NAME
    Lingua::StanfordCoreNLP - A Perl interface to Stanford's CoreNLP tool
    set.

SYNOPSIS
     # Note that Lingua::StanfordCoreNLP can't be instantiated.
     use Lingua::StanfordCoreNLP;

     # Create a new NLP pipeline (make corefs bidirectional)
     my $pipeline = new Lingua::StanfordCoreNLP::Pipeline(1);

     # Get annotator properties:
     my $props = $pipeline->getProperties();

     # These are the default annotator properties:
     $props->put('annotators', 'tokenize, ssplit, pos, lemma, ner, parse, dcoref');

     # Update properties:
     $pipeline->setProperties($props);

     # Process text
     # (Will output lots of debug info from the Java classes to STDERR.)
     my $result = $pipeline->process(
        'Jane looked at the IBM computer. She turned it off.'
     );

     my @seen_corefs;

     # Print results
     for my $sentence (@{$result->toArray}) {
        print "\n[Sentence ID: ", $sentence->getIDString, "]:\n";
        print "Original sentence:\n\t", $sentence->getSentence, "\n";

        print "Tagged text:\n";
        for my $token (@{$sentence->getTokens->toArray}) {
           printf "\t%s/%s/%s [%s]\n",
                  $token->getWord,
                  $token->getPOSTag,
                  $token->getNERTag,
                  $token->getLemma;
        }

        print "Dependencies:\n";
        for my $dep (@{$sentence->getDependencies->toArray}) {
           printf "\t%s(%s-%d, %s-%d) [%s]\n",
                  $dep->getRelation,
                  $dep->getGovernor->getWord,
                  $dep->getGovernorIndex,
                  $dep->getDependent->getWord,
                  $dep->getDependentIndex,
                  $dep->getLongRelation;
        }

        print "Coreferences:\n";
        for my $coref (@{$sentence->getCoreferences->toArray}) {
           printf "\t%s [%d, %d] <=> %s [%d, %d]\n",
                  $coref->getSourceToken->getWord,
                  $coref->getSourceSentence,
                  $coref->getSourceHead,
                  $coref->getTargetToken->getWord,
                  $coref->getTargetSentence,
                  $coref->getTargetHead;

           print "\t\t(Duplicate)\n"
              if(grep { $_->equals($coref) } @seen_corefs);

           push @seen_corefs, $coref;
        }
     }

DESCRIPTION
    This module implements a "StanfordCoreNLP" pipeline for annotating text
    with part-of-speech tags, dependencies, lemmas, named-entity tags, and
    coreferences.

    (Note that the archive contains the CoreNLP annotation models, which is
    why it's so darn big. Also note that versions before 0.10 have slightly
    different API:s than 0.10+.)

INSTALLATION
    The following should do the job:

     $ perl Build.PL
     $ ./Build test
     $ sudo ./Build install

PREREQUISITES
    Lingua::StanfordCoreNLP consists mainly of Java code, and thus needs
    Inline::Java installed to function.

ENVIRONMENT
    Lingua::StanfordCoreNLP can use the following environmental variables

  LINGUA_CORENLP_JAR_PATH
    Directory containing the CoreNLP jar-files. Normally,
    Lingua::StanfordCoreNLP expects LINGUA_CORENLP_JAR_PATH to contain the
    following files:

     stanford-corenlp-VERSION.jar
     stanford-corenlp-VERSION-models.jar
     joda-time.jar
     jollyday.jar
     xom.jar

    (Where VERSION is 1.3.4 or the value of LINGUA_CORENLP_VERSION.) If your
    filenames are different, you can add "*.jar" to the end of the path, to
    make Lingua::StanfordCoreNLP use all the jar-files in
    LINGUA_CORENLP_JAR_PATH.

  LINGUA_CORENLP_VERSION
    Version of jar-files in LINGUA_CORENLP_JAR_PATH.

  LINGUA_CORENLP_JAVA_ARGS
    Arguments to pass to JVM (via Inline::Java). Defaults to "-Xmx2000m"
    (increase max memory to 2000 MB).

EXPORTED CLASS
    Lingua::StanfordCoreNLP exports the following Java-classes via
    Inline::Java:

  Lingua::StanfordCoreNLP::Pipeline
    The main interface to "StanfordCoreNLP". This class is the only one you
    can instantiate yourself. It is, basically, a perlified
    be.fivebyfive.lingua.stanfordcorenlp.Pipeline.

    new
    new($bidirectionalCorefs)
        Creates a new "Lingua::StanfordCoreNLP::Pipeline" object. The
        optional boolean parameter $bidirectionalCorefs makes coreferences
        bidirectional; that is to say, the coreference is added to both the
        source and the target sentence of all coreferences (if the source
        and target sentence are different). $silent and $bidirectionalCorefs
        default to false.

    getProperties
        Returns a "java.util.Properties" object containing annotator
        options. By default it contains a single entry, "annotators", which
        has the value "tokenize, ssplit, pos, lemma, ner, parse, dcoref".

    setProperties($prop)
        Updates annotator options. Expects a "java.util.Properties" object.
        If you call this after having called "process", you will have to
        call "initPipeline" to update the annotator.

    getPipeline
        Returns a reference to the "StanfordCoreNLP" pipeline used for
        annotation. You probably won't want to touch this.

    initPipeline
        Reinitializes the "StanfordCoreNLP" pipeline used for annotation.

    process($str)
        Process a string. Returns a
        "Lingua::StanfordCoreNLP::PipelineSentenceList".

JAVA CLASSES
    In addition, Lingua::StanfordCoreNLP indirectly exports the following
    Java-classes, all belonging to the namespace
    "be.fivebyfive.lingua.stanfordcorenlp":

  PipelineItem
    Abstract superclass of
    "Pipeline{Coreference,Dependency,Sentence,Token}". Contains ID and
    methods for getting and comparing it.

    getID
        Returns a "java.util.UUID" object which represents the item's ID.

    getIDString
        Returns the ID as a string.

    identicalTo($b)
        Returns true if $b has an identical ID to this item.

  PipelineCoreference
    An object representing a coreference between head-word W1 in sentence S1
    and head-word W2 in sentence S2. Note that both sentences and words are
    zero-indexed, unlike the default outputs of Stanford's tools.

    getSourceSentence
        Index of sentence S1.

    getTargetSentence
        Index of sentence S2.

    getSourceHead
        Index of word W1 (in S1).

    getTargetHead
        Index of word W2 (in S2).

    getSourceToken
        The "PipelineToken" representing W1.

    getTargetToken
        The "PipelineToken" representing W2.

    equals($b)
        Returns true if this "PipelineCoreference" matches $b --- if their
        "getSourceToken" and "getTargetToken" have the same ID. Note that it
        returns true even if the orders of the coreferences are reversed (if
        "$a->getSourceToken->getID == $b->getTargetToken->getID" and
        "$a->getTargetToken->getID == $b->getSourceToken->getID").

    toCompactString
        A compact String representation of the coreference ---
        "Word/Sentence:Head <=> Word/Sentence:Head".

    toString
        A String representation of the coreference --- "Word/POS-tag
        [sentence, head] <=> Word/POS-tag [sentence, head]".

  PipelineDependency
    Represents a dependency in the Stanford Typed Dependency format. For
    example, in the fragment "Walk hard", "Walk" is the governor and "hard"
    is the dependent in the relationship "advmod" ("hard" is an adverbial
    modifier of "Walk").

    getGovernor
        The governor in the relation as a "PipelineToken".

    getGovernorIndex
        The index of the governor within the sentence.

    getDependent
        The dependent in the relation as a "PipelineToken".

    getDependentIndex
        The index of the dependent within the sentence.

    getRelation
        Short name of the relation.

    getLongRelation
        Long description of the relation.

    toCompactString
    toCompactString($includeIndices)
    toString
    toString($includeIndices)
        Returns a String representation of the dependency ---
        "relation(governor-N, dependent-N) [description]". "toCompactString"
        does not include description. The optional parameter $includeIndices
        controls whether governor and dependent indices are included, and
        defaults to true. (Note that unlike those of, e.g., the Stanford
        Parser, these indices start at zero, not one.)

  PipelineSentence
    An annotated sentence, containing the sentence itself, its dependencies,
    pos- and ner-tagged tokens, and coreferences.

    getSentence
        Returns a string containing the original sentence

    getTokens
        A "PipelineTokenList" containing the POS- and NER-tagged and
        lemmaized tokens of the sentence.

    getDependencies
        A "PipelineDependencyList" containing the dependencies found in the
        sentence.

    getCoreferences
        A "PipelineCoreferenceList" of the coreferences between this and
        other sentences.

    toCompactString
    toString
        A String representation of the sentence, its coreferences,
        dependencies, and tokens. "toCompactString" separates fields by
        "\n", whereas "toString" separates them by "\n\n".

  PipelineToken
    A token, with POS- and NER-tag and lemma.

    getWord
        The textual representation of the token (i.e. the word).

    getPOSTag
        The token's Part-of-Speech tag.

    getNERTag
        The token's Named-Entity tag.

    getLemma
        The lemma of the the token.

    toCompactString
    toCompactString($lemmaize)
        A compact String representation of the token --- "word/POS-tag". If
        the optional argument $lemmaize is true, returns "lemma/POS-tag".

    toString
        A String representation of the token --- "word/POS-tag/NER-tag
        [lemma]".

  PipelineList
  PipelineCoreferenceList
  PipelineDependencyList
  PipelineSentenceList
  PipelineTokenList
    "PipelineList" is a generic list class which extends
    "java.Util.ArrayList". It is in turn extended by
    "Pipeline{Coreference,Dependency,Sentence,Token}List" (which are the
    list-types that "Pipeline" returns). Note that all lists are
    zero-indexed.

    joinList($sep)
    joinListCompact($sep)
        Returns a string containing the output of either the "toString" or
        "toCompactString" methods of the elements in "PipelineList",
        separated by $sep.

    toArray
        Return the elements of the list as an array-reference.

    toHashMap
        Return the list as a "java.util.HashMap<String,PipelineItem>", with
        items' stringified ID:s as keys.

    toCompactString
    toString
        Returns the elements of the "PipelineList" as a string containing
        the output of either their "toCompactString" or "toString" methods,
        separated by the default separator (which is "\n" for all lists
        except "PipelineTokenList" which uses " ").

TODO
    *   Add representative mention to PipelineCoreference.

    *   Make build system also compile "LinguaSCNLP.jar".

REQUESTS & BUGS
    Please file any issues, bug-reports, or feature-requests at
    <https://github.com/raisanen/lingua-stanfordcorenlp>.

AUTHORS
    Kalle R�is�nen <kal@cpan.org>.

COPYRIGHT
  Lingua::StanfordCoreNLP (Perl bindings)
    Copyright � 2011-2013 Kalle R�is�nen.

    This program is free software: you can redistribute it and/or modify it
    under the terms of the GNU Affero General Public License as published by
    the Free Software Foundation, either version 3 of the License, or (at
    your option) any later version.

    This program is distributed in the hope that it will be useful, but
    WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero
    General Public License for more details.

    You should have received a copy of the GNU Affero General Public License
    along with this program. If not, see <http://www.gnu.org/licenses/>.

  Stanford CoreNLP tool set
    Copyright � 2010-2012 The Board of Trustees of The Leland Stanford
    Junior University.

    This program is free software; you can redistribute it and/or modify it
    under the terms of the GNU General Public License as published by the
    Free Software Foundation; either version 2 of the License, or (at your
    option) any later version.

    This program is distributed in the hope that it will be useful, but
    WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General
    Public License for more details.

    You should have received a copy of the GNU General Public License along
    with this program; if not, see <http://www.gnu.org/licenses/>.

SEE ALSO
    <http://nlp.stanford.edu/software/corenlp.shtml>,
    Text::NLP::Stanford::EntityExtract, NLP::StanfordParser, Inline::Java.