rename data -> waveform
Browse files- example.py +22 -9
- quakeflow_nc.py +15 -19
example.py
CHANGED
@@ -1,20 +1,33 @@
|
|
1 |
# %%
|
|
|
2 |
import numpy as np
|
3 |
-
from datasets import load_dataset
|
4 |
from torch.utils.data import DataLoader
|
5 |
|
6 |
-
|
7 |
-
quakeflow_nc
|
8 |
-
"
|
9 |
-
|
10 |
-
# name="
|
11 |
-
split="test",
|
12 |
download_mode="force_redownload",
|
|
|
|
|
13 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
14 |
|
15 |
# print the first sample of the iterable dataset
|
16 |
for example in quakeflow_nc:
|
17 |
-
print("\nIterable
|
|
|
18 |
print(example.keys())
|
19 |
for key in example.keys():
|
20 |
if key == "data":
|
@@ -28,7 +41,7 @@ quakeflow_nc = quakeflow_nc.with_format("torch")
|
|
28 |
dataloader = DataLoader(quakeflow_nc, batch_size=8, num_workers=0, collate_fn=lambda x: x)
|
29 |
|
30 |
for batch in dataloader:
|
31 |
-
print("\nDataloader
|
32 |
print(f"Batch size: {len(batch)}")
|
33 |
print(batch[0].keys())
|
34 |
for key in batch[0].keys():
|
|
|
1 |
# %%
|
2 |
+
import datasets
|
3 |
import numpy as np
|
|
|
4 |
from torch.utils.data import DataLoader
|
5 |
|
6 |
+
quakeflow_nc = datasets.load_dataset(
|
7 |
+
"AI4EPS/quakeflow_nc",
|
8 |
+
name="station",
|
9 |
+
split="train",
|
10 |
+
# name="station_test",
|
11 |
+
# split="test",
|
12 |
download_mode="force_redownload",
|
13 |
+
trust_remote_code=True,
|
14 |
+
num_proc=36,
|
15 |
)
|
16 |
+
# quakeflow_nc = datasets.load_dataset(
|
17 |
+
# "./quakeflow_nc.py",
|
18 |
+
# name="station",
|
19 |
+
# split="train",
|
20 |
+
# # name="statoin_test",
|
21 |
+
# # split="test",
|
22 |
+
# num_proc=36,
|
23 |
+
# )
|
24 |
+
|
25 |
+
print(quakeflow_nc)
|
26 |
|
27 |
# print the first sample of the iterable dataset
|
28 |
for example in quakeflow_nc:
|
29 |
+
print("\nIterable dataset\n")
|
30 |
+
print(example)
|
31 |
print(example.keys())
|
32 |
for key in example.keys():
|
33 |
if key == "data":
|
|
|
41 |
dataloader = DataLoader(quakeflow_nc, batch_size=8, num_workers=0, collate_fn=lambda x: x)
|
42 |
|
43 |
for batch in dataloader:
|
44 |
+
print("\nDataloader dataset\n")
|
45 |
print(f"Batch size: {len(batch)}")
|
46 |
print(batch[0].keys())
|
47 |
for key in batch[0].keys():
|
quakeflow_nc.py
CHANGED
@@ -167,7 +167,7 @@ class QuakeFlow_NC(datasets.GeneratorBasedBuilder):
|
|
167 |
):
|
168 |
features = datasets.Features(
|
169 |
{
|
170 |
-
"
|
171 |
"phase_time": datasets.Sequence(datasets.Value("string")),
|
172 |
"phase_index": datasets.Sequence(datasets.Value("int32")),
|
173 |
"phase_type": datasets.Sequence(datasets.Value("string")),
|
@@ -183,7 +183,7 @@ class QuakeFlow_NC(datasets.GeneratorBasedBuilder):
|
|
183 |
elif (self.config.name == "event") or (self.config.name == "event_train") or (self.config.name == "event_test"):
|
184 |
features = datasets.Features(
|
185 |
{
|
186 |
-
"
|
187 |
"phase_time": datasets.Sequence(datasets.Sequence(datasets.Value("string"))),
|
188 |
"phase_index": datasets.Sequence(datasets.Sequence(datasets.Value("int32"))),
|
189 |
"phase_type": datasets.Sequence(datasets.Sequence(datasets.Value("string"))),
|
@@ -224,19 +224,17 @@ class QuakeFlow_NC(datasets.GeneratorBasedBuilder):
|
|
224 |
# By default the archives will be extracted and a path to a cached folder where they are extracted is returned instead of the archive
|
225 |
urls = _URLS[self.config.name]
|
226 |
# files = dl_manager.download(urls)
|
227 |
-
files = dl_manager.download_and_extract(urls)
|
228 |
-
|
229 |
-
|
|
|
230 |
|
231 |
if self.config.name == "station" or self.config.name == "event":
|
232 |
return [
|
233 |
datasets.SplitGenerator(
|
234 |
name=datasets.Split.TRAIN,
|
235 |
# These kwargs will be passed to _generate_examples
|
236 |
-
gen_kwargs={
|
237 |
-
"filepath": files[:-1],
|
238 |
-
"split": "train",
|
239 |
-
},
|
240 |
),
|
241 |
datasets.SplitGenerator(
|
242 |
name=datasets.Split.TEST,
|
@@ -247,10 +245,7 @@ class QuakeFlow_NC(datasets.GeneratorBasedBuilder):
|
|
247 |
return [
|
248 |
datasets.SplitGenerator(
|
249 |
name=datasets.Split.TRAIN,
|
250 |
-
gen_kwargs={
|
251 |
-
"filepath": files,
|
252 |
-
"split": "train",
|
253 |
-
},
|
254 |
),
|
255 |
]
|
256 |
elif self.config.name == "station_test" or self.config.name == "event_test":
|
@@ -269,6 +264,7 @@ class QuakeFlow_NC(datasets.GeneratorBasedBuilder):
|
|
269 |
# The `key` is for legacy reasons (tfds) and is not important in itself, but must be unique for each example.
|
270 |
|
271 |
for file in filepath:
|
|
|
272 |
with fsspec.open(file, "rb") as fs:
|
273 |
with h5py.File(fs, "r") as fp:
|
274 |
event_ids = list(fp.keys())
|
@@ -292,10 +288,10 @@ class QuakeFlow_NC(datasets.GeneratorBasedBuilder):
|
|
292 |
or (self.config.name == "station_train")
|
293 |
or (self.config.name == "station_test")
|
294 |
):
|
295 |
-
|
296 |
|
297 |
for i, sta_id in enumerate(station_ids):
|
298 |
-
|
299 |
attrs = event[sta_id].attrs
|
300 |
phase_type = attrs["phase_type"]
|
301 |
phase_time = attrs["phase_time"]
|
@@ -304,7 +300,7 @@ class QuakeFlow_NC(datasets.GeneratorBasedBuilder):
|
|
304 |
station_location = [attrs["longitude"], attrs["latitude"], -attrs["elevation_m"] / 1e3]
|
305 |
|
306 |
yield f"{event_id}/{sta_id}", {
|
307 |
-
"
|
308 |
"phase_time": phase_time,
|
309 |
"phase_index": phase_index,
|
310 |
"phase_type": phase_type,
|
@@ -323,7 +319,7 @@ class QuakeFlow_NC(datasets.GeneratorBasedBuilder):
|
|
323 |
or (self.config.name == "event_test")
|
324 |
):
|
325 |
|
326 |
-
|
327 |
phase_type = []
|
328 |
phase_time = []
|
329 |
phase_index = []
|
@@ -331,7 +327,7 @@ class QuakeFlow_NC(datasets.GeneratorBasedBuilder):
|
|
331 |
station_location = []
|
332 |
|
333 |
for i, sta_id in enumerate(station_ids):
|
334 |
-
|
335 |
attrs = event[sta_id].attrs
|
336 |
phase_type.append(list(attrs["phase_type"]))
|
337 |
phase_time.append(list(attrs["phase_time"]))
|
@@ -341,7 +337,7 @@ class QuakeFlow_NC(datasets.GeneratorBasedBuilder):
|
|
341 |
[attrs["longitude"], attrs["latitude"], -attrs["elevation_m"] / 1e3]
|
342 |
)
|
343 |
yield event_id, {
|
344 |
-
"
|
345 |
"phase_time": phase_time,
|
346 |
"phase_index": phase_index,
|
347 |
"phase_type": phase_type,
|
|
|
167 |
):
|
168 |
features = datasets.Features(
|
169 |
{
|
170 |
+
"waveform": datasets.Array2D(shape=(3, self.nt), dtype="float32"),
|
171 |
"phase_time": datasets.Sequence(datasets.Value("string")),
|
172 |
"phase_index": datasets.Sequence(datasets.Value("int32")),
|
173 |
"phase_type": datasets.Sequence(datasets.Value("string")),
|
|
|
183 |
elif (self.config.name == "event") or (self.config.name == "event_train") or (self.config.name == "event_test"):
|
184 |
features = datasets.Features(
|
185 |
{
|
186 |
+
"waveform": datasets.Array3D(shape=(None, 3, self.nt), dtype="float32"),
|
187 |
"phase_time": datasets.Sequence(datasets.Sequence(datasets.Value("string"))),
|
188 |
"phase_index": datasets.Sequence(datasets.Sequence(datasets.Value("int32"))),
|
189 |
"phase_type": datasets.Sequence(datasets.Sequence(datasets.Value("string"))),
|
|
|
224 |
# By default the archives will be extracted and a path to a cached folder where they are extracted is returned instead of the archive
|
225 |
urls = _URLS[self.config.name]
|
226 |
# files = dl_manager.download(urls)
|
227 |
+
# files = dl_manager.download_and_extract(urls)
|
228 |
+
files = [f"waveform_h5/{x}" for x in _FILES]
|
229 |
+
for file in sorted(files):
|
230 |
+
print(file)
|
231 |
|
232 |
if self.config.name == "station" or self.config.name == "event":
|
233 |
return [
|
234 |
datasets.SplitGenerator(
|
235 |
name=datasets.Split.TRAIN,
|
236 |
# These kwargs will be passed to _generate_examples
|
237 |
+
gen_kwargs={"filepath": files[:-1], "split": "train"},
|
|
|
|
|
|
|
238 |
),
|
239 |
datasets.SplitGenerator(
|
240 |
name=datasets.Split.TEST,
|
|
|
245 |
return [
|
246 |
datasets.SplitGenerator(
|
247 |
name=datasets.Split.TRAIN,
|
248 |
+
gen_kwargs={"filepath": files, "split": "train"},
|
|
|
|
|
|
|
249 |
),
|
250 |
]
|
251 |
elif self.config.name == "station_test" or self.config.name == "event_test":
|
|
|
264 |
# The `key` is for legacy reasons (tfds) and is not important in itself, but must be unique for each example.
|
265 |
|
266 |
for file in filepath:
|
267 |
+
print(f"\nReading {file}")
|
268 |
with fsspec.open(file, "rb") as fs:
|
269 |
with h5py.File(fs, "r") as fp:
|
270 |
event_ids = list(fp.keys())
|
|
|
288 |
or (self.config.name == "station_train")
|
289 |
or (self.config.name == "station_test")
|
290 |
):
|
291 |
+
waveform = np.zeros([3, self.nt], dtype="float32")
|
292 |
|
293 |
for i, sta_id in enumerate(station_ids):
|
294 |
+
waveform[:, : self.nt] = event[sta_id][:, : self.nt]
|
295 |
attrs = event[sta_id].attrs
|
296 |
phase_type = attrs["phase_type"]
|
297 |
phase_time = attrs["phase_time"]
|
|
|
300 |
station_location = [attrs["longitude"], attrs["latitude"], -attrs["elevation_m"] / 1e3]
|
301 |
|
302 |
yield f"{event_id}/{sta_id}", {
|
303 |
+
"waveform": waveform,
|
304 |
"phase_time": phase_time,
|
305 |
"phase_index": phase_index,
|
306 |
"phase_type": phase_type,
|
|
|
319 |
or (self.config.name == "event_test")
|
320 |
):
|
321 |
|
322 |
+
waveform = np.zeros([len(station_ids), 3, self.nt], dtype="float32")
|
323 |
phase_type = []
|
324 |
phase_time = []
|
325 |
phase_index = []
|
|
|
327 |
station_location = []
|
328 |
|
329 |
for i, sta_id in enumerate(station_ids):
|
330 |
+
waveform[i, :, : self.nt] = event[sta_id][:, : self.nt]
|
331 |
attrs = event[sta_id].attrs
|
332 |
phase_type.append(list(attrs["phase_type"]))
|
333 |
phase_time.append(list(attrs["phase_time"]))
|
|
|
337 |
[attrs["longitude"], attrs["latitude"], -attrs["elevation_m"] / 1e3]
|
338 |
)
|
339 |
yield event_id, {
|
340 |
+
"waveform": waveform,
|
341 |
"phase_time": phase_time,
|
342 |
"phase_index": phase_index,
|
343 |
"phase_type": phase_type,
|