- include wheel into the python distribution for PyPI
- add Doc Pages content and use material theme
- document cicd pipeline, by visualizing the Github Actoins Workflow as a graph
- automatically create nav tree of API refs, from discovered docstrings in *.py
- update README
- default docker build stage includes vgg and demo cmd
- call reusable workflow to handle PyPI Publish Job
- run Docker Job from reusable workflow
CI Docker Behaviour
- Implement an ON/OFF switch that gurantees nothing gets docker build and push
- Implement an ON/OFF switch that (given above ON), guarantees docker build and push
- Implement a switch between Continuous Deployment and Continuous Delivery Modes
New Multi-Stage Dockerfile
docker build --target prod_install .
: Image with installed app; Python and CLIdocker build --target prod_ready .
: Image with installed app + baked in Pretrained VGG Image Modeldocker build --target --target prod_demo .
: Image with installed app + baked in Pretrained VGG Image Model + data to run quick NST demodocker build .
: Image with installed app + baked in Pretrained VGG Image Model + default entrypoint
One-click, containerized NST Demo run
docker-compose up
- prioritize env vars to find the Demo Content and Style Images
- document ci pipeline configuration file
- multi-stage Dockerfile, and demo-nst service in docker-compose
- define distince Docker Stages for image with or without baked-in image model vgg weights
- introduce the Ruff tool for Fast! Static Code Analysis
- apply all static checks
- on CI Docker Job, build for Stage prod_install (--target), since missing vgg weights
- accept 4 Policies to define Docker BUild/Publish decision making
- on/off switch, giving option of running docker Job, in case tests Job was skipped
- static checks with isort, black, pylint, pyrom, pyflakes, mccabe, dodgy, profile-validator
- deploy (built .tar.gz and/or .whl) 'v*' tags; dev: test.pypi.org, master: pypi.org
Revive CI Pipeline
- update badges refs and urls
- add CI Pipeline Status Badge in README
- show Demo Content Image + Style Image = Generated Image
- Upload Code Coverage Data to Codecov.io, resulted from Test Suite runs
- Prototype GUI Client
- over 90% code coverage
- heavily document what is going on in the code
- initialize same Stochastic Process on subsequent processes
- interactive GUI with Live Update of Gen Image
- nst algo - broadcast weighted costs
- add cli cmd to quickly demo algorithm on 300 x 225 Content & Style images
- running NST on indentical input (Content/Style) yields same Generated Image
- unit test the Layer bridging the backend code and the Demo CLI cmd
- devcontainer and docker-compose with tensorboard service
- install tree cli tool, inside devcontainer
- breakdown perform_nst method into smaller ones
- remove the 'utils ' local package and use the software-patterns' package from pypi
- migrate from setuptools to poetry build
- add Github Actions CI Pipeline
- on gui start up, (load and) select Demo Content/Style Images, & render UI accordingly
- test algorithm & conditionally mock production pretrained model
- document how to use the docker image hosted on docker hub
- run regression test on ci server