Summarise a directory of git repos. Regenerate them from the summary.
- Python 3 at
/usr/bin/python3 - Git 2.22 or above
- Ensure that
~/.local/bin/is in your$PATHenvironment variable jq(optional) so you can use some handy recipes shown below
curl --proto '=https' --tlsv1.2 -sSf https://raw.githubusercontent.com/mebble/sirup/main/scripts/install.sh | sh
curl --proto '=https' --tlsv1.2 -sSf https://raw.githubusercontent.com/mebble/sirup/main/scripts/remove.sh | sh
Print the usage instructions by running:
sirup help
Output:
Usage: sirup <command> <args>
Commands:
help Print these usage instructions
sum Summarise git repos and print the summary in JSON to stdout
--repos ./repos/dir The directory containing the git repos
--log [optional] Will output logs to stdout
gen Generate git repos from a summary file
--from ./sum/file Path to the summary file
--to ./dest/dir The destination directory where you want to clone the repos
For more information, visit: https://github.com/mebble/sirup/blob/main/README.md
Notes:
- Your git repositories must be kept in a flat directory structure. That is, if your personal projects lie within
./projects-personal/, they must be direct children of this directory and should not be nested within any sub directories. Future versions ofsirupmight allow nested directories.
Assuming you have this directory structure:
`-- projects
|-- birdwatch
`-- burl
Basic usage:
sirup sum --repos ./projectsWith logs:
sirup sum --repos ./projects --logPiped to jq (don't use any flags so that only JSON gets piped to jq):
sirup sum --repos ./projects | jqOutput when piped to jq:
[
{
"name": "birdwatch",
"is_clean": true,
"current_branch": {
"local_branch": "master",
"remote_branch": "origin/master",
"synced": true
},
"remotes": {
"origin": {
"fetch": "[email protected]:mebble/birdwatch.git",
"push": "[email protected]:mebble/birdwatch.git"
}
},
"size": {
"value": 11.76,
"unit": "KiB"
}
},
{
"name": "burl",
"is_clean": true,
"current_branch": {
"local_branch": "main",
"remote_branch": "origin/main",
"synced": false
},
"remotes": {
"origin": {
"fetch": "[email protected]:mebble/burl.git",
"push": "[email protected]:mebble/burl.git"
}
},
"size": {
"value": 10.90,
"unit": "KiB"
}
}
]Sort the repos by size:
sirup sum --repos . | jq '. | sort_by(.size.value) | group_by(.size.unit) | flatten | .[] | [.name, .size.value, .size.unit] | tostring' -r | tr -d '[]"' | column -t -s ','Get the names of dirty repos:
sirup sum --repos . | jq '.[] | select(.is_clean | not) | .name'Get the names of repos that aren't synced with their remotes:
sirup sum --repos . | jq '.[] | select(.current_branch.is_sync | not) | .name'Assuming you had executed:
sirup sum --repos ./projects > repos.json
Basic usage:
sirup gen --from repos.json --to output-dirOutput:
Cloning repositories to output-dir
[1/2] Cloning repo: birdwatch
[Done]
[2/2] Cloning repo: burl
[Done]
You can add the following fields to a repo object in the JSON summary that sirup gen will pick up:
| Field | Description |
|---|---|
"ignore": true |
The repo will not be cloned |
In your computer you might have a directory containing many git repositories. These could be personal projects, college assignments or codebases from work.
Now imagine you just got a new computer and need to move all these repositories to it. If you have hundreds of repos, you'd probably need to copy gigabytes of data. This could take time. Your old computer is probably slow, making this worse. You could first delete your node_modules, build files, etc, but this would be really tedious. It would also be tedious to figure out which repo is clean or dirty, which repo is synced to its remote, and so on.
Enter sirup. This tool will show you a JSON summary of all your git repositories. You could analyse the JSON using any standard tool to figure out the status of each repo. You could tweak this JSON to your liking. You could move this JSON to another computer and generate the git repositories over there.
This will hopefully save you time and data.
python3 -m unittest tests.test_shell
python3 -m unittest tests.test_git