remove lim dataset
3 changed files+4−68
benchcompress/src/benchcompress/datasets/seismic/__init__.pymodified+0−47View file
@@ -65,35 +65,6 @@ def _load_quantized_04A_04B_seismic_data():
6565 return X
6666
6767
68-def _load_lim_2024_seismic_data() -> np.ndarray:
69- """Load seismic data from the 2022 Goesan earthquake.
70-
71- Returns:
72- Array containing the loaded seismic data
73- """
74- file_path = "lim_et_al_2024.01.concat.npy"
75- if not os.path.exists(file_path):
76- # Download the numpy array file
77- url = "https://zenodo.org/records/14774624/files/lim_et_al_2024.01.concat.npy?download=1"
78- response = requests.get(url)
79- with open(file_path, "wb") as f:
80- f.write(response.content)
81- print(f"Downloaded {file_path}")
82- else:
83- print(f"{file_path} already exists locally.")
84-
85- # Load the numpy array
86- X = np.load(file_path)
87- return X
88-
89-
90-def _load_quantized_lim_2024_seismic_data():
91- X = _load_lim_2024_seismic_data()
92- step = 10
93- X = np.round(X / step).astype(np.int32)
94- return X
95-
96-
9768 datasets = [
9869 {
9970 "name": "seismic-04A-04B",
@@ -113,22 +84,4 @@ datasets = [
11384 "source_file": SOURCE_FILE,
11485 "long_description": LONG_DESCRIPTION,
11586 },
116- {
117- "name": "seismic-lim-2024-01",
118- "version": "1",
119- "description": "Seismic data from the 2022 Mw 3.8 Goesan earthquake in South Korea.",
120- "create": lambda: _load_lim_2024_seismic_data(),
121- "tags": tags_float,
122- "source_file": SOURCE_FILE,
123- "long_description": LONG_DESCRIPTION,
124- },
125- {
126- "name": "seismic-lim-2024-01-quantized",
127- "version": "3",
128- "description": "Seismic data from the 2022 Mw 3.8 Goesan earthquake in South Korea, quantized.",
129- "create": lambda: _load_quantized_lim_2024_seismic_data(),
130- "tags": tags_integer,
131- "source_file": SOURCE_FILE,
132- "long_description": LONG_DESCRIPTION,
133- },
13487 ]
benchcompress/src/benchcompress/datasets/seismic/seismic.mdmodified+0−19View file
@@ -6,10 +6,6 @@ This dataset contains marine seismic reflection data. Processed seismic reflecti
66
77 The data comes from a SEG-Y file available on Zenodo (https://zenodo.org/records/8152964). SEG-Y is a standard format for storing seismic data that includes both the recorded waveforms and metadata about the survey.
88
9-## Variants
10-
11-We provide two datasets with two versions each:
12-
139 ### Revelle RR1508 Data (seismic-04A-04B)
1410 - Original floating point values from the SEG-Y file
1511 - Contains the natural amplitude variations of the seismic waves
@@ -20,22 +16,7 @@ We provide two datasets with two versions each:
2016 - Uses a quantization step of 10000
2117 - Stored as 32-bit integers
2218
23-### Goesan Earthquake Data (seismic-lim-2024-01)
24-- Original floating point values from the continuous seismic recording
25-- Data from the 2022 Mw 3.8 earthquake in Goesan, South Korea
26-- Part of a study analyzing 42 earthquakes including foreshocks and aftershocks
27-- Records from permanent seismic networks with closest station at 8.3 km from epicenter
28-- Stored as 32-bit floating point numbers
29-
30-### Goesan Earthquake Quantized Data (seismic-lim-2024-01-quantized)
31-- Values are scaled and rounded to integers
32-- Uses a quantization step of 10000
33-- Stored as 32-bit integers
34-
3519 ## Source Details
3620
3721 ### Roger Revelle RR1508
3822 The data comes from a SEG-Y file available on Zenodo (https://zenodo.org/records/8152964). SEG-Y is a standard format for storing seismic data that includes both the recorded waveforms and metadata about the survey.
39-
40-### Goesan Earthquake 2022
41-This dataset contains seismic recordings from the 2022 Mw 3.8 Goesan earthquake in South Korea. The earthquake occurred on October 28, 2022, and was preceded by a Mw 3.3 foreshock 17 seconds before the mainshock. The study analyzed 42 earthquakes in total, including the mainshock, foreshock, and aftershocks, to understand the interactions between seismic events. The data revealed that the mainshock occurred at the southeastern tip of the hypocenter distribution of three foreshocks, with aftershocks showing a diffused pattern propagating toward both ends of the inferred lineament. The data is available on Zenodo (https://zenodo.org/records/14774624).
web-ui/src/hooks/TimeseriesDataClient.tsmodified+4−2View file
@@ -3,13 +3,14 @@ export type SupportedTypedArray =
33 | Uint16Array
44 | Uint32Array
55 | Int16Array
6- | Int32Array;
6+ | Int32Array
7+ | Float32Array;
78
89 interface ChunkCache {
910 [key: number]: SupportedTypedArray;
1011 }
1112
12-type DType = "uint8" | "uint16" | "uint32" | "int16" | "int32";
13+type DType = "uint8" | "uint16" | "uint32" | "int16" | "int32" | "float32";
1314
1415 const TypedArrayConstructors = {
1516 uint8: Uint8Array,
@@ -17,6 +18,7 @@ const TypedArrayConstructors = {
1718 uint32: Uint32Array,
1819 int16: Int16Array,
1920 int32: Int32Array,
21+ float32: Float32Array,
2022 } as const;
2123
2224 export class TimeseriesDataClient {