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
Copy file name to clipboardExpand all lines: README.md
+3-3Lines changed: 3 additions & 3 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -69,7 +69,7 @@ There are some (relatively obsolete) documents from our exploration of zarr inte
69
69
## Storage types
70
70
71
71
PyActiveStorage is designed to interact with various storage backends.
72
-
The storage backend is automatically detected, but can still be specified using the `storage_type` argument to the `Active` constructor.
72
+
The storage backend is automatically detected, but can still be specified using the `interface_type` argument to the `Active` constructor.
73
73
There are two main integration points for a storage backend:
74
74
75
75
#. Load netCDF metadata
@@ -78,7 +78,7 @@ There are two main integration points for a storage backend:
78
78
### Local file
79
79
80
80
The default storage backend is a local file.
81
-
To use a local file, use a `storage_type` of `None`, which is its default value.
81
+
To use a local file, use a `interface_type` of `None`, which is its default value.
82
82
netCDF metadata is loaded using the [netCDF4](https://pypi.org/project/netCDF4/) library.
83
83
The chunk reductions are implemented in `activestorage.storage` using NumPy.
84
84
@@ -87,7 +87,7 @@ The chunk reductions are implemented in `activestorage.storage` using NumPy.
87
87
We now have support for Active runs with netCDF4 files on S3, from [PR 89](https://github.com/NCAS-CMS/PyActiveStorage/pull/89).
88
88
To achieve this we integrate with [Reductionist](https://github.com/stackhpc/reductionist-rs), an S3 Active Storage Server.
89
89
Reductionist is typically deployed "near" to an S3-compatible object store and provides an API to perform numerical reductions on object data.
90
-
To use Reductionist, use a `storage_type` of `s3`.
90
+
To use Reductionist, use a `interface_type` of `s3`.
91
91
92
92
To load metadata, netCDF files are opened using `s3fs`, with `h5netcdf` used to put the open file (which is nothing more than a memory view of the netCDF file) into an hdf5/netCDF-like object format.
93
93
Chunk reductions are implemented in `activestorage.reductionist`, with each operation resulting in an API request to the Reductionist server.
0 commit comments