Or maybe:
- DIP: Distributed Info Package
- HIP: Hash Info Package
- CHIP: Content Hash Info Package
- SHIP: Semantic Hash Info Package
Noodling on a proposal for a data package format. Needless to say, this ia a work-in-progress and no more.
package-1-example
├── .dip
│ ├── objects
│ │ ├── 12
│ │ │ └── 3e
│ │ │ └── a45d21c9025a3d0f02de4088d4cb4e7eeb6246748edd883945b430b4a05d
│ │ ├── 2b
│ │ │ └── 44
│ │ │ └── 7b5885f9b40c864ea49594a5b31fcb6e2a321c1732346e8296fe14a1dcdc
│ │ ├── 59
│ │ │ └── 9c
│ │ │ └── 521e652062bca90071b72ffc9780e97e3bcf3213630b390c0bb703594f60
│ │ ├── 71
│ │ │ └── 29
│ │ │ └── a7fac461d084afc4476e6b72156fef11b3784286cf698742fb3a6ba7e5f2
│ │ ├── 7a
│ │ │ └── 1c
│ │ │ └── 82e3a1414dd90304cff20098bd0cc94a2134f77ae703b75c1020d4f49ed4
│ │ ├── 7b
│ │ │ └── ec
│ │ │ └── d1b7e758b87c25dba4357b2bd298ba0e14dbe546ce88b108696fd93ba6c1
│ │ ├── 7d
│ │ │ └── d8
│ │ │ └── c63363ace47dc97a34eeef6042b6023a43adb8ee36e4305b3917d545648e
│ │ ├── 8b
│ │ │ └── 1d
│ │ │ └── d9de5a651508fed4c36dfb38fba591217c6fb177de0ec6f622ceb13f855c
│ │ ├── 96
│ │ │ └── 65
│ │ │ └── 15904c84bb95535c62b7865f4be638b0720f183c01ad0d2720c5f60ea2cc
│ │ ├── b2
│ │ │ └── b5
│ │ │ └── c5e08b48339246cd935a1a810c3030518a369c170d8693e59d5a5b58feab
│ │ ├── f6
│ │ │ ├── 0f
│ │ │ │ └── 3dbecc938b5fa329ef7eaf09c13612962ada8e864adcecc3e4b104f685d4
│ │ │ └── b5
│ │ │ └── c3fd3082a3c483a705c1ab7988fe690e47232063fc6a70e5cba0ab37968d
│ │ └── fb
│ │ └── 4b
│ │ └── 3be4ad5fbba57436cc5327391df21802cfc8c14d1db1ab22fb32010189b2
│ ├── refs
│ │ └── tags
│ │ ├── 65
│ │ │ └── 30
│ │ │ └── f774a5f130efbd0d0282736a21a9b92d19e48b9ca71782be30ad37c5e756
│ │ ├── 9c
│ │ │ └── 26
│ │ │ └── 4d5caeb5a7868caa70d9364514a9149aa3c764937a35378ea2906d81bf35
│ │ ├── 9f
│ │ │ └── 98
│ │ │ └── 53b3f13444197bd1c8eedb085b4c030a02ed28b5f61f3a6a1ba27bd1ae73
│ │ ├── ac
│ │ │ └── bb
│ │ │ └── b742c7524cecdec7557d60e4a19af062346309ce5731c88485c7daf48982
│ │ ├── e0
│ │ │ └── 51
│ │ │ └── 84c1f58012abeecc03933a7cc6b1e9f87b0d23de1f96993b154064ad8dca
│ │ ├── e5
│ │ │ └── 79
│ │ │ └── a69e6ffb88004b2f1bb290313704e76c156c8cf5f24ec870dd89472eca34
│ │ └── fb
│ │ └── ea
│ │ └── e7c18667b6987518f3ae61ed8b19038e5961e8e7368597428eff76e4842a
│ └── sysmeta
│ ├── 9c
│ │ └── 26
│ │ └── 4d5caeb5a7868caa70d9364514a9149aa3c764937a35378ea2906d81bf35
│ ├── 9f
│ │ └── 98
│ │ └── 53b3f13444197bd1c8eedb085b4c030a02ed28b5f61f3a6a1ba27bd1ae73
│ ├── ac
│ │ └── bb
│ │ └── b742c7524cecdec7557d60e4a19af062346309ce5731c88485c7daf48982
│ ├── e0
│ │ └── 51
│ │ └── 84c1f58012abeecc03933a7cc6b1e9f87b0d23de1f96993b154064ad8dca
│ ├── e5
│ │ └── 79
│ │ └── a69e6ffb88004b2f1bb290313704e76c156c8cf5f24ec870dd89472eca34
│ └── fb
│ └── ea
│ └── e7c18667b6987518f3ae61ed8b19038e5961e8e7368597428eff76e4842a
├── metadata
│ ├── Greenhouse_gas_flux_measurements_at_the_zero.xml
│ └── ore-resource-map.xml
├── plotobs.csv
└── raw
├── plot-orig.csv
├── site_data.csv
└── survey_data.csv
$ dip --help
dip [-s|--store file [pid [format]]] [-t|--tag cid tag] [-l|--list-tags] [-o|--list-objects] [-c|--cat cid] [-m|--meta pid] [-i|--init]All of those commands work, and put the files and objects in the right places. The major item that is missing is generating and storing the annotation files for the blobs, tree, and package structures. I simulated that for now by manually writing out example annotation files in JSON-LD and adding them to the store.
$ dip --list-tags
7becd1b7e758b87c25dba4357b2bd298ba0e14dbe546ce88b108696fd93ba6c1 doi:10.18738/XYZ
7129a7fac461d084afc4476e6b72156fef11b3784286cf698742fb3a6ba7e5f2 ORE1
123ea45d21c9025a3d0f02de4088d4cb4e7eeb6246748edd883945b430b4a05d P3
b2b5c5e08b48339246cd935a1a810c3030518a369c170d8693e59d5a5b58feab P2
fb4b3be4ad5fbba57436cc5327391df21802cfc8c14d1db1ab22fb32010189b2 P4
8b1dd9de5a651508fed4c36dfb38fba591217c6fb177de0ec6f622ceb13f855c EML1
7a1c82e3a1414dd90304cff20098bd0cc94a2134f77ae703b75c1020d4f49ed4 P1$ dip --cat 7becd1b7e758b87c25dba4357b2bd298ba0e14dbe546ce88b108696fd93ba6c1
{
"@context": {
"so": "http://schema.org/",
"sha": "hash://sha256/",
"ore": "http://www.openarchives.org/ore/terms/",
"prov": "http://www.w3.org/ns/prov#",
"ex": "https://example.com"
},
"@type": "so:Dataset",
"so:name": "example-package",
"prov:atLocation": "example-package",
"ex:contains": "sha:966515904c84bb95535c62b7865f4be638b0720f183c01ad0d2720c5f60ea2cc",
"ex:contains": "sha:f6b5c3fd3082a3c483a705c1ab7988fe690e47232063fc6a70e5cba0ab37968d",
"ex:contains": "sha:2b447b5885f9b40c864ea49594a5b31fcb6e2a321c1732346e8296fe14a1dcdc"
}$ dip --cat f6b5c3fd3082a3c483a705c1ab7988fe690e47232063fc6a70e5cba0ab37968d
{
"@context": {
"so": "http://schema.org/",
"sha": "hash://sha256/",
"ore": "http://www.openarchives.org/ore/terms/",
"prov": "http://www.w3.org/ns/prov#",
"ex": "https://example.com"
},
"@type": "ex:Folder",
"so:name": "raw",
"prov:atLocation": "raw",
"ex:contains": "sha:7dd8c63363ace47dc97a34eeef6042b6023a43adb8ee36e4305b3917d545648e",
"ex:contains": "sha:f60f3dbecc938b5fa329ef7eaf09c13612962ada8e864adcecc3e4b104f685d4",
"ex:contains": "sha:599c521e652062bca90071b72ffc9780e97e3bcf3213630b390c0bb703594f60"
}$ dip --cat 7dd8c63363ace47dc97a34eeef6042b6023a43adb8ee36e4305b3917d545648e
{
"@context": {
"so": "http://schema.org/",
"sha": "hash://sha256/",
"ore": "http://www.openarchives.org/ore/terms/",
"prov": "http://www.w3.org/ns/prov#"
},
"@id": "sha:b2b5c5e08b48339246cd935a1a810c3030518a369c170d8693e59d5a5b58feab",
"@type": "so:DataDownload",
"ore:aggregatedBy": "sha:7129a7fac461d084afc4476e6b72156fef11b3784286cf698742fb3a6ba7e5f2",
"prov:atLocation": "raw/site_data.csv"
}-
Walking a package starts top down from the PID for the package
- Find and read the
Datasetannotation CID from the PID tag - Recurse through each member of the package (referenced by CID) and read and parse it to further descend
- Find and read the
-
Building a package starts bottom up
- For each folder, starting at the deepest folder in the tree and recursing to parents
- For each of the files in the folder (note this can be fully parallelized for each file)
- Store an object by calculating the cid hash and putting it in objects dir named with cid
- If you know the PID, create a tag linking the PID to CID
- Generate or store sysmeta associated with the PID
- Create and store a file-annotation.json referencing the CID
- Create a folder-annotation.json listing all of the file and folder annotation cids that it contains
- if not at root, move to parent folder, and repeat loop (only process a parent folder after all of its child folders have been processed)
- For each of the files in the folder (note this can be fully parallelized for each file)
- Once root is reached, add datapackage-annotation.json that lists all of the file and folder annotations in the package
- Store that package annotation using its CID
- tag the package CID with its PID (e.g., a DOI)
- set the HEAD ref to the package CID
- For each folder, starting at the deepest folder in the tree and recursing to parents
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent