**********************************
Instruction For Teaching Assistant
**********************************

Responsibility
==============
#. Assignment grading

   Best to finish grading within one week after the due date.

   + Labs
   + Projects

#. Student question answering

   + on Discord
   + by Email

Project Grading
===============
Overview
--------
Programming projects are given on Canvas as assignments. It require students to
an implement application following the instruction. The grading is mostly
functionality based, which means the 80-90% of points based on how the
application compile and run. The rest of the grades require some brief code
reviews.

Pre-requisites
--------------
+ A working development environment. Check the :doc:`/cpp/ccpp-env` document for
  more details. Preferably a local environment. Linux, Mac OS or Windows WSL
  should all work well.
+ Knowledge on C++, CLI environment, and make

Workflow for Canvas Assignment
------------------------------
#. Download from Canvas

   Canvas assignment allows the download of a single zip file for each
   assignment.

#. Unzip and organize as individual directories

   A Python 3 script is available to automate the work but it is up to the TA
   to decide whether he want to use it. Take some time to learn and set up.

#. Test run and grade

   The student submission is compiled, test-run and tested using ``make``
   commands like ``make``, ``make test-all``, ``make test-run``. You can check
   the project instruction for details.

#. Update grades on Canvas

   Open the SpeedGrader, use to provided rubric to grade and leave brief
   feedbacks.

Workflow for GitHub Assignment
------------------------------
+ You may grade based on the autograding results from the GitHub Classroom
  interface.
+ Be careful with the late submission. You may need to look into the repository
  if a submission is late as GitHub Classroom is still working on the feature
  and the behavior is not consistent.
+ You can read code on GitHub website or download the code and run locally.
+ The GitHub command line tool is currently the best way to batch download the
  student submissions. It takes some time to learn and set up.

Resources And Tools
-------------------
+ Submission prepare :download:`script</_static/tasks.py>`

  * Prepare student submissions in individual directories
  * Python 3 and PyInvoke

    - Python3 is mostly pre-installed on Linux
    - install PyInvoke with the command ``pip install invoke`` or ``pip3
      install invoke``

  * Usage

    - create a new directory
    - put the ``submissions.zip`` file downloaded from Canvas and ``tasks.py``
      in the directory
    - Change to the directory in the terminal window
    - run ``inv organize`` in the directory in terminal


Policies
--------
+ Late assignment

  Apply the penalty following this scheme: -10 < 1 day; -30 < 2 days

+ Resubmission

  Re-submissions are allowed only when the student has a good reason.

+ Appeal

  The TA can explain the grading details to the student. The appeal should be
  forwarded to the instructor.

Hints
-----
+ Only fix problems for students if it can be done in very short time and
  minimal effort.
+ The student will lose all functionality points if their code does not
  compile using the ``make`` commands described in the instruction.
+ For advanced course like DSA 1 and 2, only read source code files for
  neatness, and style rather than correctness. No need to take points off for
  style unless it hurts readability badly.