Removing __pycache__ from git repository

Question

How can I remove all __pycache__ subdirectories in my repository using .gitignore?

Answer 1

You cannot remove files from existing commits: those commits are frozen for all time. You can make sure you do not add new files to future commits, though. Simply remove the files now, with git rm -r --cached __pycache__, and list __pycache__ or __pycache__/ in your .gitignore (creating this .gitignore file if needed). Do this for each __pycache__ directory; use your OS's facilities to find these (e.g., find . -name __pycache__ -type d). Then git add .gitignore and git commit to commit the removal.

Note that any time anyone moves from any commit that has the files—where they'll be checked out—to a commit that lacks the files, they will have their entire __pycache__ directory removed if Git is able to do that; at the least, any cached files that were committed and can be removed will be removed. So the --cached in the git rm -r --cached above only speeds things up for you by avoiding the removal of the cached compiled files this time. Others will have to rebuild their cache.

To make a new and different repository in which the __pycache__ files were ever accidentally committed in the first place, use git filter-branch (which is now deprecated) or the newfangled git filter-repo (which is not yet distributed with Git). Or, see any of these various existing questions and their answers, which you should already have found before you asked this:

• Remove a file from a Git repository without deleting it from the local filesystem
• Applying .gitignore to committed files
• How can I make Git "forget" about a file that was tracked, but is now in .gitignore?
• Remove sensitive files and their commits from Git history

Answer 2

Add this in .gitignore, also in future you can use this handy generator

# Byte-compiled / optimized / DLL files
__pycache__/
*.py[cod]
*$py.class