Public and Developer Version¶
pybrid, in the form of the packages pybrid-computing and pybrid-computing-native,
is available in two versions: the public/stable and the internal/private version.
The latter is reserved for anabrid employees and project partners with an API.
Please note that all up-to-date versions of anabrid products, including
pybrid, are always compatible. Reverting even only one of the packages to an older
version can lead to major incompatibilities.
Protobuf versioning
While the protobuf protocol has an internal version number and versioning infrastructure, in the current (development) phase its version is still pinned to 0.1.0. During this phase, only version bundles of all packages that use the underlying protocol version are compatible. In doubt, please always use the most recent build across the full stack.
Recognizing build versions from the version number¶
pybrid-computing uses semantic versioning with the following conventions:
- Released versions:
X.Y.Z(e.g.,0.11.5) - these versions are generally tagged and are the only versions that are released to the public on PyPi - Development versions:
X.Y.Z.devN+g<commit>(e.g.,0.11.6.dev3+g5c75b35) - development versions used in internal development, created from a single code commit and CI/CD run.
Installing the public version¶
Assuming you have set up uv as described in the getting started guide,
both packages can now be installed simply through PyPi, i.e.,
That's it! Updates can generally be retrieved with uv pip install -U.
Installing internal versions¶
Internal versions, the "bleeding edge", is hosted on anabrid's Gitlab server.
You will need an anabrid Gitlab account and set up an API
token:
1. Go to User settings (click on your portrait) and move to Personal Access Tokens
2. Click Add new token
3. Enter a self-chosen name and select the read_repository, read_virtual_registry, read_registry, read_api.
permissions.
4. Note down the generated token.
In order to automatically retrieve your images, this information should be stored permanently so you
avoid having to export ... this in each terminal window. In your ~/.bashrc, add the following two lines:
export CI_REGISTRY_USER=<your gitlab user name, e.g. thuerck>
export CI_REGISTRY_PASSWORD=<the token you generated above>
With this in place, you can now directly install from the Gitlab PyPi infrastructure: