GraphANNIS is a library for corpus linguistic queries and these are the Python bindings to graphANNIS core library version 2.4.8.
GraphANNIS is available as Python3-compatible library from the central PyPI repository: https://pypi.org/project/graphannis/ You can install it locally with pip (https://pip.pypa.io/en/stable/).
pip install graphannis
On Ubuntu Linux systems the command might be pip3
for Python 3.
You can view the API documentation online.
Install you the python packages locally (use Python3) for your current user (remove the --user
to install it system-wide)
python3 setup.py install --user
This will automatically download the the graphANNIS binaries to the graphannis/<platform>
folder, where platform is one of the following:
Operating system | <platform> |
---|---|
Linux (64 bit) | linux-x86-64 |
MacOS X (64 bit) | darwin-x86-64 |
Windows (64 bit) | win32-x86-64 |
You can change the CORE_VERSION
field in the setup.py
to use a different released version of graphANNIS.
To compile graphANNIS on your own (e.g. for using a non-released version)
- Clone the graphANNIS library from https://github.com/korpling/graphANNIS/
- Follow the graphANNIS compile instructions
- Copy the resulting shared library file
<graphANNIS-repo>/target/release/libgraphannis_capi.so
(libgraphannis_capi.dylib
under MacOS X andgraphannis_capi.dll
under Windows) into thegraphannis/<platform>
folder (but without the_capi
suffix).
- Make sure your local working directory is a checkout of the
main
branch and up-to-date. - Update version information, by
- changing the
VERSION
field in thesetup.py
file - specifying the corresponding graphANNIS release tag in the
CORE_VERSION
field in thesetup.py
file - committing the changed files
- changing the
- Prepare the release (including executing tests) by executing
./release_prepare.sh
. - Check if there where no test errors and Commit the changes created by release preparation script.
- Tag and push the latest commit with the prefix
v
, e.g.v1.4.0
CI will automatically deploy all tagged commits on the main
branch.
- CFFI (http://cffi.readthedocs.org/) MIT License
- networkX (http://networkx.github.io/) BSD License
- Thomas Krause ([email protected])