Public API for module isaacsim.replicator.writers:#

Classes#

  • class DataVisualizationWriter(Writer)

    • BB_2D_TIGHT: str

    • BB_2D_LOOSE: str

    • BB_3D: str

    • SUPPORTED_BACKGROUNDS: List

    • def init(self, output_dir: str, bounding_box_2d_tight: bool = False, bounding_box_2d_tight_params: dict = None, bounding_box_2d_loose: bool = False, bounding_box_2d_loose_params: dict = None, bounding_box_3d: bool = False, bounding_box_3d_params: dict = None, frame_padding: int = 4)

    • def write(self, data: dict)

    • def detach(self)

  • class DOPEWriter(Writer)

    • def init(self, output_dir: str, class_name_to_index_map: Dict, semantic_types: List[str] = None, image_output_format: str = ‘png’, use_s3: bool = False, bucket_name: str = ‘’, endpoint_url: str = ‘’, s3_region: str = ‘us-east-1’)

    • def register_pose_annotator(config_data: dict)

    • def setup_writer(config_data: dict, writer_config: dict)

    • def write(self, data: dict)

    • def is_last_frame_valid(self) -> bool

  • class PoseWriter(Writer)

    • RGB_ANNOT_NAME: str

    • BB3D_ANNOT_NAME: str

    • CAM_PARAMS_ANNOT_NAME: str

    • SUPPORTED_FORMATS: Unknown

    • CUBOID_KEYPOINTS_ORDER_DEFAULT: List

    • CUBOID_KEYPOINT_ORDER_DOPE: List

    • CUBOID_KEYPOINT_COLORS: List

    • CUBOID_EDGE_COLORS: Dict

    • def init(self, output_dir: str = None, use_subfolders: bool = False, visibility_threshold: float = 0.0, skip_empty_frames: bool = True, write_debug_images: bool = False, frame_padding: int = 6, format: str = None, use_s3: bool = False, s3_bucket: str = None, s3_endpoint_url: str = None, s3_region: str = None, backend: BaseBackend = None, image_output_format: str = ‘png’)

    • def write(self, data: dict)

    • def get_current_frame_id(self)

    • def detach(self)

  • class PytorchListener

    • def init(self)

    • def write_data(self, data: dict)

    • def get_rgb_data(self) -> torch.Tensor | None

  • class PytorchWriter(Writer)

    • def init(self, listener: PytorchListener, output_dir: str = None, tiled_sensor: bool = False, device: str = ‘cuda’)

    • def write(self, data: dict)

  • class YCBVideoWriter(Writer)

    • def init(self, output_dir: str, num_frames: int, semantic_types: List[str] = None, rgb: bool = False, bounding_box_2d_tight: bool = False, semantic_segmentation: bool = False, distance_to_image_plane: bool = False, image_output_format: str = ‘png’, pose: bool = False, class_name_to_index_map: Dict = None, factor_depth: int = 10000, intrinsic_matrix: np.ndarray = None)

    • def register_pose_annotator(config_data: dict)

    • def setup_writer(config_data: dict, writer_config: dict)

    • def write(self, data: dict)

    • def save_mesh_vertices(mesh_prim: UsdGeom.Mesh, coord_prim: Usd.Prim, model_name: str, output_folder: str)

    • def is_last_frame_valid(self) -> bool