Information Service URL pattern

[zazi]

The mapping from information service specific page URLs to resource URIs requires a stable URL pattern, i.e. URLs probably have to be cleaned up (e.g. www vs. non-www, country-specific top-level domains, etc.) before they can be transformed into a resource URI.

[BarryNorton]

As extension of comment in issues/1 I'd include all accepted forms as disjuncts and/or use more sophisticated regexs, e.g.: d2rq:condition "musicbrainz.url.url LIKE 'http://%myspace.com%'"

It's important that completely non-compliant URLs (like, sorry, "http://pissfork.net") should be kicked out before calling the translator.

[zazi]

I would prefer to start a clean-up task from the information-service-specific Translator class, since such cleanup tasks can be applied on all information-service-specific base URLs (so we only need to call a cleanUp method with the base URL as parameter). Every non-valid URL that is left, will be excluded from the URL (db value) to URI transformation task, since it won't fit into the given transformation pattern (e.g. original base URI to linked data base URI).

[BarryNorton]

But doesn't the Translator class have to produce a result?

Or is there an exception mechanism that will let the query as a whole proceed?

There's simply nothing that can be done with http://pissfork.net

[zazi]

Yes, I thought that I could excluded non-valid URLs via return 'null' from the translation method. However, this doesn't seem to work. So we have to define another validation condition for the mapping as a URI regex.

[zazi]

It seems that I got it to work (with the help of returning 'null' from the translation method). (the previous error was caused from my clean-up method)

Please check (with fresh D2RS-LinkedBrainz jar, please).

[zazi]

Rechecked? (I close this issue for now). The clean-up method for the base URIs can be found at https://github.com/zazi/D2RS-LinkedBrainz/blob/master/D2RS-LinkedBrainz/src/main/java/translators/util/Utils.java ;)