Releases: acwooding/conda-ops
Releases · acwooding/conda-ops
v0.4
What's Changed
Refactor to depend on prefixes broadly in the code base. In particular, this allowed for the addition of the prefix option to the config.ini
and allow for environments to be stored at the specified relative path instead of in the standard location for conda environments via conda init -p relative/path
. Here the location would be .conda-ops/envs/relative/path
and the environment would not have a name associated with it.
Other tweaks along the way:
- re-add pip output
- fix
conda ops status
- stream conda output so you know something is happening
- update config handling for new settings in conda 23.9.0
Full Changelog: v0.3...v0.4
v0.3
- Change the parsing packages on the command line for add/install as it relates to channels to make it clearer and less cumbersome. For example,
conda ops add -e .
adds a line-e .
in the pip entries of the requirements. Previously you had to put inconda ops add -c pip "-e ."
- Take local file urls out of the
lockfile.json
and replace them with a lookup to a local file catalog. The local file catalog is added to a.conda-ops/.gitignore
file and is not shared. When the local file catalog is missing a lookup, it is generated as part ofconda ops sync
(before attempting to regenerate the lockfile). - Add
conda ops env clean
to remove any temporary environments lying around conda ops status
now checks that the initial set of configuration files are present and if they are missing or inconsistent, promptsconda ops init
conda ops status
checks theconfig.ini
for missing keys and promptsconda ops init
instead of a hard fail- If you choose to re-initialize via
conda ops init
, it actually re-initializes (before it did nothing) - Clean up temporary files used to generate the lockfile instead of leaving them as artifacts
- Improve build scripts to make it easier to release
v0.2
- Update the CLI commands to hinge on a simplified set of commands (
add
/remove
,sync
,status
) - Clean up log levels and messaging for better flow and user experience (more self-evident and self-documenting)
- Update the README with basic usage instructions
- Enable sharing across platforms by tracking the platform the lock file entries were generated for