issues to get all and clean ceur dataset

[liyakun]

• There are still some irrelevant triples. Only triples with a subject
  below http://ceur-ws.org/ should end up in our dataset. If we need to
  define CEUR-specific vocabulary terms, then we should define them in the
  http://ceur-ws.org/vocab/ namespace.
• Some persons, e.g.
  http://ceur-ws.org/resource/person/M.%20Daud%20Ahmed, are owl:sameAs a
  lot of other persons with diffferent names. Something must be wrong
  here. To see the details, try the following query:

 SELECT ?s (COUNT (?o) as ?count) {?s
<http://www.w3.org/2002/07/owl#sameAs> ?o} GROUP BY ?s HAVING (?count > 10)

• The only license that we have is
  http://choosealicense.com/licenses/no-license/. However, those volumes
  that are licensed under CC0 should also be annotated with the right
  license.
• For papers in PDF format the URI generation works fine, e.g.
  paper1.pdf translates to http://ceur-ws.org/Vol-XXX/#paper1. Please
  adapt it in the same way for papers in other formats, e.g. paper1.ps.
• Apparently the extraction tool does not apply normalisation to
  strings. Look at the dc:title of
  http://ceur-ws.org/Vol-100/#Aditya_Kalyanpur-et-al: It contains a lot
  of whitespace, which comes straight from the HTML. This whitespace has
  no meaning, it's just an artifact of how the source code was written. I
  think that in most string literals it's safe to replace any sequence of
  spaces by just one space.
• As of RDF 1.1 the default datatype for a literal is xsd:string, so
  the rdf:datatype attribute could be omitted in this case, to save some
  space.
• Some workshops (at least http://ceur-ws.org/Vol-1513/) do not have
  any papers, which is wrong.
• Some non-Latin characters, e.g., in names, are output using combining
  diacritical marks. Please check e.g. Wikipedia if you don't know what
  this is. Output should be in NFC. Which means, e.g., that ñ should be
  output as "the single UTF-8 character 'ñ'" instead of "a combining '~'
  character applied to an 'n'".
• Papers seem to have a bibo:numPages property but not start and end
  page. (It should!)

[KMax]

Hi @liyakun! Thank you for your comments! You're right, there are a lot of issues which should be fixed to improve the crawler. If you have time and desire, your help is welcome!

This code of the repo were used to participate in the challenge last year, but if you want to improve the code and participate in the next challenge, feel free to do that. We only would like to ask you to cite our last year paper if you're going to reuse our work. Thanks!