datalad create-sibling-github


datalad create-sibling-github [-h] [--dataset DATASET] [-r] [-R LEVELS] [-s NAME] [--existing MODE] [--github-login NAME] [--github-organization NAME] [--access-protocol {https|ssh}] [--publish-depends SIBLINGNAME] [--dryrun] REPONAME


Create dataset sibling on GitHub.

An existing GitHub project, or a project created via the GitHub website can be configured as a sibling with the siblings command. Alternatively, this command can create a repository under a user’s GitHub account, or any organization a user is a member of (given appropriate permissions). This is particularly helpful for recursive sibling creation for subdatasets. In such a case, a dataset hierarchy is represented as a flat list of GitHub repositories.

GitHub cannot host dataset content (but LFS special remote could be used, However, in combination with other data sources (and siblings), publishing a dataset to GitHub can facilitate distribution and exchange, while still allowing any dataset consumer to obtain actual data content from alternative sources.

For GitHub authentication a personal access token is needed. Such a token can be generated by visiting or navigating via GitHub Web UI through: Settings -> Developer settings -> Personal access tokens. We will first consult Git configuration hub.oauthtoken for tokens possibly available there, and then from the system credential store.

If you provide –github-login NAME, we will consider only tokens associated with that GitHub login from hub.oauthtoken, and store/check the token in credential store as associated with that specific login name.



GitHub repository name. When operating recursively, a suffix will be appended to this name for each subdataset. Constraints: value must be a string

-h, --help, --help-np

show this help message. –help-np forcefully disables the use of a pager for displaying the help message

--dataset DATASET, -d DATASET

specify the dataset to create the publication target for. If no dataset is given, an attempt is made to identify the dataset based on the current working directory. Constraints: Value must be a Dataset or a valid identifier of a Dataset (e.g. a path)

-r, --recursive

if set, recurse into potential subdataset.

-R LEVELS, --recursion-limit LEVELS

limit recursion into subdataset to the given number of levels. Constraints: value must be convertible to type ‘int’

-s NAME, --name NAME

name to represent the GitHub repository in the local dataset installation. Constraints: value must be a string [Default: ‘github’]

--existing MODE

desired behavior when already existing or configured siblings are discovered. In this case, a dataset can be skipped (‘skip’), the sibling configuration be updated (‘reconfigure’), or process interrupts with error (‘error’). DANGER ZONE: If ‘replace’ is used, an existing github repository will be irreversibly removed, re-initialized, and the sibling (re-)configured (thus implies ‘reconfigure’). REPLACE could lead to data loss, so use with care. To minimize possibility of data loss, in interactive mode DataLad will ask for confirmation, but it would raise an exception in non-interactive mode. Constraints: value must be one of (‘skip’, ‘error’, ‘reconfigure’, ‘replace’) [Default: ‘error’]

--github-login NAME

GitHub user name or access token. Constraints: value must be a string

--github-organization NAME

If provided, the repository will be created under this GitHub organization. The respective GitHub user needs appropriate permissions. Constraints: value must be a string

--access-protocol {https|ssh}

Which access protocol/URL to configure for the sibling. Constraints: value must be one of (‘https’, ‘ssh’) [Default: ‘https’]

--publish-depends SIBLINGNAME

add a dependency such that the given existing sibling is always published prior to the new sibling. This equals setting a configuration item ‘remote.SIBLINGNAME.datalad-publish-depends’. This option can be given more than once to configure multiple dependencies. Constraints: value must be a string


If this flag is set, no communication with GitHub is performed, and no repositories will be created. Instead would-be repository names are reported for all relevant datasets.


datalad is developed by The DataLad Team and Contributors <>.