Conda and Git for Python Rocks!
Previously I’ve used Python for smaller projects which has not required any subversion control. I’ve basically used Git to keep track of progress using the main branch – and to share code and use multiple computers with ease. This has been working like a charm for years!
Being involved with some bigger projects and working with streamlit have forced me to learn some new tricks which is always fun! I’ll want to summarize some of the key points as a future cheat-sheet however basic it might be. It’s also a way of keeping track of progress .
- Use of Anaconda to keep several python environments
- Use of pipreq to create requirements.txt for creation of said environments
- Using git for development/test and master branches for deployment of python code
- Juypiter notebooks for data exploration and prototyping.
This has helped me raise my python development skills to a new level, which indicates how low it really was but at the same time acknowledges how it has become my primary language for so many things other than smaller scripts there and there.
With cloud-based streamlit apps recompiles on when changes are pushed to the main branch and the environment is rebuilt based on requirements.txt. It is useful to have a development branch that can be tested before updates are pushed to the public main branch.
I find myself using more and more Jupyter notebooks to experiment, explore and interact with data. Having the possibility to create sliders makes it super helpful to test data.
What’s next, learning to love debuggers ?