Taming the toxic python environment on your laptop
summary
A desktop+cloud UI called conosur (cond store in the transcript) actually builds reproducible artifacts: it creates a lock file with exact hashes, produces Docker images and tar.gz archives for Linux and macOS, and lets teams publish shared environments and roll back to prior builds. The system records install logs, shows the full dependency graph (cond versus pip sections), and supports recreation on another laptop via micr Mamba or unpacking the archive so broken installs fail loudly and repeatably.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 2:16Python packaging today
- 5:27Python environment problems
- 9:19Flexibility vs reproducibility
- 12:32Python environment best practices
- 13:52Introducing conda-store
- 14:13Live demo: conda-store
- 20:43Mission behind conda-store
- 21:26Sharing reproducible environments
- 23:57Reconstructing old environments
- 26:42UI forcing best practices
- 28:54Mixing conda and pip safely
- 33:13Dependency graph conflicts
- 35:20Separate environments per library
- 36:02Future vulnerability plans
- 38:10Environment identifiers and sharing
- 40:18Python packaging consulting
- 41:00Q&A
related talks
23mReproducible Environments for Reproducible Results
30mFrom Chaos to Confidence: Solving Python's Environment Reprodu...
23mCleaning Up Your Python Environment Superfund Site
40mMaking packages and packaging "just work"
28mBuilding reproducible Python applications for secured environments
09mShareable Working Environments
13mBuilding reproducible dev environments with Docker
26mCLI, CI, DIY: Crafting Python Based Dev Tools in a DevOps Vacuum