Note: this only deals with metadata. If you want a download of actual git repositories, I recommend using python-github-backup.

Setting up

1. The easiest way is pip3 install --user git+https://github.com/karlicoss/ghexport.

   Alternatively, use git clone --recursive, or git pull && git submodules update --init. After that, you can use pip3 install --editable.

2. To use the API, you need to get a personal access token from settings. Note that you need to use repo scope.

Exporting

Usage:

Recommended: create secrets.py keeping your api parameters, e.g.:

token = "TOKEN"

After that, use:

python3 -m ghexport.export --secrets /path/to/secrets.py

That way you type less and have control over where you keep your plaintext secrets.

Alternatively, you can pass parameters directly, e.g.

python3 -m ghexport.export --token <token>

However, this is verbose and prone to leaking your keys/tokens/passwords in shell history.

You can also import ghexport.export as a module and call get_json function directly to get raw JSON.

I highly recommend checking exported files at least once just to make sure they contain everything you expect from your export. If not, please feel free to ask or raise an issue!

API limitations

WARNING: github API limits extent to which you can retrieve certain data, e.g. events you can only get events from the past 90 days, and not more than 300 events.

I highly recommend to export regularly and keep old exports. Easy way to achieve it is command like this:

python3 -m ghexport.export --secrets /path/to/secrets.py >"export-$(date -I).json"

Or, you can use arctee that automates this.

To get your older data past 90 days, you can request a manual export in your account settings.

Known Issues

The requests (and therefore PyGithub) modules on which this depends seems to sometimes fail to login if a ~.netrc file is present, see here for context.

Using the data

You can use ghexport.dal (stands for “Data Access/Abstraction Layer”) to access your exported data, even offline. I elaborate on motivation behind it here.

• main usecase is to be imported as python module to allow for programmatic access to your data.

  You can find some inspiration in =my.= package that I’m using as an API to all my personal data.

• to test it against your export, simply run: python3 -m ghexport.dal --source /path/to/export
• you can also try it interactively: python3 -m ghexport.dal --source /path/to/export --interactive

Example output:

Your events:
Counter({'PushEvent': 181,
         'WatchEvent': 27,
         'CreateEvent': 22,
         'IssueCommentEvent': 20,
         'PullRequestEvent': 15,
         'IssuesEvent': 5,
         'DeleteEvent': 5,
         'ForkEvent': 3,
         'PullRequestReviewCommentEvent': 1})