Magda is a federated data catalog system developed by Data61's Engineering & Design Group.
Magda has an extensive REST API and a rich web-based user interface. However, to support simple data operation from the command line (and to learn golang), I wrote this simple command-line tool. It only covers the subset of the Magda API I have been using. However, I'm mor than happy to receive pull requests to extend it's functionality or fix bugs.
Right now, we don't provide pre-compiled executable, but if you have go installed, you can easily build & install it with:
go install github.com/maxott/magda-cli@latest
Simply type magda-cli -h for a listing of the supported commands, or magda-cli --help-man | nroff -man | more for a more detailed description.
Some of the common information, like --host will be read from environment variables when they are set. Check for square bracketed names in the help text, like MAGDA_HOST for the --host flag:
% magda-cli -h
usage: magda [<flags>] <command> [<args> ...]
Managing records & schemas in Magda.
Flags:
-H, --host=HOST DNS name/IP of Magda host [MAGDA_HOST]
...
The most commonly set environment variables are:
MAGDA_HOST=magda.example.com
MAGDA_AUTH_ID=6a7...
MAGDA_AUTH_KEY=GL...=
MAGDA_TENANT_ID=ab...
Like many cli tools we are following the command sub-command pattern. For instance, the following will show (read) the content of a record:
magda-cli record read -i recordID
The sub commands are primarily following the CRUD verbs with some additional functionality if the API supports it.
This tool tries to stick as much as possible to the Magda API and often simply prints what is being returned by that API.
I'm using Kingpin, so you can setup shell completion with:
# If you're using Bash
eval "$(magda-cli --completion-script-bash)"
# If you're using Zsh
eval "$(magda-cli --completion-script-zsh)"
Don't forget to check out the completion for flags by pressing <TAB> after typing -- as in:
% magda-cli --<TAB>
--authID --authKey --help --host ...
This is my first golang program so I'm still trying to work out on how to best distribute it. At this point, you'll just need to follow the usual pattern:
go mod tidy
go build .
go install
We are using GoReleaser.
To quote from the manual: GoReleaser will use the latest Git tag of your repository. Create a tag and push it to GitHub:
git tag -a v0.1.0 -m "First release"
git push origin v0.1.0
To deploy releases, we also need a GITHUB_TOKEN environment variable, which should contain a valid GitHub token with the repo scope. You can create the token here.
export GITHUB_TOKEN="YOUR_GH_TOKEN"
Now you can run GoReleaser at the root of your repository:
goreleaser release
If you don't want to create a tag yet, you can also run GoReleaser without publishing based on the latest commit by using the --snapshot flag:
goreleaser --snapshot --skip-publish --rm-dist