ep, end_ep):
        print(f"[ep {ep_idx-start_ep}/{end_ep-start_ep}] abs={ep_idx}")
        for frame in process_episode(src, ep_idx):
            dst.add_frame(frame)
        dst.save_episode()

    print(f"Done. Wrote {end_ep-start_ep} episodes ({start_ep}:{end_ep}) to {target_path}")


if __name__ == "__main__":
    main()