humanfirst-module

Humanfirst module package

Create Virtual Environment to do any dev work and perform pytest work

• Remove any previously created virtual env rm -rf ./venv
• Create virtualenv & activate python3 -m venv venv
• if bash shell source venv/bin/activate In case of deactivating use "deactivate"
• Update PiP python -m pip install --upgrade pip
• install requirements pip install -r requirements.txt --no-cache

Using pytest to test everything is working fine

First you need to decide if you are going to test locally or against staging/prod. Start with staging and production.

Set your HF_USERNAME and HF_PASSWORD env variable to match the environment you wish to test against.

Running on production

This is the default.

pytest --cov ./humanfirst/ --cov-report html --cov-report term --cov-report html - produces a report in HTML page --cov-report term - prints the report in console --cov-report term:skip-covered - helps to see uncovered parts

Running on staging

Note: Staging access is available only for internal team members

Switch environment variable HF_ENVIRONMENT = "staging"

Reset HF_USERNAME and HF_PASSWORD to be the relevant staging values.

To check are running on staging set HF_LOG_CONSOLE_ENABLE = TRUE and HF_LOG_LEVEL = DEBUG

Start the test again but with the console output printed pytest --cov ./humanfirst/ --cov-report html --cov-report term -s

You should see it calling https://api-staging.humanfirst.ai:443 in the logs

Testing using Docker build for this module

Note: Doing this requires the ability to run docker commands. You may need to run this from outside your local academy workbench

Check docker is working with docker run hello-world or sudo docker run hello-world depending on whether you have a user or root docker setup.

• docker build . -t humanfirst-module:latest --no-cache To run the tests for this module we pass through the necessary env variables echo $HF_ENVIRONMENT $BASE_URL_TEST $HF_USERNAME $HF_PASSWORD

docker run \
-e "HF_ENVIRONMENT=$HF_ENVIRONMENT" \
-e "BASE_URL_TEST=$BASE_URL_TEST" \
-e "HF_USERNAME=$HF_USERNAME" \
-e "HF_PASSWORD=$HF_PASSWORD" \
--name humanfirst-module-0 \
humanfirst-module \
pytest -s --cov ./humanfirst/ --cov-report term

Build Package

Wipe your build dir first to be sure

rm -rf ./build ./dist

Then build

python setup.py sdist bdist_wheel

Test packages before uploading

Packages can be tested using TestPYPI before uploading to PYPI - https://packaging.python.org/en/latest/guides/using-testpypi/

Register in TestPYPI - https://test.pypi.org/account/register/

Enable 2 factor authentication and generate API token using Account settings

Set TestPYPI password using keyring

keyring set system __token__

It'll prompt for a password - give it the API key Theoretically this stops you having to put in your username and password each time.

Then Enter your TestPYPI API token

Publish the package to TestPYPI

twine upload --repository testpypi dist/*

Username: __token__

Password: API token

Using TestPyPI with pip

pip install --index-url https://test.pypi.org/simple/ --extra-index-url https://pypi.org/simple/ <your-package>

Upload packages to PYPI

Register in PYPI - https://pypi.org/account/register/

Enable 2 factor authentication and generate API token using Account settings

Set PYPI password using keyring

keyring set system __token__

Then Enter your PYPI API token

Publish the package to PYPI

twine upload dist/*

Username: __token__

Password: API token

To install humanfirst package locally into academy

python -m pip install -e ../humanfirst-module/

To install humanfirst package locally into humanfirst-module

python3 -m pip install -e .

To install humanfirst package locally into humanfirst-module using the dist

pip install dist/humanfirst-<version number>.tar.gz --no-cache

or from academy example change the version number and the path...

pip install ../humanfirst-module/dist/humanfirst-1.1.3.tar.gz --no-cache

CHECKS

make sure any last minute build changes committed! pytest in academy

Log handling

• HF SDK logging offers multiple options. Either can save the logs, print them in the console, do both or none
• To store the logs in a specific directory set HF_LOG_FILE_ENABLE to 'TRUE' and set the directory in HF_LOG_DIR where the log files needs to be stored.
• Log file management
  • Rotating File Handler is used
  • When the log file size exceeds 100MB (Hard coded). Automatically a new file is created and old one is saved
  • Can go to upto 4 additional log files
  • If the number of log files exceed the additional log file count + 1, then automatically the oldest log file is gets replaced with new log information
• Can set log levels using HF_LOG_LEVEL. Accepts - 'DEBUG', 'INFO', 'WARNING', 'ERROR', 'CRITICAL'
• To print the logs in console set exp to 'TRUE'
• Default - the logs are neither saved nor printed onto console

Note: Control what files can go into SDK using Manifest.in file