ba0952db817d721b12ca1e9bd91755081bc5bbac - KorAP/Krawfish-prototype

commit	ba0952db817d721b12ca1e9bd91755081bc5bbac	[log] [tgz]
author	Akron <nils@diewald-online.de>	Mon Oct 23 18:59:55 2017 +0200
committer	Akron <nils@diewald-online.de>	Mon Oct 23 18:59:55 2017 +0200
tree	1d9a41df152467700fbfa04d12144ce53cf2eb47
parent	b8aa4f795d07f8a27dd983c9d2614a609ad6757a [diff]

tree: 1d9a41df152467700fbfa04d12144ce53cf2eb47

README.md

Krawfish Prototype

The Krawfish Prototype is a testbed search backend for KorAP, to implement design concepts both for Kanalito and Krill.

Krawfish Prototype focusses on

Patterns for distribution (data aggregation, sorting, grouping ...)
Implementation of a forward index
Normalization and optimization of complex queries
Implementation of experimental features

Krawfish Prototype is still work in progress. The segment handling of Krawfish is based on Krill and therefore heavily inspired by Lucene and Lucy.

! This software is in its early stages and not stable yet! Use it on your own risk!

SETUP

Krawfish Prototype requires Perl of at least v5.10.1. The recommended environment is based on Perlbrew with App::cpanminus.

$ git clone https://github.com/KorAP/Krawfish-Prototype
$ cd Krawfish-Prototype
$ cpanm --installdeps .
$ perl Makefile.PL
$ make test

SYNOPSIS

use Krawfish::Koral;
use Krawfish::Index;

# Add documents to index
my $index = Krawfish::Index->new;
$index->introduce_field('docID' => 'de_DE');
$index->add_doc('t/data/doc1.jsonld');
$index->add_doc('t/data/doc2.jsonld');
$index->commit;

# Start KoralQuery object
my $koral = Krawfish::Koral->new;

# Define a query
# [einen|"d.*"][][Hut]
my $query = $koral->query_builder;
$koral->query(
  $query->seq(
    $query->token(
      $query->bool_or(
        'einen',
        $query->term_re('d.*')
      )
    ),
    $query->anywhere,
    $query->term('Hut')
  )
);

# Define a virtual corpus
my $corpus = $koral->corpus_builder;
$koral->corpus(
  $corpus->bool_and(
    $corpus->string('license=free'),
    $corpus->string('corpus=corpus-2')
  )
);

# Define a compilation target
my $compile = $koral->compile_builder;
$koral->compile(
  $compile->aggregate(
    $compile->a_fields('license'),
    $compile->a_frequencies
  ),
  $compile->enrich(
    $compile->e_fields('textLength')
  ),
  $compile->sort_by(
    $compile->s_field('docID')
  )
);

# Execute KoralQuery
my $request = $koral->to_query
  ->identify($index->dict)
  ->optimize($index->segment);

if ($request->next) {
  print $request->current_match->to_string;
};

COPYRIGHT AND LICENSE

Krawfish Prototype is free software published under the BSD-2 License.