ocamlpro/swhid

swhid is an OCaml library to work with persistent identifiers found in Software Heritage, also known as swhid

swhid

swhid is an OCaml library to work with persistent identifiers found in Software Heritage, also known as swhid. It provides:

• a parser and a printer for swhid (Swhid_core.Object module)
• functions to compute a swhid for a given object of any kind : content, directory, revision, release or snapshot (Swhid.Compute)
• functions to query the swh API in order to get an URL from which you can download a given object (Swhid.Download)

Installation

swhid can be installed with opam:

opam install swhid

If you don't have opam, you can install it following the how to install opam guide.

If you can't or don't want to use opam, consult the opam file for build instructions.

Quickstart

let id = "swh:1:rev:db21f0afdb54c16b265754ca599869fda0ca4bfc"

let url =
  match Swhid_core.Object.of_string id with
  | Error e -> Error e
  | Ok id -> Swhid.Download.revision id

let () =
  match url with
  | Error e -> Format.eprintf "can't get download URL: `%s`@." e
  | Ok url ->
    Format.printf "you can download the revision at the URL: `%s`@." url

For more, have a look at the example folder, at the documentation or at the test suite.

About

• LICENSE
• CHANGELOG

swhid's Issues

use the test suite from swh as for now we clearly don't test enough things

add microseconds support

provide a CLI tool to compute SWHIDs easily

Would provide two sub-commands: swhid compute and swhid download. The compute sub-command should also let the user specify which kind of content is being worked on (directory, cnt, directory_deep etc.).

cc @kit-ty-kate

@rjbou I remember you wrote a one-liner to allow computing the swhid of a directory by calling directory_identifier_deep. For some reason I can't find it again, could you post it here ? It could be useful to others until I write the CLI tool.

move to yojson

add a download module

it should be able to download the content of a given swhid

stop parsing if the scheme version <> 1

as we don't know what future versions will look like (reported by @progval)

doc still refers to the functorized version

https://github.com/OCamlPro/swhid/blob/master/src/compute.mli#L13

enable download tests on windows again once the SSL certs are fixed in CI

add CI and code coverage with bisect_ppx

proper error handling in the parser

you should raise an exception in the parser and catch it in Parser.ml (which is different of Menhir_parser.ml)

then, you'll change all functions in Parser to return a Result.t

add a README and other boilerplate

make orel pass as much tests as possible