- Updated `.gitlab-ci.yml` to use `all-deps.repos` and install additional Python dependencies via `requirements.txt`.
- Enhanced `Dockerfile` with better dependency management:
- Added `SHELL` declaration for consistent script execution.
- Installed `libgeos-dev` for geometry processing.
- Included `python3-rosdep` and `python3-colcon-common-extensions` for ROS2 tooling.
- Streamlined ROS2 and dependency installation steps.
- Refined `requirements.txt` to adjust package versions and include new dependencies such as `tensorflow` and `torch`.
- Updated repository URL for `rbs_assets_library` in `requirements.txt`.
These changes improve dependency resolution, enhance build reproducibility, and align with updated requirements.
- Added a new `real_robot` launch argument with default set to "false".
- Updated `launch_setup` to handle `real_robot` configuration, allowing selective use of `rbs_robot_setup`.
- Updated `requirements.txt` to replace a static library reference with a GitHub repository link for `rbs_assets_library`.
- Switched base image in `Dockerfile` to `nvidia/cuda:12.6.2-cudnn-devel-ubuntu22.04`.
- Added non-interactive mode for Debian and streamlined package installation.
- Removed unused JSON library build steps.
- Enhanced ROS2 installation with keyring and repository setup.
- Simplified Python package and framework dependencies installation.
- Added commands for `git-lfs` installation and cloning `rbs_assets_library`.
- Introduced a new Russian installation guide (`doc/ru/installation.md`) with detailed steps for setting up the framework and dependencies.
- Fixed import paths in `octree.py` for consistency with project structure.
- Added `all-deps.repos` and `requirements.txt` for managing dependencies.
