pangeamt-nlp

PANGEAMT NLP

How to update and upload to PyPi

1. In the terminal go to the pangeamt_nlp directory.

2. Switch to the branch that made the PR:

git checkout <source_branch>
git pull

1. Activate the virtual env:

source venv/bin/activate

1. Remove the previous packages created:

rm -r dist

1. Create the package source distribution:

python setup.py sdist

1. Install the new distribution:

pip install dist/*

1. Test the changes:

cd tests

python -m unittest

1. You can now test it inside the docker:

sudo docker run -it -p 8081:8081 -v <absolute path to dist folder>:/nlp_dist enginestore.pangeamt.com/corporate/<engine_image> /bin/bash

1. Now, inside the docker install the new distribution:

pip install /nlp_dist/*

1. Start the server:

python start_server.py

1. Test the model with the changed package from another terminal:

curl -X POST -d '{"srcs": ["some testing segment"]}' 0.0.0.0:8081/translate

1. If it passes the tests accept the PR in github webpage.

2. Change to the master branch:

git checkout master
git pull

1. Increment version number:

vim pangeamt_nlp/__init__.py
git add pangeamt_nlp/__init__.py
git commit -m "Update version number"
git push

1. Delete older distributions:

rm -r dist/*

1. Create source and binary distributions:

python setup.py sdist bdist_wheel

1. Upload distributions to PyPi:

twine upload dist/*