The entire repository was built as a docker image and the pipeline was run on a docker container. I had to face a lot of issues with installations of pangolin and g2o. Further, there were import issues with import of the python bindings. Also, there were issues with forwarding of GUI components (example a pangolin display or matlotlib plot) from inside the container to the host machine. This markdown serves as a guide on how to build the entire repository using docker and run the pipeline as if you were running it on your local without any heachaches involved in building the dependencies
Although, I have supplied a Dockerfile for building end-to-end the application. Here, I am noting few steps that I undertook to build it step-by-step for future reference.
Step 1. Build the g2o
git clone https://github.com/ElliotHYLee/g2opy
cd g2opy
docker build -t env_g2o:latest .
You will be able to see a folder g2opy inside /home/ directory
*Step 2. Build Pangolin on top of g2o docker image
git clone https://github.com/sakshamjindal/Pangolin-Python
cd Pangolin-Python
docker build -t env_g2o_pangolin:latest .
You will be able to see a folder pangolin inside /home/ directory
Step 3. Mount the local directory and forward GUI components from inside the container
xhost +local:docker
docker run -dit -P --name slam_box --net=host -e DISPLAY -v /tmp/.X11-unix -v ~/mono-slam:/home/slam -v ~/KITTI:/home/KITTI env_g2o_pangolin:latest
docker exec -it slam_box bash
You will be able to see your working directory ~/mono-slam and ~/KITTI inside /home/ directory
Some links that were helpful in step 3 were a youtube video