ayzk commited on
Commit
151fe48
·
verified ·
1 Parent(s): f948bff

Update dataset card

Browse files
Files changed (1) hide show
  1. README.md +117 -0
README.md ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ license_name: sdrbench
4
+ license_link: https://sdrbench.github.io/
5
+ pretty_name: "SDRBench: EXAFEL"
6
+ tags:
7
+ - scientific-data
8
+ - lossy-compression
9
+ - sdrbench
10
+ - hpc
11
+ - sz3
12
+ ---
13
+
14
+ # SDRBench — EXAFEL
15
+
16
+ Images from the LCLS instrument (X-ray free-electron laser); single precision, HDF5 and binary.
17
+
18
+ This repository is an **unmodified mirror** of the EXAFEL dataset from
19
+ [SDRBench](https://sdrbench.github.io/datasets.html), the Scientific Data Reduction Benchmark. The originals are hosted by
20
+ Argonne National Laboratory on Globus; every archive was unpacked and its files uploaded
21
+ byte-for-byte (sha256 verified). `metadata/` holds the original SDRBench property and template
22
+ files. The mirror is kept in sync automatically by
23
+ [szcompressor/sdrbench](https://github.com/szcompressor/sdrbench).
24
+
25
+ - **Data source:** LCLS (SLAC)
26
+ - **Total size:** 10.4 GB
27
+ - **Format:** raw little-endian binary, C order (slowest dimension first)
28
+
29
+ ## Usage
30
+
31
+ ```bash
32
+ pip install "sdrbench[sz3]" # sdrbench + pysz (SZ3)
33
+ ```
34
+
35
+ ```python
36
+ import numpy as np
37
+ import sdrbench
38
+ from pysz import sz, szConfig, szErrorBoundMode
39
+
40
+ ds = sdrbench.dataset("exafel") # default variant: 10x32x185x388
41
+ print(ds.fields)
42
+ x = ds["dark"] # numpy array, dtype <u2, shape (10, 32, 185, 388)
43
+
44
+ # compress with SZ3 (pysz) at a 1e-3 value-range-relative error bound
45
+ conf = szConfig()
46
+ conf.errorBoundMode = szErrorBoundMode.REL
47
+ conf.relErrorBound = 1e-3
48
+ compressed, ratio = sz.compress(np.ascontiguousarray(x), conf)
49
+ y, _ = sz.decompress(compressed, x.dtype.type, x.shape)
50
+ max_err, psnr, nrmse = sz.verify(np.asarray(x), y)
51
+ print(f"ratio {ratio:.1f}x, PSNR {psnr:.1f} dB, max error {max_err:.3g}")
52
+ ```
53
+
54
+ Other variants: `sdrbench.dataset("exafel", "130x1480x1552")` (all: `130x1480x1552`, `986x32x185x388`).
55
+
56
+ Files are downloaded on first use and cached; if Hugging Face is unreachable the package falls
57
+ back to the original archive on Globus.
58
+
59
+ Without the package, any file can be read with `huggingface_hub` and numpy:
60
+
61
+ ```python
62
+ from huggingface_hub import hf_hub_download
63
+ import numpy as np
64
+ p = hf_hub_download("sdrbench/exafel", "10x32x185x388/smd-cxif5315-r129-dark.u16", repo_type="dataset")
65
+ x = np.fromfile(p, dtype="<u2").reshape((10, 32, 185, 388))
66
+ ```
67
+
68
+ ## Contents
69
+
70
+
71
+ ### `10x32x185x388` (default)
72
+
73
+ 3 files, 137.8 MB, from [SDRBENCH-EXAFEL-10x32x185x388.tar.gz](https://g-d0cd3f.fd635.8443.data.globus.org/raw-data/EXAFEL/SDRBENCH-EXAFEL-10x32x185x388.tar.gz).
74
+
75
+ | Field | dtype | Shape (C order) | File |
76
+ |---|---|---|---|
77
+ | `dark` | `<u2` | 10 x 32 x 185 x 388 | `10x32x185x388/smd-cxif5315-r129-dark.u16` |
78
+ | `calib-fde` | `<u2` | 10 x 32 x 185 x 388 | `10x32x185x388/smd-cxif5315-r169-calib-fde.u16` |
79
+ | `raw` | `<u2` | 10 x 32 x 185 x 388 | `10x32x185x388/smd-cxif5315-r169-raw.u16` |
80
+
81
+ ### `130x1480x1552`
82
+
83
+ 4 files, 1.2 GB, from [SDRBENCH-EXAFEL-130x1480x1552.tar.gz](https://g-d0cd3f.fd635.8443.data.globus.org/raw-data/EXAFEL/SDRBENCH-EXAFEL-130x1480x1552.tar.gz).
84
+
85
+ | Field | dtype | Shape (C order) | File |
86
+ |---|---|---|---|
87
+ | `data` | `<f4` | 130 x 1480 x 1552 | `130x1480x1552/SDRBENCH-EXAFEL-data-130x1480x1552.f32` |
88
+ | `nPeaks` | `<i8` | 130 | `130x1480x1552/SDRBENCH-EXAFEL-nPeaks.i64` |
89
+ | `peakXPosRaw` | `<f8` | 130 x 2048 | `130x1480x1552/SDRBENCH-EXAFEL-peakXPosRaw-130x2048.d64` |
90
+ | `peakYPosRaw` | `<f8` | 130 x 2048 | `130x1480x1552/SDRBENCH-EXAFEL-peakYPosRaw-130x2048.d64` |
91
+
92
+ ### `986x32x185x388`
93
+
94
+ 1 files, 9.1 GB, from [SDRBENCH-EXAFEL-986x32x185x388.tar.gz](https://g-d0cd3f.fd635.8443.data.globus.org/raw-data/EXAFEL/SDRBENCH-EXAFEL-986x32x185x388.tar.gz).
95
+
96
+ | Field | dtype | Shape (C order) | File |
97
+ |---|---|---|---|
98
+ | `data` | `<f4` | 986 x 32 x 185 x 388 | `986x32x185x388/EXAFEL-LCLS-986x32x185x388.f32` |
99
+
100
+ ## Citation
101
+
102
+ If you use this dataset, please cite SDRBench:
103
+
104
+ ```bibtex
105
+ @inproceedings{zhao2020sdrbench,
106
+ title = {{SDRBench}: Scientific Data Reduction Benchmark for Lossy Compressors},
107
+ author = {Zhao, Kai and Di, Sheng and Liang, Xin and Li, Sihuan and Tao, Dingwen and Chen, Zizhong and Cappello, Franck},
108
+ booktitle = {2020 IEEE International Conference on Big Data (Big Data)},
109
+ pages = {2716--2724},
110
+ year = {2020}
111
+ }
112
+ ```
113
+
114
+ ## License and terms
115
+
116
+ The data are distributed under the same terms as on the [SDRBench website](https://sdrbench.github.io/datasets.html).
117
+ Rights remain with the original data providers listed above.