transfer: add naive objects transfer function #64
Conversation
Codecov Report
@@ Coverage Diff @@
## main #64 +/- ##
==========================================
+ Coverage 50.70% 53.55% +2.84%
==========================================
Files 35 37 +2
Lines 2043 2065 +22
Branches 243 245 +2
==========================================
+ Hits 1036 1106 +70
+ Misses 987 930 -57
- Partials 20 29 +9
Continue to review full report at Codecov.
|
|
|
||
|
|
||
| def transfer( | ||
| src: "ObjectDB", dest: "ObjectDB", oids: Set["str"], jobs: int = None |
There was a problem hiding this comment.
There's a question of how oids will be discovered by the caller. I guess it's a responsibility of odb.
See iterative/dvc#7874.
| from .db import ObjectDB | ||
|
|
||
|
|
||
| def transfer( |
There was a problem hiding this comment.
I'm thinking about whether we even need this function. Based on how we calculate status on dvc-data side, the logic here will not be reusable there. Although, this will be useful for the naive transfer of objects.
skshetry
force-pushed
the
transfer
branch
2 times, most recently
from
f111f03
to
23a7cc4
Compare
skshetry
changed the title
skshetry
changed the title
| for oid in new: | ||
| path = src.oid_to_path(oid) | ||
| dest.add(path, src.fs, oid) |
There was a problem hiding this comment.
Will add batching support, as soon as ObjectDB.add gets batching support.
This implements a very naive transfer mechanism without batching (which should be a straightforward change after we have batching in odb.add itself).
This implements a very naive transfer mechanism without
batching (which should be a straightforward change after
we have batching in odb.add itself).