Exciting progress from the conda community! Sharded Repodata (CEP 16) is now in beta, delivering up to 10x faster environment solves and 35x reduction in network transfer for conda-forge users. This community-driven innovation demonstrates how open source collaboration continues to advance the Python ecosystem - exactly the kind of improvement that benefits millions of practitioners worldwide. #conda #Python
Anaconda.org and conda adopt sharded repodata! 🎉 🌟 More details and beta opt-in instructions: https://lnkd.in/grmE8B3v Sharded repodata, introduced by prefix.dev in CEP 16, splits large channel repodata files into per-project shards so clients fetch only what they need. This brings significant speedups through reduced bandwidth and memory use. As part of this beta, shards that have been available on prefix.dev channels are now also served on Anaconda.org for the conda-forge channel: 🔹 Conda users can opt in to the beta and see the improvement when using conda-forge. 🔹 Pixi and rattler-build support sharded repodata by default, so they benefit automatically. This work shows what collaboration across the community can achieve! Huge shoutout to the conda-forge community, prefix.dev, Quansight, QuantStack and Anaconda, Inc. for making this possible. ❤️