Skip to content

Commit 52b0070

Browse files
heptaflarheptaflar
committed
Update: total apsorption coefficient
1 parent 1aea457 commit 52b0070

File tree

2 files changed

+173
-71
lines changed

2 files changed

+173
-71
lines changed

‎kaleidoscope/config/config.random.json‎

Lines changed: 88 additions & 14 deletions
Original file line numberDiff line numberDiff line change
@@ -38,72 +38,146 @@
3838
"adg_412": {
3939
"bias": "adg_412_bias",
4040
"rmsd": "adg_412_rmsd",
41-
"distribution": "lognormal"
41+
"distribution": "lognormal",
42+
"clip": [
43+
0.0,
44+
10.0
45+
]
4246
},
4347
"adg_443": {
4448
"bias": "adg_443_bias",
4549
"rmsd": "adg_443_rmsd",
46-
"distribution": "lognormal"
50+
"distribution": "lognormal",
51+
"clip": [
52+
0.0,
53+
10.0
54+
]
4755
},
4856
"adg_490": {
4957
"bias": "adg_490_bias",
5058
"rmsd": "adg_490_rmsd",
51-
"distribution": "lognormal"
59+
"distribution": "lognormal",
60+
"clip": [
61+
0.0,
62+
10.0
63+
]
5264
},
5365
"adg_510": {
5466
"bias": "adg_510_bias",
5567
"rmsd": "adg_510_rmsd",
56-
"distribution": "lognormal"
68+
"distribution": "lognormal",
69+
"clip": [
70+
0.0,
71+
10.0
72+
]
5773
},
5874
"adg_560": {
5975
"bias": "adg_560_bias",
6076
"rmsd": "adg_560_rmsd",
61-
"distribution": "lognormal"
77+
"distribution": "lognormal",
78+
"clip": [
79+
0.0,
80+
10.0
81+
]
6282
},
6383
"adg_665": {
6484
"bias": "adg_665_bias",
6585
"rmsd": "adg_665_rmsd",
66-
"distribution": "lognormal"
86+
"distribution": "lognormal",
87+
"clip": [
88+
0.0,
89+
10.0
90+
]
6791
},
6892
"aph_412": {
6993
"bias": "aph_412_bias",
7094
"rmsd": "aph_412_rmsd",
71-
"distribution": "lognormal"
95+
"distribution": "lognormal",
96+
"clip": [
97+
0.0,
98+
10.0
99+
]
72100
},
73101
"aph_443": {
74102
"bias": "aph_443_bias",
75103
"rmsd": "aph_443_rmsd",
76-
"distribution": "lognormal"
104+
"distribution": "lognormal",
105+
"clip": [
106+
0.0,
107+
10.0
108+
]
77109
},
78110
"aph_490": {
79111
"bias": "aph_490_bias",
80112
"rmsd": "aph_490_rmsd",
81-
"distribution": "lognormal"
113+
"distribution": "lognormal",
114+
"clip": [
115+
0.0,
116+
10.0
117+
]
82118
},
83119
"aph_510": {
84120
"bias": "aph_510_bias",
85121
"rmsd": "aph_510_rmsd",
86-
"distribution": "lognormal"
122+
"distribution": "lognormal",
123+
"clip": [
124+
0.0,
125+
10.0
126+
]
87127
},
88128
"aph_560": {
89129
"bias": "aph_560_bias",
90130
"rmsd": "aph_560_rmsd",
91-
"distribution": "lognormal"
131+
"distribution": "lognormal",
132+
"clip": [
133+
0.0,
134+
10.0
135+
]
92136
},
93137
"aph_665": {
94138
"bias": "aph_665_bias",
95139
"rmsd": "aph_665_rmsd",
96-
"distribution": "lognormal"
140+
"distribution": "lognormal",
141+
"clip": [
142+
0.0,
143+
10.0
144+
]
145+
},
146+
"atot_412": {
147+
"total": ["aph_412", "adg_412"]
148+
},
149+
"atot_443": {
150+
"total": ["aph_443", "adg_443"]
151+
},
152+
"atot_490": {
153+
"total": ["aph_490", "adg_490"]
154+
},
155+
"atot_510": {
156+
"total": ["aph_510", "adg_510"]
157+
},
158+
"atot_560": {
159+
"total": ["aph_560", "adg_560"]
160+
},
161+
"atot_665": {
162+
"total": ["aph_665", "adg_665"]
97163
},
98164
"kd_490": {
99165
"bias": "kd_490_bias",
100166
"rmsd": "kd_490_rmsd",
101-
"distribution": "lognormal"
167+
"distribution": "lognormal",
168+
"clip": [
169+
0.0,
170+
10.0
171+
]
102172
},
103173
"chlor_a": {
104174
"bias": "chlor_a_log10_bias",
105175
"rmsd": "chlor_a_log10_rmsd",
106-
"distribution": "chlorophyll"
176+
"distribution": "chlorophyll",
177+
"clip": [
178+
0.0,
179+
100.0
180+
]
107181
}
108182
},
109183
"esa-scope-exchange": {

‎kaleidoscope/operators/randomizeop.py‎

Lines changed: 85 additions & 57 deletions
Original file line numberDiff line numberDiff line change
@@ -89,77 +89,25 @@ def run(self, source: Dataset) -> Dataset: # noqa: D102
8989
:param source: The source dataset.
9090
:return: The result dataset.
9191
"""
92+
config: dict[str : dict[str:Any]] = self.config.get(
93+
self._args.source_type, {}
94+
)
9295
target: Dataset = Dataset(
9396
data_vars=source.data_vars,
9497
coords=source.coords,
9598
attrs=source.attrs,
9699
)
97-
config: dict[str : dict[str:Any]] = self.config.get(
98-
self._args.source_type, {}
99-
)
100100
for v, x in target.data_vars.items():
101101
if v not in config or self._args.selector == 0:
102102
continue
103103
get_logger().info(f"starting graph for variable: {v}")
104-
s: list[int] = self.entropy(v, self.uuid)
105-
a: dict[str:Any] = config[v]
106-
f = Randomize(m=x.ndim, dist=a["distribution"], entropy=s)
107-
if "uncertainty" in a:
108-
u = (
109-
target[a["uncertainty"]]
110-
if isinstance(a["uncertainty"], str)
111-
else DataArray(
112-
data=da.full(
113-
x.shape, a["uncertainty"], chunks=x.chunks
114-
),
115-
coords=x.coords,
116-
dims=x.dims,
117-
attrs={},
118-
)
119-
)
120-
z = f.apply_to(
121-
_decode(x.data, x.attrs),
122-
_decode(u.data, u.attrs),
123-
coverage=a.get("coverage", 1.0),
124-
relative=a.get("relative", False),
125-
clip=a.get("clip", None),
126-
)
127-
else:
128-
b = target[a["bias"]]
129-
r = target[a["rmsd"]]
130-
z = f.apply_to(
131-
_decode(x.data, x.attrs),
132-
_decode(r.data, r.attrs),
133-
_decode(b.data, b.attrs),
134-
clip=a.get("clip", None),
135-
)
136-
target[v] = DataArray(
137-
data=_encode(z, x.attrs, x.dtype),
138-
coords=x.coords,
139-
dims=x.dims,
140-
attrs=x.attrs,
141-
)
142-
if "actual_range" in target[v].attrs:
143-
target[v].attrs["actual_range"] = np.array(
144-
[
145-
da.nanmin(z).compute(),
146-
da.nanmax(z).compute(),
147-
],
148-
dtype=z.dtype,
149-
)
150-
target[v].attrs["entropy"] = np.array(s, dtype=np.int64)
151-
if get_logger().is_enabled(Logging.DEBUG):
152-
get_logger().debug(f"entropy: {s}")
153-
get_logger().debug(f"min: {da.nanmin(z).compute() :.3f}")
154-
get_logger().debug(f"max: {da.nanmax(z).compute() :.3f}")
155-
get_logger().debug(f"mean: {da.nanmean(z).compute() :.3f}")
156-
get_logger().debug(f"std: {da.nanstd(z).compute() :.3f}")
104+
self.randomize(source, target, v, x, config[v])
157105
get_logger().info(f"finished graph for variable: {v}")
158106
return target
159107

160108
@property
161109
def config(self) -> dict[str : dict[str:Any]]:
162-
"""Returns the product type configuration."""
110+
"""Returns the randomization configuration."""
163111
package = "kaleidoscope.config"
164112
name = "config.random.json"
165113
with resources.path(package, name) as resource:
@@ -187,6 +135,86 @@ def entropy(self, name: str, uuid: str, n: int = 4) -> list[int]:
187135
g = DefaultGenerator(Philox(seed))
188136
return [g.next() for _ in range(n)]
189137

138+
def randomize(
139+
self,
140+
source: Dataset,
141+
target: Dataset,
142+
v: str,
143+
x: DataArray,
144+
config: dict[str:Any],
145+
):
146+
"""
147+
Creates the graph to randomize a variable.
148+
149+
:param source: The source dataset.
150+
:param target: The target dataset.
151+
:param v: The name of the variable.
152+
:param x: The data of the variable.
153+
:param config: The randomization configuration.
154+
"""
155+
if "total" in config:
156+
s: list[int] = []
157+
z = _decode(x.data, x.attrs)
158+
for ref in config["total"]:
159+
a = _decode(target[ref].data, target[ref].attrs)
160+
b = _decode(source[ref].data, source[ref].attrs)
161+
z = z + (a - b)
162+
elif "uncertainty" in config:
163+
s: list[int] = self.entropy(v, self.uuid)
164+
f = Randomize(m=x.ndim, dist=config["distribution"], entropy=s)
165+
u = (
166+
target[config["uncertainty"]]
167+
if isinstance(config["uncertainty"], str)
168+
else DataArray(
169+
data=da.full(
170+
x.shape, config["uncertainty"], chunks=x.chunks
171+
),
172+
coords=x.coords,
173+
dims=x.dims,
174+
attrs={},
175+
)
176+
)
177+
z = f.apply_to(
178+
_decode(x.data, x.attrs),
179+
_decode(u.data, u.attrs),
180+
coverage=config.get("coverage", 1.0),
181+
relative=config.get("relative", False),
182+
clip=config.get("clip", None),
183+
)
184+
else:
185+
s: list[int] = self.entropy(v, self.uuid)
186+
f = Randomize(m=x.ndim, dist=config["distribution"], entropy=s)
187+
b = target[config["bias"]]
188+
r = target[config["rmsd"]]
189+
z = f.apply_to(
190+
_decode(x.data, x.attrs),
191+
_decode(r.data, r.attrs),
192+
_decode(b.data, b.attrs),
193+
clip=config.get("clip", None),
194+
)
195+
target[v] = DataArray(
196+
data=_encode(z, x.attrs, x.dtype),
197+
coords=x.coords,
198+
dims=x.dims,
199+
attrs=x.attrs,
200+
)
201+
if "actual_range" in target[v].attrs:
202+
target[v].attrs["actual_range"] = np.array(
203+
[
204+
da.nanmin(z).compute(),
205+
da.nanmax(z).compute(),
206+
],
207+
dtype=z.dtype,
208+
)
209+
if s:
210+
target[v].attrs["entropy"] = np.array(s, dtype=np.int64)
211+
if get_logger().is_enabled(Logging.DEBUG):
212+
get_logger().debug(f"entropy: {s}")
213+
get_logger().debug(f"min: {da.nanmin(z).compute() :.3f}")
214+
get_logger().debug(f"max: {da.nanmax(z).compute() :.3f}")
215+
get_logger().debug(f"mean: {da.nanmean(z).compute() :.3f}")
216+
get_logger().debug(f"std: {da.nanstd(z).compute() :.3f}")
217+
190218
@property
191219
def uuid(self) -> str:
192220
"""

0 commit comments

Comments
 (0)