Canvas Data CLI
A small CLI tool for syncing data from the Canvas Data API.
NOTE: this is currently in beta, please report any bugs or issues you find!
This tool should work on Linux, OSX, and Windows. The tool uses node.js runtime, which you will need to install before being able to use it.
- Install Node.js - Any version newer than 0.12.0 should work, best bet is to follow the instructions here
Install via npm
npm install -g canvas-data-cli
OR Install from github
git clone https://github.com/instructure/canvas-data-cli.git && cd canvas-data-cli && make installLocal
To generate a stub of this configuration run
canvasDataCli sampleConfig which will create a
config.js.sample file. Rename this to a file, like
Edit the file to point to where you want to save the files as well as the file used to track the state of which data exports you have already downloaded. By default the sample config file
tries to pull your API key and secret from environment variables,
CD_API_SECRET, which is more secure, however, you can also hard code the credentials in the config file.
Configuring an HTTP Proxy
canvas-data-cli has support for HTTP Proxies, both with and without basic authentication. To do this there
are three extra options you can add to your config file.
|proxyUsername||the basic auth username for the https proxy.|
|proxyPassword||the basic auth password for the https proxy.|
If you want to simply download all the data from Canva Data, the
sync command can be used to keep an up-to-date copy locally.
canvasDataCli sync -c path/to/config.js
This will start the sync process. The sync process uses the
sync api endpoint to get a list of all the files. If the file does
not exist, it will download it. Otherwise, it will skip the file. After downloading all files, it will delete any unexpected files
in the directory to remove old data.
On subsequent executions, it will only download the files it doesn't have.
This process is also resumeable, if for whatever reason you have issues, it should restart and download only the files
that previously failed. One of the ways to make this more safe is that it downloads the file to a temporary name and
renames it once the process is finished. This may leave around
gz.tmp files, but they should get deleted automatically once
you have a successful run.
If you run this daily, you should keep all of your data from Canvas Data up to date.
Fetches most up to date data for a single table from the API. This ignores any previously downloaded files and will redownload all the files associated with that table.
canvasDataCli fetch -c path/to/config.js -t user_dim
This will start the fetch process and download what is needed to get the most recent data for that table (in this case, the
On subsequent executions, this will redownload all the data for that table, ignoring any previous days data.
NOTE: This only works after properly running a
This command will unpack the gzipped files, concat any partitioned files, and add a header to the output file
canvasDataCli unpack -c path/to/config.js -f user_dim,account_dim
This command will unpack the user_dim and account_dim tables to a directory. Currently, you explictly have to give the files you want to unpack as this has the potential for creating very large files.
This subcommand is designed to allow users to make API calls directly. The main use case for which is debugging and development.
canvasDataCli api -c config.js -r /account/self/dump
Periodically requests data is regrouped into collections that span more than
just a single day. In this case, the date that the files were generated differs
from the time that the included requests were made. To make it easier to
identify which files contain the requests made during a particular time range,
we have the
canvasDataCli historical-requests -c config.js
Its output takes the form:
- Write some code
- Write tests
- Open a pull request
If you use docker, you can run tests inside a docker container
npm install .npm test