jacobbieker commited on
Commit
4c9706a
1 Parent(s): db8068a

Add raw observation Features

Browse files
Files changed (1) hide show
  1. gfs-reforecast.py +20 -13
gfs-reforecast.py CHANGED
@@ -73,10 +73,10 @@ class GFEReforecastDataset(datasets.GeneratorBasedBuilder):
73
  DEFAULT_CONFIG_NAME = "gfs_v16" # It's not mandatory to have a default configuration. Just use one if it make sense.
74
 
75
  def _info(self):
 
76
  if "v16" in self.config.name:
77
  # TODO Add the variables one with all 696 variables, potentially combined by level
78
- features = datasets.Features(
79
- {
80
  "current_state": datasets.Array3D((721,1440,696), dtype="float32"),
81
  "next_state": datasets.Array3D((721,1440,696), dtype="float32"),
82
  "timestamp": datasets.Sequence(datasets.Value("timestamp[ns]")),
@@ -84,20 +84,27 @@ class GFEReforecastDataset(datasets.GeneratorBasedBuilder):
84
  "longitude": datasets.Sequence(datasets.Value("float32"))
85
  # These are the features of your dataset like images, labels ...
86
  }
87
- )
88
- else:
89
- features = datasets.Features(
90
- {
91
- "precipitation_rate": datasets.Array3D((721,1440,696), dtype="float32"),
92
- "timestamp": datasets.Value("timestamp[ns]"),
93
  "latitude": datasets.Sequence(datasets.Value("float32")),
94
  "longitude": datasets.Sequence(datasets.Value("float32"))
95
  # These are the features of your dataset like images, labels ...
96
  }
97
- )
98
  if "raw" in self.config.name:
99
- # Add the raw observation features
100
- raw_features = {}
 
 
 
 
 
 
 
 
101
  return datasets.DatasetInfo(
102
  # This is the description that will appear on the datasets page.
103
  description=_DESCRIPTION,
@@ -165,8 +172,8 @@ class GFEReforecastDataset(datasets.GeneratorBasedBuilder):
165
  for t in range(len(dataset["time"].values)-1):
166
  data_t = dataset.isel(time=t)
167
  data_t1 = dataset.isel(time=(t+1))
168
- value = {"current_state": np.concatenate([data_t[v].values for v in sorted(data_t.data_vars)], axis=0),
169
- "next_state": np.concatenate([data_t1[v].values for v in sorted(data_t.data_vars)], axis=0),
170
  "timestamp": data_t["time"].values,
171
  "latitude": data_t["latitude"].values,
172
  "longitude": data_t["longitude"].values}
 
73
  DEFAULT_CONFIG_NAME = "gfs_v16" # It's not mandatory to have a default configuration. Just use one if it make sense.
74
 
75
  def _info(self):
76
+ features = {}
77
  if "v16" in self.config.name:
78
  # TODO Add the variables one with all 696 variables, potentially combined by level
79
+ features = {
 
80
  "current_state": datasets.Array3D((721,1440,696), dtype="float32"),
81
  "next_state": datasets.Array3D((721,1440,696), dtype="float32"),
82
  "timestamp": datasets.Sequence(datasets.Value("timestamp[ns]")),
 
84
  "longitude": datasets.Sequence(datasets.Value("float32"))
85
  # These are the features of your dataset like images, labels ...
86
  }
87
+ elif "analysis" in self.config.name:
88
+ # TODO Add the variables one with all 322 variables, potentially combined by level
89
+ features = {
90
+ "current_state": datasets.Array3D((721,1440,322), dtype="float32"),
91
+ "next_state": datasets.Array3D((721,1440,322), dtype="float32"),
92
+ "timestamp": datasets.Sequence(datasets.Value("timestamp[ns]")),
93
  "latitude": datasets.Sequence(datasets.Value("float32")),
94
  "longitude": datasets.Sequence(datasets.Value("float32"))
95
  # These are the features of your dataset like images, labels ...
96
  }
 
97
  if "raw" in self.config.name:
98
+ # Add the raw observation features, capping at 256,000 observations, padding if not enough
99
+ raw_features = {"observations": datasets.Array2D((256000,1), dtype="float32"),
100
+ "observation_type": datasets.Array2D((256000,1), dtype="string"),
101
+ "observation_lat": datasets.Array2D((256000,1), dtype="float32"),
102
+ "observation_lon": datasets.Array2D((256000,1), dtype="float32"),
103
+ }
104
+ features = features.update(raw_features)
105
+
106
+ features = datasets.Features(features)
107
+
108
  return datasets.DatasetInfo(
109
  # This is the description that will appear on the datasets page.
110
  description=_DESCRIPTION,
 
172
  for t in range(len(dataset["time"].values)-1):
173
  data_t = dataset.isel(time=t)
174
  data_t1 = dataset.isel(time=(t+1))
175
+ value = {"current_state": np.stack([data_t[v].values for v in sorted(data_t.data_vars)], axis=2),
176
+ "next_state": np.stack([data_t1[v].values for v in sorted(data_t.data_vars)], axis=2),
177
  "timestamp": data_t["time"].values,
178
  "latitude": data_t["latitude"].values,
179
  "longitude": data_t["longitude"].values}