chroma
chroma copied to clipboard
[Feature Request]: Add a `conda` install option for `chromadb`
A conda installation option from conda-forge channel could be very helpful for all downstream packages that depend on chromadb. I have already started working on this, to add chromadb to conda-forge.
Conda-forge PR:
- https://github.com/conda-forge/staged-recipes/pull/22537
The conda-forge PR has been merged, and other dependencies (such as duckdb) have also been added to conda-forge in the same PR. You will be able to install the library with conda as follows:
conda install -c conda-forge chromadb
:bulb: I will push a PR to update the docs once the package is available on conda-forge.
Good news! :tada:
So, chromadb is now available on conda-forge.
I have added @chroma-core as co-maintainer in the feedstock repo:
- https://github.com/conda-forge/chromadb-feedstock.
Check availability of chromadb on conda-forge channel.
- URL: https://anaconda.org/conda-forge/chromadb
- Command:
conda search -c conda-forge chromadb
cc: @jeffchuber @levand
@sugatoray cool! what is the process to distribute new releases when they are cut?
- New release on PyPI
- Update the recipe in the feedstock
- there is a conda-forge bot, which often picks up new versions on PyPI and creates an automated PR to update the recipe in the feedstock.
- so, unless the dependencies have changed you may not have to change much.
- maintaining the recipe is easy; I can guide you and show you what I am doing when PyPI updates happen.
P.S. I am more active on LinkedIn. Perhaps it will be nice to stay in touch over there as well.
- https://linkedin.com/in/sugatoray
cc: @jeffchuber
I hope that the client/server mode can minimize dependencies, avoid the need to install both Nvidia and Torch. Additionally, I request that a CPU installation mode be provided, which does not require the installation of Nvidia-related packages.
Thank you @sugatoray - looks like this is all working! Closing this out.