[Draft] Icechunk CLI Design Document #714
Conversation
|
|
||
| Here is a list of tasks a user might want to do with Icechunk: | ||
|
|
||
| - List my repositories |
There was a problem hiding this comment.
What should this list? Probably all repositories in the repositories config?
It would also be great to be able to point at a location and auto-discover repos.
There was a problem hiding this comment.
Auto discover would be cool... seems like an advanced feature we may not need for a while.
| - Create a new repository | ||
| - Check configuration | ||
| - Diff between two commits | ||
| - Invoke administrative tasks (garbage collection, compaction, etc) |
There was a problem hiding this comment.
This is not yet reflected in the current Interface section. Detailed API is out of scope, but it might be useful to think about this to see if the proposed command structure works.
|
|
||
| Here is a list of tasks a user might want to do with Icechunk: | ||
|
|
||
| - List my repositories |
There was a problem hiding this comment.
Listing repositories is something Icechunk cannot do today. It can only verify if a repository exists at a given location.
We don't expect to add this functionality, we tend to see anything that is outside the repo prefix as "unknown" to Icechunk.
There was a problem hiding this comment.
How about listing the repositories defined in the proposed repo config?
| - Create a new repository | ||
| - Check configuration | ||
| - Diff between two commits | ||
| - Invoke administrative tasks (garbage collection, compaction, etc) |
There was a problem hiding this comment.
I'd add:
- Print the zarr hierarchy
- Get repo statistics, example:
getsize - Fetch metadata for a node
- Update node user attributes?
- One day search metadata (IC cannot do this today)
- In the future we could include "export" functionality, like "export array foo to a zarr store"
|
|
||
| Here is a list of tasks a user might want to do with Icechunk: | ||
|
|
||
| - List my repositories |
There was a problem hiding this comment.
Auto discover would be cool... seems like an advanced feature we may not need for a while.
| icechunk snapshot diff <repo> <snapshot_id_1> <snapshot_id_2> | ||
| icechunk ref list <repo> | ||
|
|
||
| icechunk config init # init: interactive setup |
There was a problem hiding this comment.
nice! interactive could be very usefu
Thanks for the feedback! Here's the skeleton I used when I was exploring the Python integration: #716 |
Icechunk CLI Design Document
Design document for #461
Written as a design document, but this is just a proposal – feel free to disagree and offer alternatives.
TODOs
The configuration section is underdeveloped and the next part I will work on.