You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
With #279 we have the first implementation of python bindings for rattler. 🎉
For v0.1.0 of the bindings we would like to be able to achieve the following goal:
🥅 You can solve and install an environment using the Python bindings.
We want to be able to do this to make it easier for users coming from Python to start using rattler as their base library for everything conda instead of having to rely on conda or mamba.
To that end, the following code needs to be worked on:
You can fetch repodata from a remote or local channel.
The main entry point is: fetch_repo_data.
We have to determine how we will manage memory. Since RepoData takes up a significant amount of memory we want to reduce copying, duplication, etc. The solver also takes PackageRecords by reference how do we deal with that?
I am not sure I understand why we need RepodataRecord and friends. We could also decide to hide all that behind the API, right? And just have a "Repodata" object that holds the deserialized repodata?
With #279 we have the first implementation of python bindings for rattler. 🎉
For v0.1.0 of the bindings we would like to be able to achieve the following goal:
🥅 You can solve and install an environment using the Python bindings.
We want to be able to do this to make it easier for users coming from Python to start using rattler as their base library for everything conda instead of having to rely on
conda
ormamba
.To that end, the following code needs to be worked on:
fetch_repo_data
.PackageRecord
s by reference how do we deal with that?async
AuthenticatedClient
Platform
- Start wrapping everything necessary for shell activation for py-rattler #298Channel
- add channel types to py-rattler #313RepodataRecord
/PackageRecord
(we can skip implementing all the fields for now?? maybe it would be nice to be able to access some)fetch_repo_data
returns an object that uses a scope to drop a lock. How do we deal with that?Gateway
implementation unless this lands pretty soon (would simplify the whole API a lot). Might be worth putting some effort in?RepodataRecord
orPackageRecord
to a solver function to solve the environment.RepodataRecord
s.The text was updated successfully, but these errors were encountered: