tztz8 / canvasautograderforjava

EWU Java Auto Grading

Use to auto grade students Java assignments.

This start off as me learning about tools and libs to help with grading students assignments. Now it a tool that need to be rebuilt but still works for auto grading students assignments from both Canvas LMS and GitHub and pushing those grades to Canvas LMS.

Getting started using this tool

Requirements to use this tool

This tool was and used in Linux

• Students Assignments need to be in a Jet Brains basic Java project (no gradle or maven) folder format

  • *.java in src/ folder
  • JUnit testing code in tests/ folder

• For normal labs (Not checking the students JUnit test)

  Write some JUnit tests and jar them without the sol (TODO: Link to how to do so in Intellij)

  Recommended to add a time outs to the JUnit if the students code has inf loop

• Make Both secret.py and currentLab.py

  • secret.py Has the api keys for both Canvas and GitHub
  • currentLab.py Has the needed info about the current Lab
    • Starting repo that the students fork from
    • The working dir for grading the lab
    • A csv file about the students in the class (TODO: Link to how to make the CSV file)
    • The JUnit Jar
    • Class name of the JUnit tests
    • Number of class names
    • Canvas HW ID (TODO: Link to how to get this)
    • Total number of JUnit tests
    • Path to java and javac version use for running these JUnit tests

• Update grade weights grade_weight.py

Basically ready to run the tool

Warring: you are about to run the students code on your computer

Recommended: use your IDE debugger when use the script

You can now run testRunner.py to grade the students assignments

As the script is running for each step it will ask to : Countue? you can:

• Use r to run/rerun the current step (step_get_labs is required for further steps)
• Use no or n to end the script
• Use anything else to skip the step

Steps the scrip go throw

• step_get_labs

  NOTE: Make a internal ref of the students for latter (This is why it can not be skipted)

  • Download the students lab
    • First look at forks of the starting repository for students code
      • if already clone in working dir it will refresh the clone
    • If did not find a student fork it will look for a zip on Canvas LMS on the HW ID

• step_setup_labs

  • Clean up students lab
    • Remove any out folder (Normally made by Jet Brains)
    • Remove any .idea folder (Make it easyer if you want to manaly grade to the students assgnment)
    • Remove any docs folder
    • Remove any tests folder
    • Check if src there and if not see if Canvas has a zip of students src code

• step_setup_tests

  • Make copy of tools in working dir
    • junit-platform-console-standalone jar file
    • checkstyle file
    • Junit jar file

• step_run_tests

  • Compile students code (with the test JUnit jar for if there was some lib you give the students to use can be include in that JUnit jar file)
  • Run the JUnit tests
  • Run CheckStyle on studetns code

• step_read_tests

  • Read the output files from above
  • Calculate the result of the student

  Also Makes a CSV with those result in the working dir

• step_upload_grades

  • Push those grades to Canvas LMS