canvasfs / a FUSE filesystem for Canvas courses

canvasfs allows you to mount a Canvas course as if it were a local folder. Files are fetched from the network lazily when you open them and are locally cached.

Why?

• You no longer have to click through module pages to download and save individual files
• This is more convenient than simply exporting the course contents, as files will update automatically when their content is changed
• It's easy to tell when new files have been added to a course by looking at the 'Date Modified' field in the file explorer (suggestion: sort folders by most recently modified first)
• You can use tools like ripgrep and ripgrep-all on course content

Features

• Will find all files in the 'Files', 'Modules' and 'Assignment' tabs of each Canvas course
• Currently only supports read access. Write access is possible (for instructors) but I don't have the permissions to test it
• Works on Linux, or more generally wherever pyfuse3 is supported. On Windows you can try using WSL2.

Ideas, PRs and issues are welcome.

Setup

Make sure that the FUSE libraries are installed via your distro package manager. On Arch and Manjaro these are provided by the fuse3 package.

git clone https://github.com/blepabyte/canvasfs.git
cd canvasfs
pip install -r requirements.txt

You will need to obtain an API key for your Canvas account. To do this, go to the 'Settings' page of your Canvas profile, open the browser devtools, and add the following global CSS rule to the page:

.add_access_token_link {
    display: block !important;
}

You can then click on the "New Access Token" button that appears under "Approved integrations" to get your token. Now either see Configuration below, or just run the main program and follow the prompts to automatically generate the config.json file.

python canvasfs.py

Configuration

By default config parameters are stored in a config.json file in the root canvasfs folder. The first four parameters below are required.

{
    // API key
    "token": "XXXXXXXXXXXXXXXXXXXXXXX", 
    
    // Replace with your institution's domain
    "domain": "https://canvas.auckland.ac.nz",
    
    // An empty folder where files will be shown locally
    "mount_dir": "./remote",
    
    // Cache location on disk
    "cache_dir": "/mnt/storage/.canvasfs",
    
    // IDs of courses to mount
    "courses": [
        {
            "id": 11111,
            "name": "COMPSCI 000"
        },
        {
            "id": 22222
        }
    ]
}

Optional parameters

• refresh_interval: how often (in hours) it should check for new files on Canvas. The default is 4 hours. Set to -1 to disable refreshing completely.
• debug: sets logging level to DEBUG rather than INFO

Per-course configuration

You can pass a list of courses you want to mount by specifying their ID (this can be found in the URL of the course home page: https://canvas.auckland.ac.nz/courses/<ID>). By default this will be all currently "active" courses - though some courses do not set their end date properly, so you will likely end up needing to manually configure this.

• name: sets the display name of the local folder. Obviously this should be unique between courses and a valid directory name

Usage

python canvasfs.py [path-to-config-file=./config.json]

If you want, you can manually trigger a file list refresh by sending a SIGUSR1 signal to the running Python process, e.g. using btop or htop.