SoftGroup/dataset/s3dis/prepare_data_inst.py

# Copyright (c) Gorilla-Lab. All rights reserved.
import os
import os.path as osp
import glob
import argparse
from warnings import warn

import torch
import numpy as np
import pandas as pd
from sklearn.neighbors import NearestNeighbors

# import gorilla

ROOM_TYPES = {
    "conferenceRoom": 0,
    "copyRoom": 1,
    "hallway": 2,
    "office": 3,
    "pantry": 4,
    "WC": 5,
    "auditorium": 6,
    "storage": 7,
    "lounge": 8,
    "lobby": 9,
    "openspace": 10,
}

INV_OBJECT_LABEL = {
    0: "ceiling",
    1: "floor",
    2: "wall",
    3: "beam",
    4: "column",
    5: "window",
    6: "door",
    7: "chair",
    8: "table",
    9: "bookcase",
    10: "sofa",
    11: "board",
    12: "clutter",
}

OBJECT_LABEL = {name: i for i, name in INV_OBJECT_LABEL.items()}


def object_name_to_label(object_class):
    r"""convert from object name in S3DIS to an int"""
    object_label = OBJECT_LABEL.get(object_class, OBJECT_LABEL["clutter"])
    return object_label


# modify from https://github.com/nicolas-chaulet/torch-points3d/blob/master/torch_points3d/datasets/segmentation/s3dis.py
def read_s3dis_format(area_id: str,
                      room_name: str,
                      data_root: str = "./",
                      label_out: bool = True,
                      verbose: bool = False):
    r"""
    extract data from a room folder
    """
    room_type = room_name.split("_")[0]
    room_label = ROOM_TYPES[room_type]
    room_dir = osp.join(data_root, area_id, room_name)
    raw_path = osp.join(room_dir, f"{room_name}.txt")

    room_ver = pd.read_csv(raw_path, sep=" ", header=None).values
    xyz = np.ascontiguousarray(room_ver[:, 0:3], dtype="float32")
    rgb = np.ascontiguousarray(room_ver[:, 3:6], dtype="uint8")
    if not label_out:
        return xyz, rgb
    n_ver = len(room_ver)
    del room_ver
    nn = NearestNeighbors(n_neighbors=1, algorithm="kd_tree").fit(xyz)
    semantic_labels = np.zeros((n_ver, ), dtype="int64")
    room_label = np.asarray([room_label])
    instance_labels = np.ones((n_ver, ), dtype="int64") * -100
    objects = glob.glob(osp.join(room_dir, "Annotations", "*.txt"))
    i_object = 1
    for single_object in objects:
        object_name = os.path.splitext(os.path.basename(single_object))[0]
        if verbose:
            print(f"adding object {i_object} : {object_name}")
        object_class = object_name.split("_")[0]
        object_label = object_name_to_label(object_class)
        obj_ver = pd.read_csv(single_object, sep=" ", header=None).values
        _, obj_ind = nn.kneighbors(obj_ver[:, 0:3])
        semantic_labels[obj_ind] = object_label
        # if object_label < 3: # background object
        #     continue
        instance_labels[obj_ind] = i_object
        i_object = i_object + 1

    return (
        xyz,
        rgb,
        semantic_labels,
        instance_labels,
        room_label,
    )


def get_parser():
    parser = argparse.ArgumentParser(description="s3dis data prepare")
    parser.add_argument("--data-root",
                        type=str,
                        default="./Stanford3dDataset_v1.2",
                        help="root dir save data")
    parser.add_argument("--save-dir",
                        type=str,
                        default="./preprocess",
                        help="directory save processed data")
    parser.add_argument(
        "--patch",
        action="store_true",
        help="patch data or not (just patch at first time running)")
    parser.add_argument("--align",
                        action="store_true",
                        help="processing aligned dataset or not")
    parser.add_argument("--verbose",
                        action="store_true",
                        help="show processing room name or not")

    args_cfg = parser.parse_args()

    return args_cfg


# patch -ruN -p0 -d  raw < s3dis.patch
if __name__ == "__main__":
    args = get_parser()
    data_root = args.data_root
    # processed data output dir
    save_dir = args.save_dir
    os.makedirs(save_dir, exist_ok=True)
    if args.patch:
        if args.align:  # processing aligned s3dis dataset
            # os.system(f"cd {data_root} && git apply {osp.join(osp.dirname(__file__), 's3dis_align.diff')}")
            os.system(
                f"patch -ruN -p0 -d  {data_root} < {osp.join(osp.dirname(__file__), 's3dis_align.patch')}"
            )
            # rename to avoid room_name conflict
            if osp.exists(
                    osp.join(data_root, "Area_6", "copyRoom_1",
                             "copy_Room_1.txt")):
                os.rename(
                    osp.join(data_root, "Area_6", "copyRoom_1",
                             "copy_Room_1.txt"),
                    osp.join(data_root, "Area_6", "copyRoom_1",
                             "copyRoom_1.txt"))
        else:
            os.system(
                f"patch -ruN -p0 -d  {data_root} < {osp.join(osp.dirname(__file__), 's3dis.patch')}"
            )

    area_list = ["Area_1", "Area_2", "Area_3", "Area_4", "Area_5", "Area_6"]
    # area_list = ['Area_1']
    # area_list = ['Area_2']
    # area_list = ['Area_3']
    # area_list = ['Area_4']
    # area_list = ['Area_5']
    # area_list = ['Area_6']

    for area_id in area_list:
        print(f"Processing: {area_id}")
        area_dir = osp.join(data_root, area_id)
        # get the room name list for each area
        room_name_list = os.listdir(area_dir)
        try:
            room_name_list.remove(f"{area_id}_alignmentAngle.txt")
            room_name_list.remove(".DS_Store")
        except:
            pass
        for room_name in room_name_list:
            scene = f"{area_id}_{room_name}"
            if args.verbose:
                print(f"processing: {scene}")
            save_path = osp.join(save_dir, scene + "_inst_nostuff.pth")
            if osp.exists(save_path):
                continue
            (xyz, rgb, semantic_labels, instance_labels,
             room_label) = read_s3dis_format(area_id, room_name, data_root)
            rgb = (rgb / 127.5) - 1
            torch.save((xyz, rgb, semantic_labels, instance_labels, room_label,
                        scene), save_path)