Skip to content

Latest commit

 

History

History
83 lines (61 loc) · 2.01 KB

install.md

File metadata and controls

83 lines (61 loc) · 2.01 KB

SLING Installation and Building

Operating system: Linux
Languages: C++ (gcc or clang), Python 3.6+, assembler
CPU: Intel x64 or compatible
Build system: Bazel

Python

If you just want to use SLING from Python, e.g. want to try out the parser on a pre-trained model, you can install the SLING wheel with pip. On a Linux machine with Python 3 you can install a pre-built wheel:

sudo -H pip3 install https://ringgaard.com/data/dist/sling-3.0.0-py3-none-linux_x86_64.whl

You can test the installation by trying to import the sling package:

python3 -c "import sling; sling.which()"

If you want to try out the SLING parser you can download a pre-trained model from Ringgaard Research:

sling fetch --dataset caspar

You can then use the parser in Python:

import sling

parser = sling.Parser("data/e/caspar/caspar.flow")

text = input("text: ")
doc = parser.parse(text)
print(doc.frame.data(pretty=True))
for m in doc.mentions:
  print("mention", doc.phrase(m.begin, m.end))

C++

If you want to use SLING for C++ development, you need to download the source code and build it.

First, clone the GitHub repository.

git clone https://github.com/google/sling.git
cd sling

Next, run the setup.sh script to set up the SLING development environment and build the code:

./setup.sh

This will perform the following steps:

  • Install missing package dependencies, notably GCC and Python 3.
  • Install Bazel which is used as the build system for SLING.
  • Build SLING from source.
  • Set up link to the SLING development enviroment for SLING Python 3 API.

Building

You can use the buildall.sh script to build all the source code:

tools/buildall.sh

If you haven't run the setup.sh script already, you then need to link the sling Python module directly to the Python source directory to use it in "developer mode":

sudo ln -s $(realpath python) /usr/lib/python3/dist-packages/sling