Force parameter on non-OAI 2 compliant identifiers

[netsensei]

I'm confronted with an OAI-PMH endpoint that publishes records that don't have an OAI 2 compliant identifier:

$ catmandu convert OAI --url http://example.org/oai --metadataPrefix oai_dc --getRecord 1 --identifier "foobar:20000099"

Will return me this:

Invalid Request (use 'force' to force a non-conformant request): <?xml version="1.0" encoding="UTF-8"?>
<error xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" code="badArgument">Bad argument (identifier): Identifier not in OAI format, must match regexp ^[[:alpha:]][[:alnum:]\+\-\.]*:.+</error>
 (stopped after 0 retries)

So, Catmandu hands me an error passed on from HTTP::OAI. This library has a force flag which allows one to skip the validate_request subroutine that validates the identifier.

It wasn't implemented in Catmandu::OAI so it's not possible to use this flag from the command line like this:

$ catmandu convert OAI --url http://example.org/oai --metadataPrefix oai_dc --getRecord 1 --identifier "foobar:20000099" --force 1

I'm going to file a PR with a proposed addition that makes this flag available. Of course, there are good design reasons to not implement this in the CLI of Catmandu.

The use case here is harvesters being confronted with OAI-PMH endpoints that are managed by third parties that don't comply to the strict specification nor have any intention to do so in the near future (i.e. due to strategic or operational circumstances that hamper a better implementation on short notice).

[phochste]

@netsensei I've changed your pull request to a --strict option and have force by default and switching it off optionally. The intent of Catmandu is to be quite liberal what to accept as data

[netsensei]

@phochste Looking good! You're right about the acceptance of data per Postel's Law.

I'm okay with merging this and pushing a new release.

[phochste]

@netsensei Version 0.19 is on its way to CPAN