ocf-data-sampler 0.0.45__py3-none-any.whl → 0.0.47__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.

@@ -9,7 +9,6 @@ Example:
9
9
  """
10
10
 
11
11
  import json
12
-
13
12
  from pathlib import Path
14
13
  from typing import Union
15
14
 
@@ -18,7 +17,6 @@ import yaml
18
17
 
19
18
  from ocf_data_sampler.config import Configuration
20
19
 
21
-
22
20
  def save_yaml_configuration(
23
21
  configuration: Configuration,
24
22
  filename: Union[str, Path],
@@ -35,7 +33,7 @@ def save_yaml_configuration(
35
33
  Path: The path where the configuration was saved
36
34
 
37
35
  Raises:
38
- ValueError: If filename is None or if writing to the specified path fails
36
+ ValueError: If filename is None, directory doesn't exist, or if writing to the specified path fails
39
37
  TypeError: If the configuration cannot be serialized
40
38
  """
41
39
  if filename is None:
@@ -50,24 +48,37 @@ def save_yaml_configuration(
50
48
 
51
49
  filepath = Path(filename)
52
50
 
53
- # For local files, check if directory exists before proceeding
51
+ # For local paths, check if parent directory exists before attempting to create
54
52
  if filepath.is_absolute():
55
- directory = filepath.parent
56
- if not directory.exists():
53
+ if not filepath.parent.exists():
57
54
  raise ValueError("Directory does not exist")
55
+
56
+ # Only try to create directory if it's in a writable location
57
+ try:
58
+ filepath.parent.mkdir(parents=True, exist_ok=True)
59
+ except PermissionError:
60
+ raise ValueError(f"Permission denied when accessing directory {filepath.parent}")
58
61
 
59
62
  # Serialize configuration to JSON-compatible dictionary
60
63
  config_dict = json.loads(configuration.model_dump_json())
61
64
 
62
- # Save to YAML file using fsspec
63
- with fsspec.open(str(filepath), mode='w') as yaml_file:
64
- yaml.safe_dump(config_dict, yaml_file, default_flow_style=False)
65
+ # Write to file directly for local paths
66
+ if filepath.is_absolute():
67
+ try:
68
+ with open(filepath, 'w') as f:
69
+ yaml.safe_dump(config_dict, f, default_flow_style=False)
70
+ except PermissionError:
71
+ raise ValueError(f"Permission denied when writing to {filename}")
72
+ else:
73
+ # Use fsspec for cloud storage
74
+ with fsspec.open(str(filepath), mode='w') as yaml_file:
75
+ yaml.safe_dump(config_dict, yaml_file, default_flow_style=False)
65
76
 
66
77
  return filepath
67
78
 
68
79
  except json.JSONDecodeError as e:
69
80
  raise TypeError(f"Failed to serialize configuration: {str(e)}") from e
70
- except PermissionError as e:
71
- raise ValueError(f"Permission denied when writing to {filename}") from e
72
81
  except (IOError, OSError) as e:
82
+ if "Permission denied" in str(e):
83
+ raise ValueError(f"Permission denied when writing to {filename}") from e
73
84
  raise ValueError(f"Failed to write configuration to {filename}: {str(e)}") from e
@@ -13,7 +13,7 @@ class SiteSampleKey:
13
13
  solar_elevation = "site_solar_elevation"
14
14
  id = "site_id"
15
15
 
16
-
16
+ # TODO update to include trig datetime + solar coords
17
17
  def convert_site_to_numpy_sample(da: xr.DataArray, t0_idx: int | None = None) -> dict:
18
18
  """Convert from Xarray to NumpySample"""
19
19
 
@@ -22,6 +22,7 @@ def convert_site_to_numpy_sample(da: xr.DataArray, t0_idx: int | None = None) ->
22
22
  SiteSampleKey.generation: da.values,
23
23
  SiteSampleKey.capacity_kwp: da.isel(time_utc=0)["capacity_kwp"].values,
24
24
  SiteSampleKey.time_utc: da["time_utc"].values.astype(float),
25
+ SiteSampleKey.id: da["site_id"].values,
25
26
  }
26
27
 
27
28
  if t0_idx is not None:
@@ -5,16 +5,114 @@ import pandas as pd
5
5
  import pkg_resources
6
6
  import xarray as xr
7
7
  from torch.utils.data import Dataset
8
-
9
8
  from ocf_data_sampler.config import Configuration, load_yaml_configuration
10
9
  from ocf_data_sampler.load.load_dataset import get_dataset_dict
11
10
  from ocf_data_sampler.select import fill_time_periods, Location, slice_datasets_by_space, slice_datasets_by_time
12
11
  from ocf_data_sampler.utils import minutes
13
- from ocf_data_sampler.torch_datasets.process_and_combine import process_and_combine_datasets, compute
14
- from ocf_data_sampler.torch_datasets.valid_time_periods import find_valid_time_periods
12
+ from ocf_data_sampler.torch_datasets.utils.valid_time_periods import find_valid_time_periods
13
+ from ocf_data_sampler.constants import NWP_MEANS, NWP_STDS, RSS_MEAN, RSS_STD
14
+ from ocf_data_sampler.numpy_sample import (
15
+ convert_nwp_to_numpy_sample,
16
+ convert_satellite_to_numpy_sample,
17
+ convert_gsp_to_numpy_sample,
18
+ make_sun_position_numpy_sample,
19
+ )
20
+ from ocf_data_sampler.torch_datasets.utils.merge_and_fill_utils import (
21
+ merge_dicts,
22
+ fill_nans_in_arrays,
23
+ )
24
+ from ocf_data_sampler.numpy_sample.gsp import GSPSampleKey
25
+ from ocf_data_sampler.numpy_sample.nwp import NWPSampleKey
26
+ from ocf_data_sampler.select.geospatial import osgb_to_lon_lat
15
27
 
16
28
  xr.set_options(keep_attrs=True)
17
29
 
30
+ def process_and_combine_datasets(
31
+ dataset_dict: dict,
32
+ config: Configuration,
33
+ t0: pd.Timestamp,
34
+ location: Location,
35
+ target_key: str = 'gsp'
36
+ ) -> dict:
37
+
38
+ """Normalise and convert data to numpy arrays"""
39
+ numpy_modalities = []
40
+
41
+ if "nwp" in dataset_dict:
42
+
43
+ nwp_numpy_modalities = dict()
44
+
45
+ for nwp_key, da_nwp in dataset_dict["nwp"].items():
46
+ # Standardise
47
+ provider = config.input_data.nwp[nwp_key].provider
48
+ da_nwp = (da_nwp - NWP_MEANS[provider]) / NWP_STDS[provider]
49
+
50
+ # Convert to NumpyBatch
51
+ nwp_numpy_modalities[nwp_key] = convert_nwp_to_numpy_sample(da_nwp)
52
+
53
+ # Combine the NWPs into NumpyBatch
54
+ numpy_modalities.append({NWPSampleKey.nwp: nwp_numpy_modalities})
55
+
56
+
57
+ if "sat" in dataset_dict:
58
+ # Standardise
59
+ da_sat = dataset_dict["sat"]
60
+ da_sat = (da_sat - RSS_MEAN) / RSS_STD
61
+
62
+ # Convert to NumpyBatch
63
+ numpy_modalities.append(convert_satellite_to_numpy_sample(da_sat))
64
+
65
+ gsp_config = config.input_data.gsp
66
+
67
+ if "gsp" in dataset_dict:
68
+ da_gsp = xr.concat([dataset_dict["gsp"], dataset_dict["gsp_future"]], dim="time_utc")
69
+ da_gsp = da_gsp / da_gsp.effective_capacity_mwp
70
+
71
+ numpy_modalities.append(
72
+ convert_gsp_to_numpy_sample(
73
+ da_gsp,
74
+ t0_idx=-gsp_config.interval_start_minutes / gsp_config.time_resolution_minutes
75
+ )
76
+ )
77
+
78
+ # Add coordinate data
79
+ # TODO: Do we need all of these?
80
+ numpy_modalities.append(
81
+ {
82
+ GSPSampleKey.gsp_id: location.id,
83
+ GSPSampleKey.x_osgb: location.x,
84
+ GSPSampleKey.y_osgb: location.y,
85
+ }
86
+ )
87
+
88
+ if target_key == 'gsp':
89
+ # Make sun coords NumpySample
90
+ datetimes = pd.date_range(
91
+ t0+minutes(gsp_config.interval_start_minutes),
92
+ t0+minutes(gsp_config.interval_end_minutes),
93
+ freq=minutes(gsp_config.time_resolution_minutes),
94
+ )
95
+
96
+ lon, lat = osgb_to_lon_lat(location.x, location.y)
97
+
98
+ numpy_modalities.append(
99
+ make_sun_position_numpy_sample(datetimes, lon, lat, key_prefix=target_key)
100
+ )
101
+
102
+ # Combine all the modalities and fill NaNs
103
+ combined_sample = merge_dicts(numpy_modalities)
104
+ combined_sample = fill_nans_in_arrays(combined_sample)
105
+
106
+ return combined_sample
107
+
108
+ def compute(xarray_dict: dict) -> dict:
109
+ """Eagerly load a nested dictionary of xarray DataArrays"""
110
+ for k, v in xarray_dict.items():
111
+ if isinstance(v, dict):
112
+ xarray_dict[k] = compute(v)
113
+ else:
114
+ xarray_dict[k] = v.compute(scheduler="single-threaded")
115
+ return xarray_dict
18
116
 
19
117
  def find_valid_t0_times(
20
118
  datasets_dict: dict,
@@ -48,7 +146,7 @@ def get_gsp_locations(gsp_ids: list[int] | None = None) -> list[Location]:
48
146
 
49
147
  # Load UK GSP locations
50
148
  df_gsp_loc = pd.read_csv(
51
- pkg_resources.resource_filename(__name__, "../data/uk_gsp_locations.csv"),
149
+ pkg_resources.resource_filename(__name__, "../../data/uk_gsp_locations.csv"),
52
150
  index_col="gsp_id",
53
151
  )
54
152
 
@@ -17,8 +17,8 @@ from ocf_data_sampler.select import (
17
17
  slice_datasets_by_time, slice_datasets_by_space
18
18
  )
19
19
  from ocf_data_sampler.utils import minutes
20
- from ocf_data_sampler.torch_datasets.valid_time_periods import find_valid_time_periods
21
- from ocf_data_sampler.torch_datasets.process_and_combine import merge_dicts, fill_nans_in_arrays
20
+ from ocf_data_sampler.torch_datasets.utils.valid_time_periods import find_valid_time_periods
21
+ from ocf_data_sampler.torch_datasets.utils.merge_and_fill_utils import merge_dicts, fill_nans_in_arrays
22
22
  from ocf_data_sampler.numpy_sample import (
23
23
  convert_site_to_numpy_sample,
24
24
  convert_satellite_to_numpy_sample,
@@ -257,6 +257,8 @@ class SitesDataset(Dataset):
257
257
  )
258
258
  combined_sample_dataset = xr.merge([combined_sample_dataset, sun_position_features_xr])
259
259
 
260
+ # TODO include t0_index in xr dataset?
261
+
260
262
  # Fill any nan values
261
263
  return combined_sample_dataset.fillna(0.0)
262
264
 
@@ -317,6 +319,26 @@ class SitesDataset(Dataset):
317
319
 
318
320
  # ----- functions to load presaved samples ------
319
321
 
322
+ def convert_netcdf_to_numpy_sample(ds: xr.Dataset) -> dict:
323
+ """Convert a netcdf dataset to a numpy sample"""
324
+
325
+ # convert the single dataset to a dict of arrays
326
+ sample_dict = convert_from_dataset_to_dict_datasets(ds)
327
+
328
+ if "satellite" in sample_dict:
329
+ # rename satellite to satellite actual # TODO this could be improves
330
+ sample_dict["sat"] = sample_dict.pop("satellite")
331
+
332
+ # process and combine the datasets
333
+ sample = convert_to_numpy_and_combine(
334
+ dataset_dict=sample_dict,
335
+ )
336
+
337
+ # TODO think about normalization, maybe its done not in sample creation, maybe its done afterwards,
338
+ # to allow it to be flexible
339
+
340
+ return sample
341
+
320
342
  def convert_from_dataset_to_dict_datasets(combined_dataset: xr.Dataset) -> dict[str, xr.DataArray]:
321
343
  """
322
344
  Convert a combined sample dataset to a dict of datasets for each input
@@ -360,26 +382,6 @@ def nest_nwp_source_dict(d: dict, sep: str = "/") -> dict:
360
382
  new_dict["nwp"] = nwp_subdict
361
383
  return new_dict
362
384
 
363
- def convert_netcdf_to_numpy_sample(ds: xr.Dataset) -> dict:
364
- """Convert a netcdf dataset to a numpy sample"""
365
-
366
- # convert the single dataset to a dict of arrays
367
- sample_dict = convert_from_dataset_to_dict_datasets(ds)
368
-
369
- if "satellite" in sample_dict:
370
- # rename satellite to satellite actual # TODO this could be improves
371
- sample_dict["sat"] = sample_dict.pop("satellite")
372
-
373
- # process and combine the datasets
374
- sample = convert_to_numpy_and_combine(
375
- dataset_dict=sample_dict,
376
- )
377
-
378
- # TODO think about normalization, maybe its done not in sample creation, maybe its done afterwards,
379
- # to allow it to be flexible
380
-
381
- return sample
382
-
383
385
  def convert_to_numpy_and_combine(
384
386
  dataset_dict: dict,
385
387
  ) -> dict:
@@ -406,7 +408,6 @@ def convert_to_numpy_and_combine(
406
408
 
407
409
  if "site" in dataset_dict:
408
410
  da_sites = dataset_dict["site"]
409
- sites_sample = convert_site_to_numpy_sample(da_sites)
410
411
 
411
412
  numpy_modalities.append(
412
413
  convert_site_to_numpy_sample(
@@ -414,8 +415,6 @@ def convert_to_numpy_and_combine(
414
415
  )
415
416
  )
416
417
 
417
- numpy_modalities.append(sites_sample)
418
-
419
418
  # Combine all the modalities and fill NaNs
420
419
  combined_sample = merge_dicts(numpy_modalities)
421
420
  combined_sample = fill_nans_in_arrays(combined_sample)
@@ -0,0 +1,25 @@
1
+ import numpy as np
2
+
3
+ def merge_dicts(list_of_dicts: list[dict]) -> dict:
4
+ """Merge a list of dictionaries into a single dictionary"""
5
+ # TODO: This doesn't account for duplicate keys, which will be overwritten
6
+ combined_dict = {}
7
+ for d in list_of_dicts:
8
+ combined_dict.update(d)
9
+ return combined_dict
10
+
11
+ def fill_nans_in_arrays(sample: dict) -> dict:
12
+ """Fills all NaN values in each np.ndarray in the sample dictionary with zeros.
13
+
14
+ Operation is performed in-place on the sample.
15
+ """
16
+ for k, v in sample.items():
17
+ if isinstance(v, np.ndarray) and np.issubdtype(v.dtype, np.number):
18
+ if np.isnan(v).any():
19
+ sample[k] = np.nan_to_num(v, copy=False, nan=0.0)
20
+
21
+ # Recursion is included to reach NWP arrays in subdict
22
+ elif isinstance(v, dict):
23
+ fill_nans_in_arrays(v)
24
+
25
+ return sample
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.2
2
2
  Name: ocf_data_sampler
3
- Version: 0.0.45
3
+ Version: 0.0.47
4
4
  Summary: Sample from weather data for renewable energy prediction
5
5
  Author: James Fulton, Peter Dudfield, and the Open Climate Fix team
6
6
  Author-email: info@openclimatefix.org
@@ -4,7 +4,7 @@ ocf_data_sampler/utils.py,sha256=rKA0BHAyAG4f90zEcgxp25EEYrXS-aOVNzttZ6Mzv2k,250
4
4
  ocf_data_sampler/config/__init__.py,sha256=YXnAkgHViHB26hSsjiv32b6EbpG-A1kKTkARJf0_RkY,212
5
5
  ocf_data_sampler/config/load.py,sha256=4f7vPHAIAmd-55tPxoIzn7F_TI_ue4NxkDcLPoVWl0g,943
6
6
  ocf_data_sampler/config/model.py,sha256=sXmh7IadwXDT-7lxEl5_b3vjovZgZYR77EXy4GHaf4w,7276
7
- ocf_data_sampler/config/save.py,sha256=JhPyX1DJAW-mi3GfkDF4nlrEXNt-yfX9ibBy8CZ6ENI,2464
7
+ ocf_data_sampler/config/save.py,sha256=gB44isAZWUlCe3L6VBkLkngWC9GFpcCfAM57gy-0dkg,3156
8
8
  ocf_data_sampler/data/uk_gsp_locations.csv,sha256=RSh7DRh55E3n8lVAaWXGTaXXHevZZtI58td4d4DhGos,10415772
9
9
  ocf_data_sampler/load/__init__.py,sha256=MjgfxilTzyz1RYFoBEeAXmE9hyjknLvdmlHPmlAoiQY,44
10
10
  ocf_data_sampler/load/gsp.py,sha256=Gcr1JVUOPKhFRDCSHtfPDjxx0BtyyEhXrZvGEKLPJ5I,759
@@ -24,7 +24,7 @@ ocf_data_sampler/numpy_sample/datetime_features.py,sha256=U-9uRplfZ7VYFA4qBduI8O
24
24
  ocf_data_sampler/numpy_sample/gsp.py,sha256=5UaWO_aGRRVQo82wnDaT4zBKHihOnIsXiwgPjM8vGFM,1005
25
25
  ocf_data_sampler/numpy_sample/nwp.py,sha256=_seQNWsut3IzPsrpipqImjnaM3XNHZCy5_5be6syivk,1297
26
26
  ocf_data_sampler/numpy_sample/satellite.py,sha256=8OaTvkPjzSjotcdKsa6BKmmlBKDBunbhDN4Pjo0Grxs,910
27
- ocf_data_sampler/numpy_sample/site.py,sha256=PIfmCtPA37dqpC8GArkryVqFrNAwqacj0iW2ikBOdSk,789
27
+ ocf_data_sampler/numpy_sample/site.py,sha256=cOVpFN_EVRD0d4TJtmPdNYcWjiWuWr8eswktC97KR8Q,890
28
28
  ocf_data_sampler/numpy_sample/sun_position.py,sha256=UklhucCxCT6GMlAhCWL6c4cfWrdc1cWgegrYaqUoHOY,1611
29
29
  ocf_data_sampler/select/__init__.py,sha256=E4AJulEbO2K-o0UlG1fgaEteuf_1ZFjHTvrotXSb4YU,332
30
30
  ocf_data_sampler/select/dropout.py,sha256=HCx5Wzk8Oh2Z9vV94Jy-ALJsHtGduwvMaQOleQXp5z0,1142
@@ -36,21 +36,21 @@ ocf_data_sampler/select/select_spatial_slice.py,sha256=WNxwur9Q5oetvogATw8-hNejD
36
36
  ocf_data_sampler/select/select_time_slice.py,sha256=D5P_cSvnv8Qs49K5au7lPxDr9U_VmDn42s5leMzHt0k,6122
37
37
  ocf_data_sampler/select/spatial_slice_for_dataset.py,sha256=3tRrMBXr7s4CnClbVSIq7hpls3H4Y3qYTDwswcxCCCE,1763
38
38
  ocf_data_sampler/select/time_slice_for_dataset.py,sha256=LMw8KnOCKnPjD0m4UubAWERpaiQtzRKkI2cSh5a0A-M,4335
39
- ocf_data_sampler/torch_datasets/__init__.py,sha256=nJUa2KzVa84ZoM0PT2AbDz26ennmAYc7M7WJVfypPMs,85
40
- ocf_data_sampler/torch_datasets/process_and_combine.py,sha256=BaZZwGUKQf--X-3cNHcxZRn6iTNXg-sZyaYeRzQ7V68,4306
41
- ocf_data_sampler/torch_datasets/pvnet_uk_regional.py,sha256=QRFqbdfNchVWj4y70n-rJdFvFGvQj-WpZLdFqWjnOTw,5543
42
- ocf_data_sampler/torch_datasets/site.py,sha256=bBqKQt0lhk90-d1ZkBCeHHEIAFCnoQbRs2AQYyulW40,15381
43
- ocf_data_sampler/torch_datasets/valid_time_periods.py,sha256=Qo65qUHtle_bW5tLTYr7empHTRv-lpjvfx_6GNJj3Xg,4371
39
+ ocf_data_sampler/torch_datasets/datasets/__init__.py,sha256=nJUa2KzVa84ZoM0PT2AbDz26ennmAYc7M7WJVfypPMs,85
40
+ ocf_data_sampler/torch_datasets/datasets/pvnet_uk_regional.py,sha256=xxeX4Js9LQpydehi3BS7k9psqkYGzgJuM17uTYux40M,8742
41
+ ocf_data_sampler/torch_datasets/datasets/site.py,sha256=75M0oDstOLyLZBySVIS6fLJSbEjfxcWBlgGP_ewui7s,15334
42
+ ocf_data_sampler/torch_datasets/utils/merge_and_fill_utils.py,sha256=hIbekql64eXsNDFIoEc--GWxwdVWrh2qKegdOi70Bow,874
43
+ ocf_data_sampler/torch_datasets/utils/valid_time_periods.py,sha256=Qo65qUHtle_bW5tLTYr7empHTRv-lpjvfx_6GNJj3Xg,4371
44
44
  scripts/refactor_site.py,sha256=asZ27hQ4IyXgCCUaFJqcz1ObBNcV2W3ywqHBpSXA_fc,1728
45
45
  tests/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
46
- tests/conftest.py,sha256=68hH-HPdHPLvLrtYJU8bjfkdGKbhPfNveLKvUs6_Lr0,7970
47
- tests/config/test_config.py,sha256=eaye_F7-el4tTP4n2vRME8qlV0b2jaKUX4HhgOUpa7E,5203
46
+ tests/conftest.py,sha256=DfrH0Pm552Tnl35eZn2UHCfOn2lHRiEQCcUCJIhycSU,8021
47
+ tests/config/test_config.py,sha256=Vq_kTL5tJcwEP-hXD_Nah5O6cgafo99iX6Fw1AN5NDY,5288
48
48
  tests/config/test_save.py,sha256=rA_XVxP1pOxB--5Ebujz4T5o-VbcrCbg2VSlSq2iI0o,1318
49
49
  tests/load/test_load_gsp.py,sha256=aT_nqaSXmUTcdHzuTT7AmXJr3R31k4OEN-Fv3eLxlQE,424
50
50
  tests/load/test_load_nwp.py,sha256=3qyyDkB1q9t3tyAwogfotNrxqUOpXXimco1CImoEWGg,753
51
51
  tests/load/test_load_satellite.py,sha256=STX5AqqmOAgUgE9R1xyq_sM3P1b8NKdGjO-hDhayfxM,524
52
52
  tests/load/test_load_sites.py,sha256=T9lSEnGPI8FQISudVYHHNTHeplNS62Vrx48jaZ6J_Jo,364
53
- tests/numpy_sample/test_collate.py,sha256=oWVNRR0LyF0iqaxjsq6hPMmkSGIyvGVlAM6q2rqvXqQ,840
53
+ tests/numpy_sample/test_collate.py,sha256=ngbJ8vIewnAvkXx-PpfuSMVNM82_SYaZPLhJkZZw7s0,867
54
54
  tests/numpy_sample/test_datetime_features.py,sha256=o4t3KeKFdGrOBQ77rNFcDuDMQSD23ileCS5T5AP3wG4,1769
55
55
  tests/numpy_sample/test_gsp.py,sha256=FLlq4SlJ-9cSRAepf4_ksA6PsUVKegnKEAc5pUojCJ0,1458
56
56
  tests/numpy_sample/test_nwp.py,sha256=yf4u7mAU0E3FQ4xAH6YjuHuHBzzFoXjHSFNkOVJUdSM,1455
@@ -63,11 +63,11 @@ tests/select/test_location.py,sha256=_WZk2FPYeJ-nIfCJS6Sp_yaVEEo7m31DmMFoZzgyCts
63
63
  tests/select/test_select_spatial_slice.py,sha256=7EX9b6g-pMdACQx3yefjs5do2s-Rho2UmKevV4oglsU,5147
64
64
  tests/select/test_select_time_slice.py,sha256=K1EJR5TwZa9dJf_YTEHxGtvs398iy1xS2lr1BgJZkoo,9603
65
65
  tests/torch_datasets/conftest.py,sha256=eRCzHE7cxS4AoskExkCGFDBeqItktAYNAdkfpMoFCeE,629
66
- tests/torch_datasets/test_process_and_combine.py,sha256=mbjQdqzLhox-U2sc1Ec68xLz95b3XOyPa7WchgxUM88,4256
67
- tests/torch_datasets/test_pvnet_uk_regional.py,sha256=dLY861PMyQ_buTksP8d0UXzfKsZ_CFNgceSYVGXRIRs,2134
68
- tests/torch_datasets/test_site.py,sha256=PGsGgg4jCDasxBg09hW9DA6rOGf60ECIh5YnBWyyzDE,4675
69
- ocf_data_sampler-0.0.45.dist-info/LICENSE,sha256=F-Q3UFCR-BECSocV55BFDpn4YKxve9PKrm-lTt6o_Tg,1073
70
- ocf_data_sampler-0.0.45.dist-info/METADATA,sha256=5QNcC3hPXLfehwBcxUVlidlPwSlMEoQWp5ln9JE_mzQ,11788
71
- ocf_data_sampler-0.0.45.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
72
- ocf_data_sampler-0.0.45.dist-info/top_level.txt,sha256=Faob6N6cFdPc5eUpCTYcXgCaNhi4XLLteUL5W5ayYmg,31
73
- ocf_data_sampler-0.0.45.dist-info/RECORD,,
66
+ tests/torch_datasets/test_merge_and_fill_utils.py,sha256=ueA0A7gZaWEgNdsU8p3CnKuvSnlleTUjEhSw2HUUROM,1229
67
+ tests/torch_datasets/test_pvnet_uk_regional.py,sha256=FCiFueeFqrsXe7gWguSjBz5ZeUrvyhGbGw81gaVvkHM,5087
68
+ tests/torch_datasets/test_site.py,sha256=0tnjgx6z4VlzjoF_V2p3Y2t2Z1d0o_07Vwb-FH_c3tU,4640
69
+ ocf_data_sampler-0.0.47.dist-info/LICENSE,sha256=F-Q3UFCR-BECSocV55BFDpn4YKxve9PKrm-lTt6o_Tg,1073
70
+ ocf_data_sampler-0.0.47.dist-info/METADATA,sha256=x4HkuNvlIxd5LzUtf5keEgRPFoK8BuV3kpMjogShv5w,11788
71
+ ocf_data_sampler-0.0.47.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
72
+ ocf_data_sampler-0.0.47.dist-info/top_level.txt,sha256=Faob6N6cFdPc5eUpCTYcXgCaNhi4XLLteUL5W5ayYmg,31
73
+ ocf_data_sampler-0.0.47.dist-info/RECORD,,
@@ -2,7 +2,7 @@ import tempfile
2
2
 
3
3
  import pytest
4
4
  from pydantic import ValidationError
5
-
5
+ from pathlib import Path
6
6
  from ocf_data_sampler.config import (
7
7
  load_yaml_configuration,
8
8
  Configuration,
@@ -21,39 +21,37 @@ def test_load_yaml_configuration(test_config_filename):
21
21
  Test that yaml loading works for 'test_config.yaml'
22
22
  and fails for an empty .yaml file
23
23
  """
24
-
25
- # check we get an error if loading a file with no config
26
- with tempfile.NamedTemporaryFile(suffix=".yaml") as fp:
27
- filename = fp.name
28
-
29
- # check that temp file can't be loaded
24
+ # Create temporary directory instead of file
25
+ with tempfile.TemporaryDirectory() as temp_dir:
26
+ # Create path for empty file
27
+ empty_file = Path(temp_dir) / "empty.yaml"
28
+
29
+ # Create an empty file
30
+ empty_file.touch()
31
+
32
+ # Test loading empty file
30
33
  with pytest.raises(TypeError):
31
- _ = load_yaml_configuration(filename)
32
-
33
- # test can load test_config.yaml
34
- config = load_yaml_configuration(test_config_filename)
35
-
36
- assert isinstance(config, Configuration)
37
-
34
+ _ = load_yaml_configuration(str(empty_file))
38
35
 
39
36
  def test_yaml_save(test_config_filename):
40
37
  """
41
38
  Check configuration can be saved to a .yaml file
42
39
  """
43
-
44
40
  test_config = load_yaml_configuration(test_config_filename)
45
-
46
- with tempfile.NamedTemporaryFile(suffix=".yaml") as fp:
47
- filename = fp.name
48
-
49
- # save default config to file
50
- save_yaml_configuration(test_config, filename)
51
-
52
- # check the file can be loaded back
53
- tmp_config = load_yaml_configuration(filename)
54
-
55
- # check loaded configuration is the same as the one passed to save
56
- assert test_config == tmp_config
41
+
42
+ with tempfile.TemporaryDirectory() as temp_dir:
43
+ # Create path for config file
44
+ config_path = Path(temp_dir) / "test_config.yaml"
45
+
46
+ # Save configuration
47
+ saved_path = save_yaml_configuration(test_config, config_path)
48
+
49
+ # Verify file exists
50
+ assert saved_path.exists()
51
+
52
+ # Test loading saved configuration
53
+ loaded_config = load_yaml_configuration(str(saved_path))
54
+ assert loaded_config == test_config
57
55
 
58
56
 
59
57
  def test_extra_field_error():
tests/conftest.py CHANGED
@@ -1,10 +1,10 @@
1
1
  import os
2
-
3
2
  import numpy as np
4
3
  import pandas as pd
5
4
  import pytest
6
5
  import xarray as xr
7
6
  import tempfile
7
+ from typing import Generator
8
8
 
9
9
  from ocf_data_sampler.config.model import Site
10
10
  from ocf_data_sampler.config import load_yaml_configuration, save_yaml_configuration
@@ -201,7 +201,7 @@ def ds_uk_gsp():
201
201
 
202
202
 
203
203
  @pytest.fixture(scope="session")
204
- def data_sites() -> Site:
204
+ def data_sites() -> Generator[Site, None, None]:
205
205
  """
206
206
  Make fake data for sites
207
207
  Returns: filename for netcdf file, and csv metadata
@@ -1,6 +1,6 @@
1
1
  from ocf_data_sampler.numpy_sample import GSPSampleKey, SatelliteSampleKey
2
2
  from ocf_data_sampler.numpy_sample.collate import stack_np_samples_into_batch
3
- from ocf_data_sampler.torch_datasets import PVNetUKRegionalDataset
3
+ from ocf_data_sampler.torch_datasets.datasets.pvnet_uk_regional import PVNetUKRegionalDataset
4
4
 
5
5
 
6
6
  def test_pvnet(pvnet_config_filename):
@@ -0,0 +1,42 @@
1
+ import numpy as np
2
+
3
+ from ocf_data_sampler.torch_datasets.utils.merge_and_fill_utils import (
4
+ merge_dicts,
5
+ fill_nans_in_arrays,
6
+ )
7
+
8
+ def test_merge_dicts():
9
+ """Test merge_dicts function"""
10
+ dict1 = {"a": 1, "b": 2}
11
+ dict2 = {"c": 3, "d": 4}
12
+ dict3 = {"e": 5}
13
+
14
+ result = merge_dicts([dict1, dict2, dict3])
15
+ assert result == {"a": 1, "b": 2, "c": 3, "d": 4, "e": 5}
16
+
17
+ # Test key overwriting
18
+ dict4 = {"a": 10, "f": 6}
19
+ result = merge_dicts([dict1, dict4])
20
+ assert result["a"] == 10
21
+
22
+
23
+ def test_fill_nans_in_arrays():
24
+ """Test the fill_nans_in_arrays function"""
25
+ array_with_nans = np.array([1.0, np.nan, 3.0, np.nan])
26
+ nested_dict = {
27
+ "array1": array_with_nans,
28
+ "nested": {
29
+ "array2": np.array([np.nan, 2.0, np.nan, 4.0])
30
+ },
31
+ "string_key": "not_an_array"
32
+ }
33
+
34
+ result = fill_nans_in_arrays(nested_dict)
35
+
36
+ assert not np.isnan(result["array1"]).any()
37
+ assert np.array_equal(result["array1"], np.array([1.0, 0.0, 3.0, 0.0]))
38
+ assert not np.isnan(result["nested"]["array2"]).any()
39
+ assert np.array_equal(result["nested"]["array2"], np.array([0.0, 2.0, 0.0, 4.0]))
40
+ assert result["string_key"] == "not_an_array"
41
+
42
+
@@ -1,11 +1,88 @@
1
- import pytest
1
+ import numpy as np
2
+ import pandas as pd
3
+ import xarray as xr
4
+ import dask.array as da
2
5
  import tempfile
3
6
 
4
- from ocf_data_sampler.torch_datasets import PVNetUKRegionalDataset
5
- from ocf_data_sampler.config import load_yaml_configuration, save_yaml_configuration
7
+ from ocf_data_sampler.torch_datasets.datasets.pvnet_uk_regional import PVNetUKRegionalDataset
8
+ from ocf_data_sampler.config.save import save_yaml_configuration
9
+ from ocf_data_sampler.config.load import load_yaml_configuration
6
10
  from ocf_data_sampler.numpy_sample import NWPSampleKey, GSPSampleKey, SatelliteSampleKey
11
+ from ocf_data_sampler.torch_datasets.datasets.pvnet_uk_regional import process_and_combine_datasets, compute
12
+ from ocf_data_sampler.select.location import Location
7
13
 
14
+ def test_process_and_combine_datasets(pvnet_config_filename):
8
15
 
16
+ # Load in config for function and define location
17
+ config = load_yaml_configuration(pvnet_config_filename)
18
+ t0 = pd.Timestamp("2024-01-01 00:00")
19
+ location = Location(coordinate_system="osgb", x=1234, y=5678, id=1)
20
+
21
+ nwp_data = xr.DataArray(
22
+ np.random.rand(4, 2, 2, 2),
23
+ dims=["time_utc", "channel", "y", "x"],
24
+ coords={
25
+ "time_utc": pd.date_range("2024-01-01 00:00", periods=4, freq="h"),
26
+ "channel": ["t2m", "dswrf"],
27
+ "step": ("time_utc", pd.timedelta_range(start='0h', periods=4, freq='h')),
28
+ "init_time_utc": pd.Timestamp("2024-01-01 00:00")
29
+ }
30
+ )
31
+
32
+ sat_data = xr.DataArray(
33
+ np.random.rand(7, 1, 2, 2),
34
+ dims=["time_utc", "channel", "y", "x"],
35
+ coords={
36
+ "time_utc": pd.date_range("2024-01-01 00:00", periods=7, freq="5min"),
37
+ "channel": ["HRV"],
38
+ "x_geostationary": (["y", "x"], np.array([[1, 2], [1, 2]])),
39
+ "y_geostationary": (["y", "x"], np.array([[1, 1], [2, 2]]))
40
+ }
41
+ )
42
+
43
+ # Combine as dict
44
+ dataset_dict = {
45
+ "nwp": {"ukv": nwp_data},
46
+ "sat": sat_data
47
+ }
48
+
49
+ # Call relevant function
50
+ result = process_and_combine_datasets(dataset_dict, config, t0, location)
51
+
52
+ # Assert result is dict - check and validate
53
+ assert isinstance(result, dict)
54
+ assert NWPSampleKey.nwp in result
55
+ assert result[SatelliteSampleKey.satellite_actual].shape == (7, 1, 2, 2)
56
+ assert result[NWPSampleKey.nwp]["ukv"][NWPSampleKey.nwp].shape == (4, 1, 2, 2)
57
+
58
+ def test_compute():
59
+ """Test compute function with dask array"""
60
+ da_dask = xr.DataArray(da.random.random((5, 5)))
61
+
62
+ # Create a nested dictionary with dask array
63
+ nested_dict = {
64
+ "array1": da_dask,
65
+ "nested": {
66
+ "array2": da_dask
67
+ }
68
+ }
69
+
70
+ # Ensure initial data is lazy - i.e. not yet computed
71
+ assert not isinstance(nested_dict["array1"].data, np.ndarray)
72
+ assert not isinstance(nested_dict["nested"]["array2"].data, np.ndarray)
73
+
74
+ # Call the compute function
75
+ result = compute(nested_dict)
76
+
77
+ # Assert that the result is an xarray DataArray and no longer lazy
78
+ assert isinstance(result["array1"], xr.DataArray)
79
+ assert isinstance(result["nested"]["array2"], xr.DataArray)
80
+ assert isinstance(result["array1"].data, np.ndarray)
81
+ assert isinstance(result["nested"]["array2"].data, np.ndarray)
82
+
83
+ # Ensure there no NaN values in computed data
84
+ assert not np.isnan(result["array1"].data).any()
85
+ assert not np.isnan(result["nested"]["array2"].data).any()
9
86
 
10
87
  def test_pvnet(pvnet_config_filename):
11
88
 
@@ -1,8 +1,6 @@
1
1
  import pandas as pd
2
-
3
- from ocf_data_sampler.torch_datasets import SitesDataset
4
- from ocf_data_sampler.torch_datasets.site import convert_from_dataset_to_dict_datasets
5
2
  import numpy as np
3
+ from ocf_data_sampler.torch_datasets.datasets.site import SitesDataset, convert_from_dataset_to_dict_datasets
6
4
  from xarray import Dataset, DataArray
7
5
 
8
6
 
@@ -1,132 +0,0 @@
1
- import numpy as np
2
- import pandas as pd
3
- import xarray as xr
4
- from typing import Optional
5
-
6
- from ocf_data_sampler.config import Configuration
7
- from ocf_data_sampler.constants import NWP_MEANS, NWP_STDS,RSS_MEAN,RSS_STD
8
- from ocf_data_sampler.numpy_sample import (
9
- convert_nwp_to_numpy_sample,
10
- convert_satellite_to_numpy_sample,
11
- convert_gsp_to_numpy_sample,
12
- make_sun_position_numpy_sample,
13
-
14
- )
15
- from ocf_data_sampler.numpy_sample.gsp import GSPSampleKey
16
- from ocf_data_sampler.numpy_sample.nwp import NWPSampleKey
17
- from ocf_data_sampler.select.geospatial import osgb_to_lon_lat
18
- from ocf_data_sampler.select.location import Location
19
- from ocf_data_sampler.utils import minutes
20
-
21
-
22
- def process_and_combine_datasets(
23
- dataset_dict: dict,
24
- config: Configuration,
25
- t0: Optional[pd.Timestamp] = None,
26
- location: Optional[Location] = None,
27
- target_key: str = 'gsp'
28
- ) -> dict:
29
-
30
- """Normalise and convert data to numpy arrays"""
31
- numpy_modalities = []
32
-
33
- if "nwp" in dataset_dict:
34
-
35
- nwp_numpy_modalities = dict()
36
-
37
- for nwp_key, da_nwp in dataset_dict["nwp"].items():
38
- # Standardise
39
- provider = config.input_data.nwp[nwp_key].provider
40
- da_nwp = (da_nwp - NWP_MEANS[provider]) / NWP_STDS[provider]
41
- # Convert to NumpySample
42
- nwp_numpy_modalities[nwp_key] = convert_nwp_to_numpy_sample(da_nwp)
43
-
44
- # Combine the NWPs into NumpySample
45
- numpy_modalities.append({NWPSampleKey.nwp: nwp_numpy_modalities})
46
-
47
-
48
- if "sat" in dataset_dict:
49
- # Standardise
50
- da_sat = dataset_dict["sat"]
51
- da_sat = (da_sat - RSS_MEAN) / RSS_STD
52
-
53
- # Convert to NumpySample
54
- numpy_modalities.append(convert_satellite_to_numpy_sample(da_sat))
55
-
56
-
57
- gsp_config = config.input_data.gsp
58
-
59
- if "gsp" in dataset_dict:
60
- da_gsp = xr.concat([dataset_dict["gsp"], dataset_dict["gsp_future"]], dim="time_utc")
61
- da_gsp = da_gsp / da_gsp.effective_capacity_mwp
62
-
63
- numpy_modalities.append(
64
- convert_gsp_to_numpy_sample(
65
- da_gsp,
66
- t0_idx=-gsp_config.interval_start_minutes / gsp_config.time_resolution_minutes
67
- )
68
- )
69
-
70
- # Add coordinate data
71
- # TODO: Do we need all of these?
72
- numpy_modalities.append(
73
- {
74
- GSPSampleKey.gsp_id: location.id,
75
- GSPSampleKey.x_osgb: location.x,
76
- GSPSampleKey.y_osgb: location.y,
77
- }
78
- )
79
-
80
- if target_key == 'gsp':
81
- # Make sun coords NumpySample
82
- datetimes = pd.date_range(
83
- t0+minutes(gsp_config.interval_start_minutes),
84
- t0+minutes(gsp_config.interval_end_minutes),
85
- freq=minutes(gsp_config.time_resolution_minutes),
86
- )
87
-
88
- lon, lat = osgb_to_lon_lat(location.x, location.y)
89
-
90
- numpy_modalities.append(
91
- make_sun_position_numpy_sample(datetimes, lon, lat, key_prefix=target_key)
92
- )
93
-
94
- # Combine all the modalities and fill NaNs
95
- combined_sample = merge_dicts(numpy_modalities)
96
- combined_sample = fill_nans_in_arrays(combined_sample)
97
-
98
- return combined_sample
99
-
100
- def merge_dicts(list_of_dicts: list[dict]) -> dict:
101
- """Merge a list of dictionaries into a single dictionary"""
102
- # TODO: This doesn't account for duplicate keys, which will be overwritten
103
- combined_dict = {}
104
- for d in list_of_dicts:
105
- combined_dict.update(d)
106
- return combined_dict
107
-
108
- def fill_nans_in_arrays(sample: dict) -> dict:
109
- """Fills all NaN values in each np.ndarray in the sample dictionary with zeros.
110
-
111
- Operation is performed in-place on the sample.
112
- """
113
- for k, v in sample.items():
114
- if isinstance(v, np.ndarray) and np.issubdtype(v.dtype, np.number):
115
- if np.isnan(v).any():
116
- sample[k] = np.nan_to_num(v, copy=False, nan=0.0)
117
-
118
- # Recursion is included to reach NWP arrays in subdict
119
- elif isinstance(v, dict):
120
- fill_nans_in_arrays(v)
121
-
122
- return sample
123
-
124
-
125
- def compute(xarray_dict: dict) -> dict:
126
- """Eagerly load a nested dictionary of xarray DataArrays"""
127
- for k, v in xarray_dict.items():
128
- if isinstance(v, dict):
129
- xarray_dict[k] = compute(v)
130
- else:
131
- xarray_dict[k] = v.compute(scheduler="single-threaded")
132
- return xarray_dict
@@ -1,126 +0,0 @@
1
- import numpy as np
2
- import pandas as pd
3
- import xarray as xr
4
- import dask.array as da
5
-
6
- from ocf_data_sampler.config import load_yaml_configuration
7
- from ocf_data_sampler.select.location import Location
8
- from ocf_data_sampler.numpy_sample import NWPSampleKey, GSPSampleKey, SatelliteSampleKey
9
- from ocf_data_sampler.torch_datasets import PVNetUKRegionalDataset
10
-
11
- from ocf_data_sampler.torch_datasets.process_and_combine import (
12
- process_and_combine_datasets,
13
- merge_dicts,
14
- fill_nans_in_arrays,
15
- compute,
16
- )
17
-
18
-
19
- def test_process_and_combine_datasets(pvnet_config_filename):
20
-
21
- # Load in config for function and define location
22
- config = load_yaml_configuration(pvnet_config_filename)
23
- t0 = pd.Timestamp("2024-01-01 00:00")
24
- location = Location(coordinate_system="osgb", x=1234, y=5678, id=1)
25
-
26
- nwp_data = xr.DataArray(
27
- np.random.rand(4, 2, 2, 2),
28
- dims=["time_utc", "channel", "y", "x"],
29
- coords={
30
- "time_utc": pd.date_range("2024-01-01 00:00", periods=4, freq="h"),
31
- "channel": ["t2m", "dswrf"],
32
- "step": ("time_utc", pd.timedelta_range(start='0h', periods=4, freq='h')),
33
- "init_time_utc": pd.Timestamp("2024-01-01 00:00")
34
- }
35
- )
36
-
37
- sat_data = xr.DataArray(
38
- np.random.rand(7, 1, 2, 2),
39
- dims=["time_utc", "channel", "y", "x"],
40
- coords={
41
- "time_utc": pd.date_range("2024-01-01 00:00", periods=7, freq="5min"),
42
- "channel": ["HRV"],
43
- "x_geostationary": (["y", "x"], np.array([[1, 2], [1, 2]])),
44
- "y_geostationary": (["y", "x"], np.array([[1, 1], [2, 2]]))
45
- }
46
- )
47
-
48
- # Combine as dict
49
- dataset_dict = {
50
- "nwp": {"ukv": nwp_data},
51
- "sat": sat_data
52
- }
53
-
54
- # Call relevant function
55
- result = process_and_combine_datasets(dataset_dict, config, t0, location)
56
-
57
- # Assert result is dict - check and validate
58
- assert isinstance(result, dict)
59
- assert NWPSampleKey.nwp in result
60
- assert result[SatelliteSampleKey.satellite_actual].shape == (7, 1, 2, 2)
61
- assert result[NWPSampleKey.nwp]["ukv"][NWPSampleKey.nwp].shape == (4, 1, 2, 2)
62
-
63
-
64
- def test_merge_dicts():
65
- """Test merge_dicts function"""
66
- dict1 = {"a": 1, "b": 2}
67
- dict2 = {"c": 3, "d": 4}
68
- dict3 = {"e": 5}
69
-
70
- result = merge_dicts([dict1, dict2, dict3])
71
- assert result == {"a": 1, "b": 2, "c": 3, "d": 4, "e": 5}
72
-
73
- # Test key overwriting
74
- dict4 = {"a": 10, "f": 6}
75
- result = merge_dicts([dict1, dict4])
76
- assert result["a"] == 10
77
-
78
-
79
- def test_fill_nans_in_arrays():
80
- """Test the fill_nans_in_arrays function"""
81
- array_with_nans = np.array([1.0, np.nan, 3.0, np.nan])
82
- nested_dict = {
83
- "array1": array_with_nans,
84
- "nested": {
85
- "array2": np.array([np.nan, 2.0, np.nan, 4.0])
86
- },
87
- "string_key": "not_an_array"
88
- }
89
-
90
- result = fill_nans_in_arrays(nested_dict)
91
-
92
- assert not np.isnan(result["array1"]).any()
93
- assert np.array_equal(result["array1"], np.array([1.0, 0.0, 3.0, 0.0]))
94
- assert not np.isnan(result["nested"]["array2"]).any()
95
- assert np.array_equal(result["nested"]["array2"], np.array([0.0, 2.0, 0.0, 4.0]))
96
- assert result["string_key"] == "not_an_array"
97
-
98
-
99
- def test_compute():
100
- """Test compute function with dask array"""
101
- da_dask = xr.DataArray(da.random.random((5, 5)))
102
-
103
- # Create a nested dictionary with dask array
104
- nested_dict = {
105
- "array1": da_dask,
106
- "nested": {
107
- "array2": da_dask
108
- }
109
- }
110
-
111
- # Ensure initial data is lazy - i.e. not yet computed
112
- assert not isinstance(nested_dict["array1"].data, np.ndarray)
113
- assert not isinstance(nested_dict["nested"]["array2"].data, np.ndarray)
114
-
115
- # Call the compute function
116
- result = compute(nested_dict)
117
-
118
- # Assert that the result is an xarray DataArray and no longer lazy
119
- assert isinstance(result["array1"], xr.DataArray)
120
- assert isinstance(result["nested"]["array2"], xr.DataArray)
121
- assert isinstance(result["array1"].data, np.ndarray)
122
- assert isinstance(result["nested"]["array2"].data, np.ndarray)
123
-
124
- # Ensure there no NaN values in computed data
125
- assert not np.isnan(result["array1"].data).any()
126
- assert not np.isnan(result["nested"]["array2"].data).any()