Fix URI encoding problems about hdt-cpp HOT 12 CLOSED

There are lots of problems with the built-in parser, including incorrect interpretation of non-ASCII and/or encoded characters. My suggested course of action would be to get rid of the built-in parser and only use SERD. SERD gives spec-compatible results, and encoding is handled fine.

Only downside: direct reading of gzipped files only works with the built-in parser; piping is not supported either.

from hdt-cpp.

Have you used SERD and -f turtle to create the HDT file? All other parsers are broken.

from hdt-cpp.

The N-Triples parser is broken (especially encoding things). SERD is in the Makefile, but might need to be switched on.

from hdt-cpp.

No experience with Raptor. Would be great if SERD supported quads as well… can't be too hard to implement.

from hdt-cpp.

Di you manage to compile the code? m1ci

from hdt-cpp.

yes, I use the hdt-cpp lib.

from hdt-cpp.

yeah, HDT has encoding issues. Are you using UTF-8?

from hdt-cpp.

yeah, HDT has encoding issues. Are you using UTF-8?

I use subject and the object of the triples are IRIs not URIs.

<http://ru.dbpedia.org/resource/Список_римско-католических_епархий_(структурный_вид)> <http://dbpedia.org/ontology/wikiPageWikiLink> <http://ru.dbpedia.org/resource/Епархия_Абаэтетубы> .

Not sure if this answers your question.

Have you used SERD and -f turtle to create the HDT file? All other parsers are broken.

Oh, I used the default n-triples. Maybe -f turtle will fix the problem. Not sure if I used SERD.

from hdt-cpp.

Have you used SERD and -f turtle to create the HDT file? All other parsers are broken.

@RubenVerborgh I'd take a stab this week at addressing the parser situation. Could you elaborate, please, on your experience and observations regarding the different parsers that the library (ostensibly) supports at present?

from hdt-cpp.

@RubenVerborgh I'd also be in favor of removing the built-in parser. Serd is a high-quality library that's the right thing to do for Turtle and N-Triples support. Perhaps the gzip support could be contributed to Serd directly, benefiting a wider base of users.

Any observations regarding the Raptor integration? As HDT moves to support quads, we'll need N-Quads format support in hdt-cpp, and it probably makes sense to keep the optional Raptor dependency around for that, unless it is significantly problematic.

from hdt-cpp.

@RubenVerborgh True, I might contribute N-Quads support to Serd just so as to avoid the complicated and tangled web of Raptor's dependencies, which runs rather deep.

from hdt-cpp.

Fixed in #31.

from hdt-cpp.