canvasgrading

This repo contains several scripts that use the Canvas API to manage quizzes and assignments on Canvas.

All scripts will receive an argument that is based on an access token associated to your account (https://canvas.ubc.ca/profile/settings, see Approved Integrations). This token can be provided directly as a string in the command-line using the -t TOKEN option, or saved into a text file and provided with the -f TOKENFILE option.

All scripts will also include an optional argument: -c COURSE, corresponding to the ID of the course on Canvas. If you don't specify it, the script lists all courses you have access to and asks for a course. The ID can be obtained by looking at the URL on Canvas, it's the number after 'courses/'.

All scripts that involve quizzes will also include the -q QUIZ argument, corresponding to the ID of the quiz. If you don't specify it, the script lists all quizzes from the course you selected and asks for a quiz. The ID can be obtained by looking at the URL on Canvas when you open the quiz, it's the number after 'quizzes/'.

For assignment-based quizzes, the -a ASSIGNMENT argument is also provided and works in the same way as the quiz argument above.

Please submit any questions or comments to [email protected].

quiz2pdf.py

This is a script that converts a Canvas quiz to a PDF suitable for Gradescope. In addition to the -t/-f, -c and -q arguments listed above, it also receives the following arguments, all optional:

1. -l CLASSLIST: The name of a classlist CSV file containing at least the columns SNUM and ACCT. You can generate it on the department computers with a command like: classlist <course_num> -T -f "%SN,%ACCT". If provided, the PDF will use the student account as identifier in the first page. If not provided, the PDF will use the student's name and student number for identification. Note that, due to FIPPA, the name/student number option is only suitable for use in gradescope.ca, though its use gives significantly better recognition by Canvas automatic matching algorithm.

2. -p PREFIX: A string used as prefix for the output files. Can also be used to specify a different directory to store the data in (e.g., ~/cs299/Final/FinalExam). If not specified, a prefix is generated based on the quiz name, removing any non-alphanumeric characters (spaces, etc.).

3. --only-question QID QID ...: A list of question IDs to include. If not specified, all questions are included. The ID of the question can be obtained by running this script with all questions (without this argument), the number provided before the text of each question is the question ID. Alternatively, you may obtain the question ID by using the JSON sync script below.

4. --not-question QID QID ...: A list of question IDs to exclude. If not specified, all questions are included. The ID of the question can be obtained by running this script with all questions (without this argument), the number provided before the text of each question is the question ID. Alternatively, you may obtain the question ID by using the JSON sync script below.

5. -d: Enter debug mode. In this mode, only the first 10 submissions are loaded from Canvas, and an extra debug.json file is generated. This can be used to test the configuration, style or the script itself.

6. --css: An additional CSS file to be used for this quiz. The existing canvasquiz.css file in this repo is always added, this CSS file can be used for question-specific or exam-specific formatting.

7. --template-only: If provided, only the template is created, with no student submission.

The script requires the use of the weasyprint Python library. As system package installation may be required, see https://weasyprint.readthedocs.io/en/stable/install.html before installing with something like:

pip3 install --user weasyprint

The script will connect to the Canvas API to get the latest responses from Canvas itself. The program will generate a file XXX_template.html and XXX_template.html.pdf, where XXX is the string from the first argument; the PDF file can be used as a template for a quiz on Gradescope. The script will also create a series of files named XXX_exams_YY.html and XXX_exams_YY.html.pdf, where YY is a counter; each file contains up to 20 exams, which can be uploaded to Gradescope as exams.

The script also creates a file named XXX_raw_answers.zip containing a file for each essay question in the quiz. For quizzes with multiple attempts allowed, these files will be created for all attempts. This file is named answer_<QID>_<STUD>_v<ATT>.html, where <QID>, <STUD> and <ATT> are the question ID, student number (or account) and attempt number, respectively. These can be useful in cases where either the answer is too long to fit in a PDF page, or if the marker wants to copy that answer to test it (e.g., in an IDE). Also, for all file upload type questions, the uploads will be saved in the same file, in a file named answer_<QID>_<STUD>_v<ATT>_<FN>, where FN is the file name originally used by the submitter.

The script supports all question types allowed in Canvas classic quizzes, including text-only, essay, file upload, fill-in the blank (or multiple blanks), multiple choice (including true/false and multiple answers), multiple dropdowns, matching, numerical answers and formula (calculated) questions. It also supports question groups, in which case each question version is saved in a different page, and any question submission that doesn't include a specific question will list the answer as "NO SUBMISSION", with an explanation text. For file upload questions, the answer will list the file names included in the answer, which can be found in the raw answers file listed above. For all other questions, the answer will be listed in the PDF itself. If the question plus answer don't fit a single page, it will be truncated, and you may find the answer in the raw answers file above.

Given some limitations of the Canvas API, this script does not support question groups linked to question banks. If using question groups, the questions must be included in the quiz itself instead of in the bank only.

quiz2txt.py

This script generates a ZIP file containing HTML files for each essay-type answer, as well as file uploads. It receives arguments -t/-f, -p, -c, -q, -d, --only-question and --not-question, with the same format and meaning as the equivalent arguments in quiz2pdf.py above.

The script will connect to the Canvas API to get the latest responses from Canvas itself. The script also creates one ZIP file per question, named <XXX>_<QNAME>_<QID>.zip, containing a file for each essay question in the quiz, where <XXX>, <QNAME> and <QID> are the prefix, question name and question ID, respectively. This file is named <QID>_<STUD>_v<ATT>.html, where <QID>, <STUD> and <ATT> are the question ID, student number and attempt number, respectively. Each answer will also include the question itself at the top of the file. For quizzes with multiple attempts allowed, these files will be created for all attempts. Also, for all file upload type questions, the uploads will be saved in the same file, in a file named <QID>_<STUD>_v<ATT>_<FN>, where FN is the file name originally used by the submitter.

In addition to the answers, the ZIP files can also create a rubric file for every student. In that case, a template must be provided for each question, with the name <XXX>_rubtempl_q<QID>.txt. For question groups, this rubric may be named <XXX>_rubtempl_qg<QGID>.txt, where <QGID> is the question group ID. This template will be copied as is to the ZIP file, once for every submission, with the name <QID>_<STUD>_v<ATT>_rubric.txt.

Unlike quiz2pdf.py, this script supports question groups linked to question banks.

json2quiz.py

This script can be used to read, change and push changes to Canvas quizzes using JSON files. This can be suitable for cases where changing a quiz on a text editor is easier, or to create quizzes/questions programatically.

The script receives arguments -f/-t, -c and -q as described above. It also receives a required argument (positional, i.e., with no prefix required) corresponding to the JSON file to use for synchronization. The script can work in one of three ways:

1. If the -l option is provided, the quiz is loaded from Canvas and saved into the JSON file.

2. If the -p option is provided, the JSON file is read and pushed to Canvas as a new or updated quiz.

3. If both the -p and the -l options are provided, then both operations are performed: the Canvas quiz is updated based on the JSON file, and the new values from Canvas (including IDs) are loaded back to the JSON file.

In addition to the options above, the -s option (used when -l is provided) controls the format of the output JSON file. If not provided, all fields received by the Canvas API are saved into the JSON file. If the option is given, only the fields that can be sent back to Canvas for updates are kept, as well as the IDs for information only.

The -d option is also provided to enter a debug mode, though at this point it provides no functionality.

The script can update existing questions, which can be done by simply changing the values associated to the question without changing its key. It can also create new questions, by creating a new item in the JSON file with a key not currently associated to an existing question or group (e.g., a non-numeric key). The script currently has no support for deleting questions.

Question groups are also supported, but only if there is at least one question in the group. Due to limitations of the Canvas API, question groups with no question directly associated to the group cannot be loaded from Canvas. This includes groups linked to question banks, even if the question bank has questions. To add a question to an existing group, use the group ID itself in the 'quiz_group_id' field. It is also possible to create new groups, in which case a non-numeric string should be used as a key, and questions to be added to that group should use the string key as group ID. You are strongly encouraged to maintain at least one question associated to each group, as groups with no question can't be loaded back to the file.

It is also possible to reorder items in the quiz by changing the 'order' item. Note that the name used in this item is for information purposes only, and will be ignored during updates.

dupquiz.py

This script duplicates an existing quiz on Canvas onto a new quiz with the same properties and the same questions. The script receives arguments -f/-t, -c and -q as described above, where the -q ID corresponds to the source quiz. The title of the new quiz is the same as the original quiz, adding the suffix " (copy)".

By default, the new quiz is created as unpublished. If the --published option is provided, the quiz is instead created as published.

If the --practice option is provided, the quiz is changed to be a practice quiz. This means that the title receives the suffix " (Practice Version)" instead of copy. Some changes are also done in the quiz properties, such as removing the due date and lock date, setting the release date to the lock date of the original quiz, removing time limits and attempt limits, and allowing correct answers to be shown on submission.

pushquizgrade.py

This script pushes quiz grades based on a CSV file. Documentation pending.

pushasggrade.py

This script pushes assignment grades from a CSV file to a Canvas rubric associated to an assignment. Documentation pending.

License

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.