20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221 | class CamelsDk(HydroDataset):
"""CAMELS_DK dataset class extending RainfallRunoff.
This class provides access to the CAMELS_DK dataset, which contains hourly
hydrological and meteorological data for various watersheds.
Attributes:
region: Geographic region identifier
download: Whether to download data automatically
ds_description: Dictionary containing dataset file paths
"""
_COL_MAP = {
"precipitation": "pcp_mm",
"temperature": "airtemp_c_mean",
"pet": "pet_mm",
"DKM_dtp": "dkm_dtp",
"DKM_eta": "aet_mm",
"DKM_wcr": "dkm_wcr",
"DKM_sdr": "dkm_sdr",
"DKM_sre": "dkm_sre",
"DKM_gwh": "dkm_gwh",
"Qdkm": "qdkm",
"DKM_irr": "dkm_irr",
"Abstraction": "abstraction",
"Qobs": "q_cms_obs",
}
# AquaFetch _static_data combines only these five (signature_obs/sim are
# excluded: they share Q_mean/Q5/Q95 column names and would collide).
# soil's first column is "Id15_model" (values match catch_id), so every
# file is read with index_col=0.
_ATTR_FILES = [
"CAMELS_DK_topography.csv",
"CAMELS_DK_climate.csv",
"CAMELS_DK_geology.csv",
"CAMELS_DK_landuse.csv",
"CAMELS_DK_soil.csv",
]
def __init__(
self, uri: str, region: Optional[str] = None, download: bool = False
) -> None:
super().__init__(uri)
self.region = region
self.download = download
if not str(uri).startswith("s3://"):
self.aqua_fetch = CAMELS_DK(uri)
def read_object_ids(self) -> np.ndarray:
uri = str(self.data_source_dir).rstrip("/")
ts_rel = "CAMELS_DK/Gauged_catchments/Gauged_catchments"
if self._is_cloud():
fs = self._make_s3fs()
names = [p.split("/")[-1] for p in fs.ls(f"{uri}/{ts_rel}".removeprefix("s3://"))]
else:
names = os.listdir(os.path.join(uri, *ts_rel.split("/")))
ids = sorted(
n.replace("CAMELS_DK_obs_based_", "").replace(".csv", "")
for n in names if n.startswith("CAMELS_DK_obs_based_")
)
return np.array(ids)
def cache_attributes_to_zarr(self) -> None:
import zarr
fs = self._make_s3fs()
uri = str(self.data_source_dir).rstrip("/")
dfs = []
for fname in self._ATTR_FILES:
path = f"{uri}/CAMELS_DK/{fname}".removeprefix("s3://")
try:
with fs.open(path) as fh:
df = pd.read_csv(fh, index_col=0)
df.index = df.index.astype(str)
dfs.append(df)
except Exception as e:
print(f" WARN {fname}: {e}")
static = pd.concat(dfs, axis=1)
static = static.loc[~static.index.duplicated(keep="first")]
stations = self.read_object_ids().tolist()
static = static.reindex(stations)
static.columns = self._clean_feature_names(list(static.columns))
static = static.rename(columns={"catch_area": "area_km2", "catch_outlet_lat": "lat"})
zarr_name = self._attributes_cache_filename.replace(".nc", ".zarr")
out, opts = self._zarr_path_and_opts(zarr_name)
n = len(stations)
root = zarr.open_group(out, mode="w", storage_options=opts, zarr_format=2)
for col in static.columns:
vals = static[col].values.astype(str) if static[col].dtype == object else static[col].values
arr = root.create_array(col, shape=(n,), chunks=(n,), dtype=vals.dtype)
arr[:] = vals
arr.attrs["_ARRAY_DIMENSIONS"] = ["basin"]
basin_arr = root.create_array("basin", shape=(n,), chunks=(n,), dtype=str)
basin_arr[:] = stations
basin_arr.attrs["_ARRAY_DIMENSIONS"] = ["basin"]
root.attrs["coordinates"] = "basin"
self._write_zarr_units(root, "static")
print(f"Attributes zarr written to: {out}")
def cache_timeseries_to_zarr(self) -> None:
import zarr
from tqdm import tqdm
fs = self._make_s3fs()
uri = str(self.data_source_dir).rstrip("/")
ts_base = f"{uri}/CAMELS_DK/Gauged_catchments/Gauged_catchments"
stations = self.read_object_ids().tolist()
all_times = pd.date_range(self.default_t_range[0], self.default_t_range[1], freq="D")
n, nt = len(stations), len(all_times)
times_ns = all_times.asi8
all_vars = list(self._COL_MAP.values())
data: dict[str, np.ndarray] = {vn: np.full((n, nt), np.nan) for vn in all_vars}
print(f"Reading {n} stations from OSS...")
for i, stn in enumerate(tqdm(stations, desc="CAMELS_DK zarr")):
path = f"{ts_base}/CAMELS_DK_obs_based_{stn}.csv".removeprefix("s3://")
try:
with fs.open(path) as fh:
df = pd.read_csv(fh, index_col="time", parse_dates=True)
df = df.reindex(all_times)
for raw_col, zarr_vn in self._COL_MAP.items():
if raw_col in df.columns:
data[zarr_vn][i] = df[raw_col].values.astype(float)
except Exception as e:
print(f" WARN {stn}: {e}")
zarr_name = self._timeseries_cache_filename.replace(".nc", ".zarr")
out, opts = self._zarr_path_and_opts(zarr_name)
root = zarr.open_group(out, mode="w", storage_options=opts, zarr_format=2)
for vn in all_vars:
arr = root.create_array(vn, shape=(n, nt), chunks=(min(n, 100), min(nt, 365)), dtype="float64")
arr[:] = data[vn]
arr.attrs["_ARRAY_DIMENSIONS"] = ["basin", "time"]
time_arr = root.create_array("time", shape=(nt,), chunks=(min(nt, 365),), dtype="int64")
time_arr[:] = times_ns
time_arr.attrs["_ARRAY_DIMENSIONS"] = ["time"]
time_arr.attrs["units"] = "nanoseconds since 1970-01-01"
time_arr.attrs["calendar"] = "proleptic_gregorian"
basin_arr = root.create_array("basin", shape=(n,), chunks=(n,), dtype=str)
basin_arr[:] = stations
basin_arr.attrs["_ARRAY_DIMENSIONS"] = ["basin"]
root.attrs["coordinates"] = "basin time"
self._write_zarr_units(root, "dynamic")
print(f"Timeseries zarr written to: {out}")
@property
def _attributes_cache_filename(self):
return "camels_dk_attributes.nc"
@property
def _timeseries_cache_filename(self):
return "camels_dk_timeseries.nc"
@property
def default_t_range(self):
return ["1989-01-02", "2023-12-31"]
# get the information of features from dataset file"Data_description.pdf"
_subclass_static_definitions = {
"p_mean": {"specific_name": "p_mean", "unit": "mm/day"},
"area": {"specific_name": "area_km2", "unit": "km^2"},
"gauge_lat": {"specific_name": "lat", "unit": "degree"},
"gauge_lon": {"specific_name": "long", "unit": "degree"},
"elev_mean": {"specific_name": "dem_mean", "unit": "m"},
"pet_mean": {"specific_name": "pet_mean", "unit": "mm/day"},
}
_dynamic_variable_mapping = {
StandardVariable.STREAMFLOW: {
"default_source": "dkmodel",
"sources": {
"dkmodel": {"specific_name": "q_cms_obs", "unit": "m^3/s"},
},
},
StandardVariable.PRECIPITATION: {
"default_source": "dmi",
"sources": {
"dmi": {"specific_name": "pcp_mm", "unit": "mm/day"},
},
},
StandardVariable.TEMPERATURE_MEAN: {
"default_source": "dmi",
"sources": {
"dmi": {"specific_name": "airtemp_c_mean", "unit": "°C"},
},
},
StandardVariable.POTENTIAL_EVAPOTRANSPIRATION: {
"default_source": "dmi",
"sources": {
"dmi": {"specific_name": "pet_mm", "unit": "mm/day"},
},
},
StandardVariable.EVAPOTRANSPIRATION: {
"default_source": "dkm_model",
"sources": {
"dkm_model": {"specific_name": "aet_mm", "unit": "mm/day"},
},
},
}
|