File size: 2,200 Bytes
a7335ea
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
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
---
license: cc0-1.0
pretty_name: IceCube point-source data, years 2008-2018 (FrontierPhysics mirror)
viewer: false
tags:
  - physics
  - astroparticle-physics
  - neutrino-astronomy
  - icecube
  - frontierphysics
---

# IceCube data for neutrino point-source searches, years 2008-2018 (mirror)

This is a byte-identical mirror of version 2.0 of the IceCube Collaboration's public
point-source release, as served by Harvard Dataverse in its original (as-uploaded)
file format:

- Dataverse: [doi:10.7910/DVN/VKL316](https://doi.org/10.7910/DVN/VKL316), version 2.0, CC0 1.0
- Data DOI: [10.21234/CPKQ-K003](https://doi.org/10.21234/CPKQ-K003)
- Paper: IceCube Collaboration, "IceCube Data for Neutrino Point-Source Searches: Years 2008-2018",
  [arXiv:2101.09836](https://arxiv.org/abs/2101.09836)

The FrontierPhysics task `icecube-pointsource-unbinned` downloads this file when it builds its
image, pinned to a commit of this repository, and falls back to Dataverse when the mirror is
unavailable. Either way, the build checks every file against `icecube_release.sha256`.

## Files

| File | Contents |
|---|---|
| `icecube-pointsource-2008-2018.zip` | The 36 release files: `events/`, `irfs/`, `uptime/` and `readme.txt` (409,178,788 bytes unpacked) |
| `icecube_release.sha256` | sha256 of each of the 36 files |
| `make_zip.py` | The script that built the zip |

## Rebuilding the zip

```bash
curl -L -o bundle.zip "https://dataverse.harvard.edu/api/access/dataset/:persistentId/versions/2.0?persistentId=doi:10.7910/DVN/VKL316&format=original"
mkdir release && unzip -q bundle.zip -d release
(cd release && sha256sum -c ../icecube_release.sha256)
python3 make_zip.py release icecube-pointsource-2008-2018.zip icecube_release.sha256
```

`format=original` matters: without it, Dataverse 6.12 and later serve the ingested CSV files
in their archival `.tab` form. Dataverse builds its bundle on demand, so the bundle's own
digest changes between downloads. The zip here is built from sorted entries with a fixed
timestamp, and two independent downloads gave the same zip (sha256 `01f1f73e…`).

## Citation

Please cite the IceCube release and paper above. This mirror adds no data of its own.