resources.coref.pleonasm.jape Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of annie Show documentation
ANNIE is a general purpose information extraction system that provides the building blocks of many other GATE applications.
The newest version!
/*
*  pleonasm.jape
*
* Copyright (c) 1998-2004, The University of Sheffield.
*
*  This file is part of GATE (see http://gate.ac.uk/), and is free
*  software, licenced under the GNU Library General Public License,
*  Version 2, June 1991 (in the distribution as file licence.html,
*  and also available at http://gate.ac.uk/gate/licence.html).
*
*  Marin DImitrov, 20/Dec/2001
*
*  $Id: pleonasm.jape 11880 2009-11-03 13:29:39Z markagreenwood $
*/


Phase:	pleonastic_it
Input: Token
Options: control = appelt


Macro: IT
(
    {Token.category=="PRP",Token.string=="It"} |
    {Token.category=="PRP",Token.string=="it"}
)


Macro: BE
(
    {Token.category=="VB",Token.string=="be"} |
    {Token.category=="VBP",Token.string=="be"} |
    {Token.category=="VBP",Token.string=="become"} |
    {Token.category=="VBP",Token.string=="remain"} |
    {Token.category=="VBZ",Token.string=="is"} |
    {Token.category=="VBZ",Token.string=="Is"} |
    {Token.category=="VBZ",Token.string=="becomes"} |
    {Token.category=="VBZ",Token.string=="remains"} |
    {Token.category=="VBD",Token.string=="was"} |
    {Token.category=="VBD",Token.string=="Was"} |
    {Token.category=="VBD",Token.string=="became"} |
    {Token.category=="VBD",Token.string=="remained"}
)

Macro: BEEN
(
   {Token.category=="VBN",Token.string=="been"} |
   {Token.category=="VBN",Token.string=="become"}
)


Macro: MODAL-VERB
(
    {Token.category=="MD"}
    ({Token.category=="RB",Token.string=="'nt"})?
)

Macro: HAS
(
   {Token.category=="VBZ",Token.string=="has"} |
   {Token.category=="VBZ",Token.string=="Has"} |
   {Token.category=="VBZ",Token.string=="had"} |
   {Token.category=="VBZ",Token.string=="Had"}
)


Macro: IT-S
(
    (IT)
    {Token.category=="VBZ",Token.orth=="apostrophe",Token.string=="'s"}
)

Macro: TO
(
    {Token.category=="TO" , Token.string=="to"}
)


Macro: NOT
(
    {Token.category=="RB",Token.string=="not"} |
    {Token.category=="RB",Token.string=="never"}
)


Macro: CONJ
(
    {Token.category=="IN",Token.string=="for"} |
    {Token.category=="IN" , Token.string=="that"} |
    {Token.category=="DT" , Token.string=="that"} |
    {Token.category=="IN" , Token.string=="if"}    |
    {Token.category=="IN" , Token.string=="whether"} |
    {Token.category=="WRB" , Token.string=="when"}
)



Macro: ADVERB-01
(
    {Token.category=="RB", Token.string=="highly"} |
    {Token.category=="RB", Token.string=="very"} |
    {Token.category=="RB", Token.string=="still"} |
    {Token.category=="RB", Token.string=="increasingly"} |
    {Token.category=="RB", Token.string=="certainly"} |
    {Token.category=="RB", Token.string=="absolutely"} |
    {Token.category=="RB", Token.string=="especially"} |
    {Token.category=="RB", Token.string=="entirely"} |
    {Token.category=="RB", Token.string=="simply"} |
    {Token.category=="RB", Token.string=="particularly"} |
    {Token.category=="RB", Token.string=="quite"} |
    {Token.category=="RB", Token.string=="also"} |
    {Token.category=="RB", Token.string=="yet"} |
    {Token.category=="RB", Token.string=="even"} |
    {Token.category=="RBR", Token.string=="more"} |
    {Token.category=="RBS", Token.string=="most"} |
    {Token.category=="RB", Token.string=="often"} |
    {Token.category=="RB", Token.string=="rarely"}
)


Macro: TIME-MODIF
(
    {Token.category=="RB", Token.string=="just"} |
    {Token.category=="IN", Token.string=="about"}
)

Macro: MODALADJ
(
    {Token.category=="JJ" , Token.string=="advisable"} |
    {Token.category=="JJ" , Token.string=="unadvisable"} |
    {Token.category=="JJ" , Token.string=="inadvisable"} |
    {Token.category=="JJ" , Token.string=="suggested"} |
    {Token.category=="JJ" , Token.string=="recommended"} |
    {Token.category=="JJ" , Token.string=="sensible"} |
    {Token.category=="JJ" , Token.string=="resonalble"} |
    {Token.category=="JJ" , Token.string=="unresonalble"} |
    {Token.category=="JJ" , Token.string=="wise"} |
    {Token.category=="JJ" , Token.string=="unwise"} |

    {Token.category=="JJ" , Token.string=="convenient"} |
    {Token.category=="JJ" , Token.string=="inconvenient"} |

    {Token.category=="JJ" , Token.string=="desirable"} |
    {Token.category=="JJ" , Token.string=="undesirable"} |
    {Token.category=="JJ" , Token.string=="suitable"} |
    {Token.category=="JJ" , Token.string=="unsuitable"} |
    {Token.category=="JJ" , Token.string=="worthy"} |
    {Token.category=="JJ" , Token.string=="unworthy"} |

    {Token.category=="JJ" , Token.string=="difficult"} |
    {Token.category=="JJ" , Token.string=="hard"} |

    {Token.category=="JJ" , Token.string=="easy"} |
    {Token.category=="JJ" , Token.string=="painless"} |

    {Token.category=="JJ" , Token.string=="economical"} |
    {Token.category=="JJ" , Token.string=="efficient"} |
    {Token.category=="JJ" , Token.string=="inefficient"} |

    {Token.category=="JJ" , Token.string=="certain"} |
    {Token.category=="JJ" , Token.string=="uncertain"} |
    {Token.category=="JJ" , Token.string=="sure"} |
    {Token.category=="JJ" , Token.string=="unsure"} |

    {Token.category=="JJ" , Token.string=="good"} |
    {Token.category=="JJ" , Token.string=="bad"} |
    {Token.category=="JJ" , Token.string=="acceptable"} |
    {Token.category=="JJ" , Token.string=="unacceptable"} |

    {Token.category=="JJ" , Token.string=="important"} |
    {Token.category=="JJ" , Token.string=="unimportant"} |
    {Token.category=="JJ" , Token.string=="useful"} |
    {Token.category=="JJ" , Token.string=="useless"} |
    {Token.category=="JJ" , Token.string=="practical"} |
    {Token.category=="JJ" , Token.string=="impractical"} |

    {Token.category=="JJ" , Token.string=="legal"} |
    {Token.category=="JJ" , Token.string=="illegal"} |
    {Token.category=="JJ" , Token.string=="valid"} |
    {Token.category=="JJ" , Token.string=="invalid"} |

    {Token.category=="JJ" , Token.string=="likely"} |
    {Token.category=="JJ" , Token.string=="unlikely"} |
    {Token.category=="JJ" , Token.string=="probable"} |
    {Token.category=="JJ" , Token.string=="improbable"} |
    {Token.category=="JJ" , Token.string=="expected"} |
    {Token.category=="JJ" , Token.string=="unexpected"} |
    {Token.category=="JJ" , Token.string=="plausible"} |
    {Token.category=="JJ" , Token.string=="implausible"} |

    {Token.category=="JJ" , Token.string=="necessary"} |
    {Token.category=="JJ" , Token.string=="unnecessary"} |
    {Token.category=="JJ" , Token.string=="essential"} |
    {Token.category=="JJ" , Token.string=="inessential"} |
    {Token.category=="JJ" , Token.string=="obligatory"} |
    {Token.category=="JJ" , Token.string=="mandatory"} |

    {Token.category=="JJ" , Token.string=="possible"} |
    {Token.category=="JJ" , Token.string=="impossible"} |
    {Token.category=="JJ" , Token.string=="feasible"} |

    {Token.category=="JJ" , Token.string=="sufficient"} |
    {Token.category=="JJ" , Token.string=="insufficient"} |
    {Token.category=="JJ" , Token.string=="adequate"} |
    {Token.category=="JJ" , Token.string=="inadequate"} |
    {Token.category=="JJ" , Token.string=="enough"} |
    {Token.category=="RB" , Token.string=="enough"} |

    {Token.category=="JJ" , Token.string=="clear"} |
    {Token.category=="JJ" , Token.string=="unclear"} |
    {Token.category=="JJ" , Token.string=="tough"} |
    {Token.category=="JJ" , Token.string=="typical"} |
    {Token.category=="JJ" , Token.string=="untypical"} |
    {Token.category=="JJ" , Token.string=="imperative"} |
    {Token.category=="JJ" , Token.string=="appropriate"} |
    {Token.category=="JJ" , Token.string=="inappropriate"} |
    {Token.category=="JJ" , Token.string=="fair"} |
    {Token.category=="JJ" , Token.string=="unfair"} |
    {Token.category=="JJ" , Token.string=="common"} |
    {Token.category=="JJ" , Token.string=="uncommon"} |
    {Token.category=="JJ" , Token.string=="rare"} |
    {Token.category=="JJ" , Token.string=="wrong"} |
    {Token.category=="JJ" , Token.string=="safe"} |
    {Token.category=="JJ" , Token.string=="unsafe"} |
    {Token.category=="JJ" , Token.string=="usual"} |
    {Token.category=="JJ" , Token.string=="unusual"} |
    {Token.category=="JJR" , Token.string=="better"} |
    {Token.category=="JJR" , Token.string=="great"}
)



Macro: COGV-ED
(
    {Token.category=="VBN" , Token.string=="anticipated"} |
    {Token.category=="VBN" , Token.string=="expected"}	|
    {Token.category=="VBN" , Token.string=="foreseen"}	|
    {Token.category=="VBN" , Token.string=="predicted"} |
    {Token.category=="VBN" , Token.string=="promised"}	|

    {Token.category=="VBN" , Token.string=="assumed"}  |
    {Token.category=="VBN" , Token.string=="accepted"}	|

    {Token.category=="VBN" , Token.string=="believed"}	|

    {Token.category=="VBN" , Token.string=="required"}	|
    {Token.category=="VBN" , Token.string=="demanded"}	|

    {Token.category=="VBN" , Token.string=="known"}  |
    {Token.category=="VBN" , Token.string=="recognized"}  |
    {Token.category=="VBN" , Token.string=="realized"}	|

    {Token.category=="VBN" , Token.string=="recommended"}  |
    {Token.category=="VBN" , Token.string=="proposed"}	|
    {Token.category=="VBN" , Token.string=="suggested"}  |

    {Token.category=="VBN" , Token.string=="thought"}  |
    {Token.category=="VBN" , Token.string=="considered"}  |
    {Token.category=="VBN" , Token.string=="intended"} |

    {Token.category=="VBN" , Token.string=="reported"} |
    {Token.category=="VBN" , Token.string=="estimated"} |
    {Token.category=="VBN" , Token.string=="denied"} |
    {Token.category=="VBN" , Token.string=="announced"} |
    {Token.category=="VBN" , Token.string=="calculated"} |
    {Token.category=="VBN" , Token.string=="declared"}
)

Macro: DO
(
    {Token.category=="VB" , Token.string=="do"} |
    {Token.category=="VBZ" , Token.string=="does"} |
    {Token.category=="VBD" , Token.string=="did"}
)

Macro: DO-NEG
(
    {Token.category=="RB" , Token.string=="n't"}
)

Macro: SEEM-VERB
(
    {Token.category=="VBZ" , Token.string=="seems"} |
    {Token.category=="VB" , Token.string=="seem"} |
    {Token.category=="VBD" , Token.string=="seemed"} |
    {Token.category=="VBZ" , Token.string=="appears"} |
    {Token.category=="VB" , Token.string=="appear"} |
    {Token.category=="VBD" , Token.string=="appeared"} |
    {Token.category=="VBZ" , Token.string=="looks"} |
    {Token.category=="VB" , Token.string=="look"} |
    {Token.category=="VBD" , Token.string=="looked"} |
    {Token.category=="VBZ" , Token.string=="means"} |
    {Token.category=="VB" , Token.string=="mean"} |
    {Token.category=="VBD" , Token.string=="meant"} |
    {Token.category=="VBZ" , Token.string=="happens"} |
    {Token.category=="VB" , Token.string=="happen"} |
    {Token.category=="VBD" , Token.string=="happened"} |
    {Token.category=="VBZ" , Token.string=="sounds"} |
    {Token.category=="VB" , Token.string=="sound"} |
    {Token.category=="VBD" , Token.string=="sounded"}
)

Macro: CONJ-02
(
    {Token.category=="IN" , Token.string=="that"} |
    {Token.category=="DT" , Token.string=="that"} |
    {Token.category=="IN" , Token.string=="if"}    |
    {Token.category=="IN" , Token.string=="as"}    |
    {Token.category=="IN" , Token.string=="like"}
)

Macro: IT-BE
(
   ( (MODAL-VERB)? (IT) (MODAL-VERB)? (BE) ) |
   ( (HAS)? (DO-NEG)? (IT) (HAS)? (DO-NEG)? (BEEN) ) |
   ( (IT-S) (BEEN)? ) |
   ( (BE)(IT) )
)

Macro: FIND
(
    {Token.category=="VBZ" , Token.string=="finds"} |
    {Token.category=="VB" , Token.string=="find"} |
    {Token.category=="VBD" , Token.string=="found"} |
    {Token.category=="VBN" , Token.string=="found"} |
    {Token.category=="VBZ" , Token.string=="makes"} |
    {Token.category=="VB" , Token.string=="make"} |
    {Token.category=="VBD" , Token.string=="made"} |
    {Token.category=="VBN" , Token.string=="made"} |
    {Token.category=="VBZ" , Token.string=="considers"} |
    {Token.category=="VB" , Token.string=="consider"} |
    {Token.category=="VBD" , Token.string=="considered"} |
    {Token.category=="VBN" , Token.string=="considered"}
)

/*
    IT_IS_ADVERB_MODADJ
*/
Rule:Rule_A_B
Priority: 99
(
   (IT-BE)
   (NOT)?
   (ADVERB-01)?
   (NOT)?
   (MODALADJ)
   ((CONJ) | (TO) | {Token.category == "DT",Token.string == "the"})
) :pleonasm
-->
{
	gate.AnnotationSet found_pleonasm = (gate.AnnotationSet)bindings.get("pleonasm");
	Annotation pleonasm = (Annotation)found_pleonasm.iterator().next();

	gate.FeatureMap features = Factory.newFeatureMap();
	features.put("firedBy","ruleA+B");
	outputAS.add(found_pleonasm.firstNode(), found_pleonasm.lastNode(), "PleonasticIt", features);

    //try to get doc feature
    Integer cnt = (Integer)doc.getFeatures().get("PLEONASM");

    if (null == cnt) {
	cnt = new Integer(1);
    }
    else {
	cnt = new Integer(cnt.intValue() + 1);
    }

    doc.getFeatures().put("PLEONASM",cnt);

}




/*
    IT_IS_COGV-ED_THAT
*/

Rule:Rule_C
Priority: 77
(
   (
      ( (IT)(BE) ) |
      ( (BE)(IT) )
   )
   (NOT)?
   (ADVERB-01)?
   (COGV-ED)
   ({Token.category=="IN" , Token.string=="that"})?
) :pleonasm
-->
{
	gate.AnnotationSet found_pleonasm = (gate.AnnotationSet)bindings.get("pleonasm");
	Annotation pleonasm = (Annotation)found_pleonasm.iterator().next();

	gate.FeatureMap features = Factory.newFeatureMap();
	features.put("firedBy","ruleC");
	outputAS.add(found_pleonasm.firstNode(), found_pleonasm.lastNode(), "PleonasticIt", features);

    //try to get doc feature
    Integer cnt = (Integer)doc.getFeatures().get("PLEONASM");

    if (null == cnt) {
	cnt = new Integer(1);
    }
    else {
	cnt = new Integer(cnt.intValue() + 1);
    }

    doc.getFeatures().put("PLEONASM",cnt);

}


/*
    IT_[ADVERB-02]_VERB-01_[THAT,TO]_S
*/

Rule:Rule_D
Priority: 66
(
   (DO)?
   (DO-NEG)?
   (IT)
   (ADVERB-01)?
   (DO)?
   (DO-NEG)?
   (SEEM-VERB)
   ( (CONJ-02) | (TO) )?
) :pleonasm
-->
{
	gate.AnnotationSet found_pleonasm = (gate.AnnotationSet)bindings.get("pleonasm");
	Annotation pleonasm = (Annotation)found_pleonasm.iterator().next();

	gate.FeatureMap features = Factory.newFeatureMap();
	features.put("firedBy","ruleD");
	outputAS.add(found_pleonasm.firstNode(), found_pleonasm.lastNode(), "PleonasticIt", features);

    //try to get doc feature
    Integer cnt = (Integer)doc.getFeatures().get("PLEONASM");

    if (null == cnt) {
	cnt = new Integer(1);
    }
    else {
	cnt = new Integer(cnt.intValue() + 1);
    }

    doc.getFeatures().put("PLEONASM",cnt);

}

Rule:Rule_E
Priority: 55
(
   ( (DO) | (HAS) )?
   (DO-NEG)?
   (FIND)
   (IT)
   (ADVERB-01)?
   (MODALADJ)
   ( (CONJ) | (TO) )?
) :pleonasm
-->
{
	gate.AnnotationSet found_pleonasm = (gate.AnnotationSet)bindings.get("pleonasm");
	Annotation pleonasm = (Annotation)found_pleonasm.iterator().next();

	gate.FeatureMap features = Factory.newFeatureMap();
	features.put("firedBy","ruleE");
	outputAS.add(found_pleonasm.firstNode(), found_pleonasm.lastNode(), "PleonasticIt", features);

    //try to get doc feature
    Integer cnt = (Integer)doc.getFeatures().get("PLEONASM");

    if (null == cnt) {
	cnt = new Integer(1);
    }
    else {
	cnt = new Integer(cnt.intValue() + 1);
    }

    doc.getFeatures().put("PLEONASM",cnt);

}


/*
    IT_IS_[about]_time_[that,to]
*/
/*
Rule:Rule_D
Priority: 55
(
    (
	(   ((IT) (BE)) |
	    (IT-S)
	)
    (TIME-MODIF)
    {Token.category=="NN" , Token.string=="time"}
    ( (THAT) | (TO) | (FOR))
    ) :pleonasm
)
-->
{
	gate.AnnotationSet found_pleonasm = (gate.AnnotationSet)bindings.get("pleonasm");
	Annotation pleonasm = (Annotation)found_pleonasm.iterator().next();

	gate.FeatureMap features = Factory.newFeatureMap();
	features.put("firedBy","ruleD");
	outputAS.add(found_pleonasm.firstNode(), found_pleonasm.lastNode(), "PleonasticIt", features);

    //try to get doc feature
    Integer cnt = (Integer)doc.getFeatures().get("PLEONASM");

    if (null == cnt) {
	cnt = new Integer(1);
    }
    else {
	cnt = new Integer(cnt.intValue() + 1);
    }

    doc.getFeatures().put("PLEONASM",cnt);

}
*/