rosbags version

[ZJW-HUST]

dockerfile默认下载0.11版本的rosbags
读取vbr数据集时需修改如下：

from rosbags.typesys.stores.ros1_noetic import sensor_msgs__msg__PointCloud2 as PointCloud2
from rosbags.typesys.stores.ros1_noetic import sensor_msgs__msg__PointField as PointField
# from rosbags.typesys.types import sensor_msgs__msg__PointCloud2 as PointCloud2
# from rosbags.typesys.types import sensor_msgs__msg__PointField as PointField

[ZJW-HUST]

Hello , I am using Docker to run programs. I would like to ask how to use rerun for GUI since I am not familiar with it.
When I run rerun in host machine, the program crashes.

[EmanueleGiacomini]

Hi, by tomorrow I'll investigate and provide an answer, sorry for the delay

[EmanueleGiacomini]

Hi, concerning the first issue related to rosbags version, I'll push a fix on the Dockerfile to freeze the rosbags version to 0.10.9.

Concerning the second issue, Rerun GUI is not currently supported in Dockerized applications. I suggest installing rerun-sdk outside of Docker (be sure to match the versions) and enabling rerun logging in serve-grpc mode.

Within your configuration file:

logging:
  enable: true

  rerun_spawn: false                      # Do not spawn the GUI
  rerun_serve_grpc: true               # serve log-data over gRPC
  rerun_connect_grpc_url: false 

When launching Slam through Docker with:

python3 run.py slam <path_to_your_config>

You should notice the following (or similar) logging message:

11:03:28 INFO     rerun+http://127.0.0.1:9876/proxy                 rerun_logging.py:40

From outside the container, you can launch a GUI that connects with the in-container rerun server:

rerun --connect rerun+http://127.0.0.1:9876/proxy

This should allow you to visualize the logging data produced by the slam script.

EDIT: d82680d edits the Dockerfile to pin rosbags to version 0.10.9 and rerun-sdk to 0.23.3.