ocf-data-sampler 0.5.24__py3-none-any.whl → 0.5.25__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Potentially problematic release.


This version of ocf-data-sampler might be problematic. Click here for more details.

@@ -75,7 +75,7 @@ def _tensostore_open_zarr_paths(zarr_path: str | list[str], time_dim: str) -> xr
75
75
  zarr_path = sorted(glob(zarr_path))
76
76
 
77
77
  if isinstance(zarr_path, list | tuple):
78
- ds = open_zarrs(zarr_path, concat_dim=time_dim).sortby(time_dim)
78
+ ds = open_zarrs(zarr_path, concat_dim=time_dim, data_source="nwp").sortby(time_dim)
79
79
  else:
80
80
  ds = open_zarr(zarr_path)
81
81
  return ds
@@ -14,6 +14,7 @@ References:
14
14
  [2] https://www.apache.org/licenses/LICENSE-2.0
15
15
  """
16
16
 
17
+ import logging
17
18
  import os.path
18
19
  import re
19
20
 
@@ -26,6 +27,7 @@ from xarray_tensorstore import (
26
27
  _TensorStoreAdapter,
27
28
  )
28
29
 
30
+ logger = logging.getLogger(__name__)
29
31
 
30
32
  def _zarr_spec_from_path(path: str, zarr_format: int) -> ...:
31
33
  if re.match(r"\w+\://", path): # path is a URI
@@ -127,6 +129,7 @@ def open_zarrs(
127
129
  concat_dim: str,
128
130
  context: ts.Context | None = None,
129
131
  mask_and_scale: bool = True,
132
+ data_source: str = "unknown",
130
133
  ) -> xr.Dataset:
131
134
  """Open multiple zarrs with TensorStore.
132
135
 
@@ -135,6 +138,7 @@ def open_zarrs(
135
138
  concat_dim: Dimension along which to concatenate the data variables.
136
139
  context: TensorStore context.
137
140
  mask_and_scale: Whether to mask and scale the data.
141
+ data_source: Which data source is being opened. Used for warning context.
138
142
 
139
143
  Returns:
140
144
  Concatenated Dataset with all data variables opened via TensorStore.
@@ -143,13 +147,28 @@ def open_zarrs(
143
147
  context = ts.Context()
144
148
 
145
149
  ds_list = [xr.open_zarr(p, mask_and_scale=mask_and_scale, decode_timedelta=True) for p in paths]
146
- ds = xr.concat(
147
- ds_list,
148
- dim=concat_dim,
149
- data_vars="minimal",
150
- compat="equals",
151
- combine_attrs="drop_conflicts",
152
- )
150
+ try:
151
+ ds = xr.concat(
152
+ ds_list,
153
+ dim=concat_dim,
154
+ data_vars="minimal",
155
+ compat="equals",
156
+ combine_attrs="drop_conflicts",
157
+ join="exact",
158
+ )
159
+ except ValueError:
160
+ logger.warning(f"Coordinate mismatch found in {data_source} input data. "
161
+ f"The coordinates will be overwritten! "
162
+ f"This might be fine for satellite data. "
163
+ f"Proceed with caution.")
164
+ ds = xr.concat(
165
+ ds_list,
166
+ dim=concat_dim,
167
+ data_vars="minimal",
168
+ compat="equals",
169
+ combine_attrs="drop_conflicts",
170
+ join="override",
171
+ )
153
172
 
154
173
  if mask_and_scale:
155
174
  _raise_if_mask_and_scale_used_for_data_vars(ds)
@@ -19,7 +19,7 @@ def open_sat_data(zarr_path: str | list[str]) -> xr.DataArray:
19
19
  """
20
20
  # Open the data
21
21
  if isinstance(zarr_path, list | tuple):
22
- ds = open_zarrs(zarr_path, concat_dim="time")
22
+ ds = open_zarrs(zarr_path, concat_dim="time", data_source="satellite")
23
23
  else:
24
24
  ds = open_zarr(zarr_path)
25
25
 
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: ocf-data-sampler
3
- Version: 0.5.24
3
+ Version: 0.5.25
4
4
  Author: James Fulton, Peter Dudfield
5
5
  Author-email: Open Climate Fix team <info@openclimatefix.org>
6
6
  License: MIT License
@@ -9,8 +9,8 @@ ocf_data_sampler/data/uk_gsp_locations_20250109.csv,sha256=XZISFatnbpO9j8LwaxNKF
9
9
  ocf_data_sampler/load/__init__.py,sha256=-vQP9g0UOWdVbjEGyVX_ipa7R1btmiETIKAf6aw4d78,201
10
10
  ocf_data_sampler/load/gsp.py,sha256=zsQ39dZBS45qd86lGfCZUjheLRTtMzIUozj-j8c87UQ,3058
11
11
  ocf_data_sampler/load/load_dataset.py,sha256=K8rWykjII-3g127If7WRRFivzHNx3SshCvZj4uQlf28,2089
12
- ocf_data_sampler/load/open_xarray_tensorstore.py,sha256=kAqlIavGe1dcCPkzAtoZo2dFS-tW36E-wRE_3w1HMfg,5620
13
- ocf_data_sampler/load/satellite.py,sha256=B-m0_Py_D0GwzwX5o-ixyeXntV5Z4k4MbmMBHZLUWMM,1831
12
+ ocf_data_sampler/load/open_xarray_tensorstore.py,sha256=YglCBeKa4mSjUU5qlcMOLZXUtFrPFWVKDeKHLjs_YbA,6353
13
+ ocf_data_sampler/load/satellite.py,sha256=5o5SfcplQfZFlm3JJq73j8_m_cWKpFtKk0tTKGjjCuE,1856
14
14
  ocf_data_sampler/load/site.py,sha256=WtOy20VMHJIY0IwEemCdcecSDUGcVaLUown-4ixJw90,2147
15
15
  ocf_data_sampler/load/utils.py,sha256=AGL0aOOQPrgqNBTjlBtR7Qg1PyQov3DFJo-y198u8pY,2044
16
16
  ocf_data_sampler/load/nwp/__init__.py,sha256=SmcrnbygO5xtCKmGR4wtHrj-HI7nOAvnAtfuvRufBGQ,25
@@ -21,7 +21,7 @@ ocf_data_sampler/load/nwp/providers/ecmwf.py,sha256=P7JqfssmQq8eHKKXaBexsxts325A
21
21
  ocf_data_sampler/load/nwp/providers/gfs.py,sha256=h6vm-Rfz1JGOE4P_fP1_XQJ3bugNbeNAIyt56N8B1Dc,1066
22
22
  ocf_data_sampler/load/nwp/providers/icon.py,sha256=iVZwLKRr_D74_kAu5MHir6pRKEfbTmIxFRZAxzmiYdI,1257
23
23
  ocf_data_sampler/load/nwp/providers/ukv.py,sha256=2i32VM9gnmWUpbL0qBSp_AKzuyKucXZPS8yklbcGlbc,1039
24
- ocf_data_sampler/load/nwp/providers/utils.py,sha256=IjJ3w7zDgXNFaVa4TMk8yVCvdzfrIRu5tn1OaaQ7Zso,2304
24
+ ocf_data_sampler/load/nwp/providers/utils.py,sha256=Pyp1U3kUpUoqv6ArEtYBfEfzjO2JGQFlUOL-ndsEmlY,2323
25
25
  ocf_data_sampler/numpy_sample/__init__.py,sha256=5bdpzM8hMAEe0XRSZ9AZFQdqEeBsEPhaF79Y8bDx3GQ,407
26
26
  ocf_data_sampler/numpy_sample/collate.py,sha256=hoxIc5SoHoIs3Nx37aRZzWChpswjy9lHUgaKgHIoo80,2039
27
27
  ocf_data_sampler/numpy_sample/common_types.py,sha256=9CjYHkUTx0ObduWh43fhsybZCTXvexql7qC2ptMDoek,377
@@ -59,7 +59,7 @@ ocf_data_sampler/torch_datasets/utils/valid_time_periods.py,sha256=xcy75cVxl0Wrg
59
59
  ocf_data_sampler/torch_datasets/utils/validation_utils.py,sha256=YqmT-lExWlI8_ul3l0EP73Ik002fStr_bhsZh9mQqEU,4735
60
60
  scripts/download_gsp_location_data.py,sha256=rRDXMoqX-RYY4jPdxhdlxJGhWdl6r245F5UARgKV6P4,3121
61
61
  scripts/refactor_site.py,sha256=skzvsPP0Cn9yTKndzkilyNcGz4DZ88ctvCJ0XrBdc2A,3135
62
- ocf_data_sampler-0.5.24.dist-info/METADATA,sha256=r5w9nbLOb6vB8rIALgrp02EY7HM6KJ43hE0xp3xLGEE,13541
63
- ocf_data_sampler-0.5.24.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
64
- ocf_data_sampler-0.5.24.dist-info/top_level.txt,sha256=deUxqmsONNAGZDNbsntbXH7BRA1MqWaUeAJrCo6q_xA,25
65
- ocf_data_sampler-0.5.24.dist-info/RECORD,,
62
+ ocf_data_sampler-0.5.25.dist-info/METADATA,sha256=9JPhGShknN0Rr_r09f-UqymH1i_DIssLHR2Q8N1bfSE,13541
63
+ ocf_data_sampler-0.5.25.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
64
+ ocf_data_sampler-0.5.25.dist-info/top_level.txt,sha256=deUxqmsONNAGZDNbsntbXH7BRA1MqWaUeAJrCo6q_xA,25
65
+ ocf_data_sampler-0.5.25.dist-info/RECORD,,