hypothe/ai4ro2_2 repository overview

⁠Brief

This docker image contains code from popf-tif⁠ and Dr. Antony Thomas⁠ for the basic setup of the AI4RO2 second assignment. They are the ones to be thanked for all the code here presented! This minimal setup contains the planner and the files provided on Aulaweb, already compiled and built so you can start working right away.

Please notice that this image contains the barebones of the system, being built from the Ubuntu-18.04 Docker Official Image⁠ and, as such, no graphical interface (as is customary for docker images). To edit the planning files already contained in the image two options can be used:

• mount a shared directory between your host system and the container

  docker run -dit -v path/to_your/host/folder:path/to/container/folder  --name your_name hypothe/ai4ro2_2
  

  Copy
  This means you can modify the files (problem, domain, c++, anything you need) in your host system and, simply by placing them in the folder you shared, they will be visible from the docker in the destination folder you specified (but remember to correctly specify their path when calling the planner!)
• use Visual Studio Code⁠ to interact with an otherwise isolate docker container. Please refer to their tutorial "Developing inside a Container"⁠ to start getting familiar with the system. However, here are the broad steps of the workflow:
  • install the necessary docker extensions in your VSCode if you haven't already (Docker, Remote - Containers)
  • launch the container from the terminal (PowerShell or directly from within VSCode) with

      docker run -dit --name your_name hypothe/ai4ro2_2
  

  Copy

  NOTE: you can also start a container from within VSCode but it will be removed automatically when you stop it, hence losing all your data not saved elsewhere! If you are not sure of what you're doing don't risk it.

  • from VSCode click on the green icon in the lowermost left corner of the window, it will open a search bar: from there select Attach to Running Container and, once it loads, the name of the running container (eg. your_name) will appear; simply click on it.
  • a new VSCode window will open and, after few seconds, it will be connected to your running container: you can explore it from the terminal in VSCode itself, open files with its editor and even run the planner from that same terminal!
  • (optional) from the VSCode window running inside the container context you can install other extensions which might be useful, such as those for C++ or PDDL (which will not be installed in the VSCode running in your host system)

To check that everything is running simply try to execute the planner with the external attachments unmodified from within the container (so start it in whichever terminal you prefer), with:

    ~/ai4ro2/visits_domain# popf3-clp -x dom1.pddl prob1.pddl ../visits_module/build/libVisits.so region_poses

It should prompt you a log of the correct planning, with total cost 8 (due to the placeholder function present in the source file for the semantic attachment at the moment).

Also, as a QoL choice, the planner executable is already added to the PATH, meaning you can run popf3-clp from any directory without having to specify the full path to it each time.

⁠Thanks

Dr. Antony Thomas, for the domain, problem and semantic attachment scripts
Francesco Ganci, for a list of the needed dependencies