h5grove, core utilities to serve HDF5 file contents¶
h5grove is a Python package that provides utilities to design backends serving HDF5 file content: attributes, metadata and data. HDF5 files are accessed with h5py.
Documentation¶
The documentation of the latest release is available here.
Rationale¶
There are several packages out there that can serve HDF5 files. However, they are dedicated to their usecases and settle on one implementation, hampering reusability.
In addition, some problems arise constantly when designing HDF5 backends. To name a few:
Resolving external links
Dealing with compression and slicing of datasets
Encoding data efficiently and consistently (looking at you,
NaN,Infinityin JSON)
h5grove aims at providing building blocks that solve these common problems and can be reused in existing or new backends.
Installation¶
pip install h5grove
You can use h5grove low-level utilities whatever the backend implementation you choose. Additional utilities are provided for several Python web frameworks but require additional packages that can be installed the following way:
pip install h5grove[fastapi]
pip install h5grove[flask]
pip install h5grove[tornado]
Using h5grove¶
h5grove can be used in several ways.
To simply serve HDF5 file contents, you can look at the example implementations we provide and their API.
You might be interested to add HDF5 serving capabilities into an existing application
If the API does not suit you, you can make use of the low-level modules to design your own backend, letting h5grove handle low-level tasks such as link resolution and encoding.
Changelog¶
The release history is tracked on the Releases page.