[LINUX] [For organizing] Python development environment
A memo of my Python development environment
Use
- Deep learning / development of image processing
- Remote sensing
policy
- Introduced as easily as possible
- Management should be at least possible
- Easy to write code
- Completion, format, etc.
OS: Ubuntu 18.04
- Installation of libraries etc. is easier than Windows
- Not so recently, but some of them don't work without Ubuntu.
Environmental management: Anaconda
- All you need for machine learning is installed + Because you can manage the minimum environment
- The main libraries are as follows
- Omit numpy or Anaconda
- You can't do it without geopandas
#For deep learning
pytorch
torchvision
tensorboard
#For image processing
opencv
#For remote sensing
gdal
qgis
geopandas
Editor: VSCode
- Enhancement of Extension
- Remote Development
- Because it is developed on a remote server (GPU server)
- Extension should be usable at remote destinations
- ms-python
- Official Microsoft Extension
- Complements to the extent that it is not frustrating
- Bracket Pair Colorizer
- Corresponding parenthesis coloring
- Easy to set Linter and Formatter
- Notebook file can be edited
Linter: Flake8
- Feeling not too strict and not too loose
- Used with
max-line-length = 120, max-complexity = 10
Formatter: yapf
- Because the feeling after Format came the best
- Used with
column_limit = 120 as in Flake8
in conclusion
- I would like to add albumentations and other conveniences in the future
- I would appreciate it if you could tell me if there are other good things.