From 8f86545e14f99bbf47ab83bf202e26af7a2716c4 Mon Sep 17 00:00:00 2001
From: Bartlomiej Niton <bartek.niton@gmail.com>
Date: Wed, 25 Jan 2017 15:36:01 +0100
Subject: [PATCH] Cleaning unused experimental code.
---
src/main/java/pl/waw/ipipan/zil/core/md/Main.java | 42 ++++--------------------------------------
src/main/java/pl/waw/ipipan/zil/core/md/detection/Cleaner.java | 74 +-------------------------------------------------------------------------
src/main/java/pl/waw/ipipan/zil/core/md/detection/Constants.java | 8 ++++++++
src/main/java/pl/waw/ipipan/zil/core/md/detection/Detector.java | 545 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
src/main/java/pl/waw/ipipan/zil/core/md/entities/Mention.java | 65 ++++++-----------------------------------------------------------
src/main/java/pl/waw/ipipan/zil/core/md/entities/Sentence.java | 71 ++++-------------------------------------------------------------------
src/main/java/pl/waw/ipipan/zil/core/md/entities/SyntacticGroup.java | 98 ++++++++++++++++++++++++++++++++------------------------------------------------------------------
src/main/java/pl/waw/ipipan/zil/core/md/entities/SyntacticWord.java | 31 +++++++++++++------------------
src/main/java/pl/waw/ipipan/zil/core/md/io/tei/TeiLoader.java | 1 -
9 files changed, 233 insertions(+), 702 deletions(-)
diff --git a/src/main/java/pl/waw/ipipan/zil/core/md/Main.java b/src/main/java/pl/waw/ipipan/zil/core/md/Main.java
index 4740d94..e70f152 100644
--- a/src/main/java/pl/waw/ipipan/zil/core/md/Main.java
+++ b/src/main/java/pl/waw/ipipan/zil/core/md/Main.java
@@ -33,9 +33,8 @@ public class Main {
private static final boolean GZIP_OUTPUT = true;
private static final String DEFAULT_ZERO_SUBJECT_MODEL = "/zero_subject_model.bin";
- private static final String DEFAULT_VERBS_VALENCE = "/walenty_20170117_verbs_all.txt";
- private static final String DEFAULT_NOUNS_VALENCE = "/walenty_20170117_nouns_all.txt";
- private static final String COMPLEX_PREPS = "/complex_preps_Walenty_USJP.txt";
+ private static final String DEFAULT_VERBS_VALENCE = "/walenty_20170117_verbs_all_with_realizations.txt";
+ private static final String DEFAULT_NOUNS_VALENCE = "/walenty_20170117_nouns_all_with_realizations.txt";
private static ZeroSubjectDetector zeroSubjectModel;
@@ -46,8 +45,6 @@ public class Main {
private static Map<ValenceDicts,Map<String,ArrayList<String>>> valence =
new EnumMap(ValenceDicts.class);
-
- private static final ArrayList<String> complexPreps;
static {
InputStream zeroSubjectDetectionModelStream = Main.class.getResourceAsStream(DEFAULT_ZERO_SUBJECT_MODEL);
@@ -58,9 +55,6 @@ public class Main {
InputStream walentyNounsStream = Main.class.getResourceAsStream(DEFAULT_NOUNS_VALENCE);
valence.put(ValenceDicts.NounsValence, readWalenty(walentyNounsStream));
-
- InputStream complexPrepositionsStream = Main.class.getResourceAsStream(COMPLEX_PREPS);
- complexPreps = readValues(complexPrepositionsStream);
}
@@ -125,34 +119,6 @@ public class Main {
return false;
}
-
- public static ArrayList<String> readValues(InputStream stream) {
- ArrayList<String> values;
- try {
- BufferedReader br=new BufferedReader(new InputStreamReader(stream));
- values = new ArrayList<String>();
- String line;
- boolean firstLine = true;
- while((line = br.readLine()) != null) {
- if (firstLine) {
- line = line.replace("\uFEFF", ""); // remove BOM character
- firstLine = false;
- }
-
- if (!line.startsWith("%")) {
- String value = line.trim();
- if (!value.isEmpty()) {
- values.add(value);
- }
- }
- }
- br.close();
- } catch (IOException ex) {
- ex.printStackTrace();
- throw new RuntimeException(ex);
- }
- return values;
- }
private Main() {
}
@@ -244,7 +210,7 @@ public class Main {
*/
public static void annotateThriftText(TText thriftText) throws MultiserviceException {
Text responseText = ThriftLoader.loadTextFromThrift(thriftText);
- Detector.findMentionsInText(responseText, zeroSubjectModel, valence, complexPreps);
+ Detector.findMentionsInText(responseText, zeroSubjectModel, valence);
ThriftSaver.updateThriftText(responseText, thriftText);
}
@@ -257,7 +223,7 @@ public class Main {
*/
public static void annotateTeiText(TEICorpusText teiText) throws TEIException {
Text responseText = TeiLoader.loadTextFromTei(teiText);
- Detector.findMentionsInText(responseText, zeroSubjectModel, valence, complexPreps);
+ Detector.findMentionsInText(responseText, zeroSubjectModel, valence);
TeiSaver.updateTeiText(responseText, teiText);
}
diff --git a/src/main/java/pl/waw/ipipan/zil/core/md/detection/Cleaner.java b/src/main/java/pl/waw/ipipan/zil/core/md/detection/Cleaner.java
index 23a83b3..9eb2ec8 100644
--- a/src/main/java/pl/waw/ipipan/zil/core/md/detection/Cleaner.java
+++ b/src/main/java/pl/waw/ipipan/zil/core/md/detection/Cleaner.java
@@ -1,6 +1,5 @@
package pl.waw.ipipan.zil.core.md.detection;
-import pl.waw.ipipan.zil.core.md.Main.ValenceDicts;
import pl.waw.ipipan.zil.core.md.entities.Mention;
import pl.waw.ipipan.zil.core.md.entities.Sentence;
import pl.waw.ipipan.zil.core.md.entities.SyntacticGroup;
@@ -164,33 +163,6 @@ public class Cleaner {
}
}
- /*private static void removeWalentyFramedMentions(Sentence sentence,
- ArrayList<Mention> mentions,
- ArrayList<String> schemata) {
- ArrayList<Mention> mentionsToRemove = new ArrayList<Mention>();
- for (Mention mention : mentions) {
- int mentionStart = mention.getFirstSegment().getSentencePosition();
- int mentionEnd = mention.getLastSegment().getSentencePosition();
- SyntacticGroup startGroup = sentence.getFirstGroup(mentionStart, mentionEnd);
- SyntacticGroup endGroup = sentence.getLastGroup(mentionStart, mentionEnd);
- if (startGroup != null && endGroup != null
- && startGroup.compareTo(endGroup) != 0) {
- ArrayList<String> startGroupRealizations = startGroup.getWalentyRealizations();
- ArrayList<String> endGroupRealizations = endGroup.getWalentyRealizations();
- for (String schema : schemata) {
- if (isProperSchema(schema, startGroupRealizations, endGroupRealizations)) {
- mentionsToRemove.add(mention);
- break;
- }
- }
- }
- }
-
- for (Mention mentionToRemove : mentionsToRemove) {
- sentence.removeMention(mentionToRemove);
- }
- }*/
-
private static boolean isProperSchema(String schema, ArrayList<String> group1Types,
ArrayList<String> group2Types) {
for (String group1Type : group1Types) {
@@ -207,7 +179,7 @@ public class Cleaner {
String phraseType2) {
boolean phrType1Found = false;
boolean phrType2Found = false;
- for (String position : schema.split("\\+")) {
+ for (String position : schema.split("\\s\\+\\s")) {
position = position.trim();
position = position.substring(1, position.length()-1);
for (String phrT : position.split(";")) {
@@ -226,34 +198,6 @@ public class Cleaner {
return false;
}
-
- // wykrywa, ze "wszystkim" jest czescia "przede wszystkim" (kublika Qub)
- public static void cleanQubs(Sentence sentence) {
- ArrayList<Mention> mentionsToRemove = new ArrayList<Mention>();
- for (Mention mention : sentence.getMentions()) {
- if (mention.isPartOfQub()) {
- mentionsToRemove.add(mention);
- }
- }
-
- for (Mention mentionToRemove : mentionsToRemove) {
- sentence.removeMention(mentionToRemove);
- }
- }
-
- public static void cleanPreps(Sentence sentence) {
- ArrayList<Mention> mentionsToRemove = new ArrayList<Mention>();
- for (Mention mention : sentence.getMentions()) {
- if (mention.isPartOfPrep()) {
- mentionsToRemove.add(mention);
- }
- }
-
- for (Mention mentionToRemove : mentionsToRemove) {
- sentence.removeMention(mentionToRemove);
- }
- }
-
public static void cleanFrazeos(Sentence sentence) {
ArrayList<Mention> mentionsToRemove = new ArrayList<Mention>();
for (Mention mention : sentence.getMentions()) {
@@ -267,20 +211,4 @@ public class Cleaner {
}
}
- // wyrzuca wzmianki bedace czescia przyimkow zlozonych
- public static void cleanComplexPreps(Sentence sentence,
- ArrayList<String> complexPreps) {
-
- ArrayList<Mention> mentionsToRemove = new ArrayList<Mention>();
- for (Mention mention : sentence.getMentions()) {
- if (mention.isPartOfComplexPrep(complexPreps)) {
- mentionsToRemove.add(mention);
- }
- }
-
- for (Mention mentionToRemove : mentionsToRemove) {
- sentence.removeMention(mentionToRemove);
- }
- }
-
}
diff --git a/src/main/java/pl/waw/ipipan/zil/core/md/detection/Constants.java b/src/main/java/pl/waw/ipipan/zil/core/md/detection/Constants.java
index 19e6d56..cc7aa2a 100644
--- a/src/main/java/pl/waw/ipipan/zil/core/md/detection/Constants.java
+++ b/src/main/java/pl/waw/ipipan/zil/core/md/detection/Constants.java
@@ -1,5 +1,8 @@
package pl.waw.ipipan.zil.core.md.detection;
+import java.util.Arrays;
+import java.util.List;
+
public class Constants {
public static final String MORPHO_NOUN_CTAGS = "subst|depr|ger";
public static final String MORPHO_VERB_CTAGS = "fin|bedzie|aglt|impt";
@@ -7,6 +10,11 @@ public class Constants {
public static final String MORPHO_CTAGS = MORPHO_NOUN_CTAGS + "|"
+ MORPHO_PRONOUN_CTAGS;
public static final String WORDS_CTAGS = "Noun|Ppron.*";
+
+ public static final List<String> FRAZEO_CTAGS = Arrays.asList("Prep", "Qub", "Adv", "Interj",
+ "Adj", "Conj", "Comp");
+
+ public static final List<String> VERB_CTAGS = Arrays.asList("Inf", "Verbfin");
private Constants() {
}
diff --git a/src/main/java/pl/waw/ipipan/zil/core/md/detection/Detector.java b/src/main/java/pl/waw/ipipan/zil/core/md/detection/Detector.java
index c8c89b2..399e9e3 100644
--- a/src/main/java/pl/waw/ipipan/zil/core/md/detection/Detector.java
+++ b/src/main/java/pl/waw/ipipan/zil/core/md/detection/Detector.java
@@ -22,24 +22,22 @@ public class Detector {
public static void findMentionsInText(Text text,
ZeroSubjectDetector zeroSubjectModel,
- Map<ValenceDicts,Map<String,ArrayList<String>>> valence,
- ArrayList<String> complexPreps) {
+ Map<ValenceDicts,Map<String,ArrayList<String>>> valence) {
text.clearMentions();
logger.debug("Detecting mentions in text " + text.getId());
for (Paragraph p : text)
for (Sentence s : p)
- detectMentionsInSentence(s, zeroSubjectModel, valence, complexPreps);
+ detectMentionsInSentence(s, zeroSubjectModel, valence);
}
private static void detectMentionsInSentence(Sentence sentence,
ZeroSubjectDetector zeroSubjectModel,
- Map<ValenceDicts,Map<String,ArrayList<String>>> valence,
- ArrayList<String> complexPreps) {
+ Map<ValenceDicts,Map<String,ArrayList<String>>> valence) {
// adding mentions
addMentionsByTokenCtag(sentence);
addMentionsBySyntacticWordsCtag(sentence);
addMentionsByNamedEntities(sentence);
- addMentionsByGroups(sentence, valence, complexPreps);
+ addMentionsByGroups(sentence, valence);
addSpeakerMentionsInSpoken(sentence);
// zero subject detection
@@ -47,12 +45,9 @@ public class Detector {
// removing mentions
removeTo(sentence);
+ Cleaner.cleanWalentyFramedMentions(sentence, valence.get(ValenceDicts.VerbsValence));
Cleaner.cleanUnnecessarySentenceMentions(sentence);
- //Cleaner.cleanQubs(sentence);
- //Cleaner.cleanPreps(sentence);
- //Cleaner.cleanComplexPreps(sentence, complexPreps);
Cleaner.cleanFrazeos(sentence);
- Cleaner.cleanWalentyFramedMentions(sentence, valence.get(ValenceDicts.VerbsValence));
// updating mention heads
updateMentionHeads(sentence);
@@ -108,294 +103,64 @@ public class Detector {
* @param sentence
*/
private static void addMentionsByGroups(Sentence sentence,
- Map<ValenceDicts,Map<String,ArrayList<String>>> valence,
- ArrayList<String> complexPreps) {
- List<SyntacticGroup> groups = sentence.getGroups();
- for (int i = 0; i < groups.size(); i++) {
- SyntacticGroup thisGroup = groups.get(i);
-
- /*SyntacticGroup nearPrepNG = null;
- SyntacticGroup nextNG = null;*/
-
- SyntacticGroup nextGroup = thisGroup.getFollowingGroup();
-
- /*if (thisGroup.getType().startsWith("NG")) {
- nearPrepNG = getFollowingPrepNGs(thisGroup.getSentencePositionEnd(),
- sentence);
- nextNG = thisGroup.getNextNG();
- }*/
-
- /*if (nextNG != null) {
- int prepStart = thisGroup.getSentencePositionEnd() + 1;
- int prepEnd = nextNG.getSentencePositionStart() - 1;
- String prep = sentence.getTextInsideSpan(prepStart, prepEnd);
- if (complexPreps.contains(prep)) {
- String cos = "";
- }
- }*/
+ Map<ValenceDicts,Map<String,ArrayList<String>>> valence) {
+
+ for (SyntacticGroup group : sentence.getGroups()) {
+ SyntacticGroup nextGroup = group.getFollowingGroup();
+ SyntacticGroup nextnextGroup = null;
+ SyntacticGroup nextnextnextGroup = null;
+ if (nextGroup != null) {
+ nextnextGroup = nextGroup.getFollowingGroup();
+ if (nextnextGroup != null) {
+ nextnextnextGroup = nextnextGroup.getFollowingGroup();
+ }
+ }
- /*if (thisGroup.getType().startsWith("NG") && nearPrepNG != null &&
- //!isPartOfPrepNG(thisGroup, sentence) &&
- //getFollowingPrepNGs(nearPrepNG.getSentencePositionEnd(), sentence) == null &&
- precedingWordIsVerb(thisGroup, sentence) &&
- //!precedingVerbModifyPrepNG(thisGroup, nearPrepNG, sentence, valence.get(ValenceDicts.VerbsValence)) &&
- !precedingVerbModifyNG(thisGroup, nearPrepNG, sentence, valence.get(ValenceDicts.VerbsValence)) &&
- !sameSemanticHeads(thisGroup, nearPrepNG)) {
- List<Token> heads = thisGroup.getSemanticHeadTokens();
- List<Token> segments = thisGroup.getTokens();
- segments.addAll(nearPrepNG.getTokens());
-
- sentence.addMention(new Mention(segments, heads));
- }*/
- /*if (thisGroup.getType().startsWith("NG") && nearPrepNG != null &&
- // !precedingWordIsVerb(thisGroup, sentence) &&
- !isPartOfPrepNG(thisGroup, sentence) &&
- getFollowingPrepNGs(nearPrepNG.getSentencePositionEnd(), sentence) == null &&
- //!precedingWordIsVerb(thisGroup, sentence) &&
- !precedingVerbModifyPrepNG(thisGroup, nearPrepNG, sentence, valence.get(ValenceDicts.VerbsValence)) &&
- //!precedingVerbModifyNG(thisGroup, nearPrepNG, sentence, valence.get(ValenceDicts.VerbsValence)) &&
- !sameSemanticHeads(thisGroup, nearPrepNG)) {
- List<Token> heads = thisGroup.getSemanticHeadTokens();
- List<Token> segments = thisGroup.getTokens();
- segments.addAll(nearPrepNG.getTokens());
-
- sentence.addMention(new Mention(segments, heads));
- }*/
- if (thisGroup.getType().startsWith("NG") &&
- nextGroup != null && nextGroup.getType().startsWith("PrepNG") &&
- NGPrepNGValenceCompatibility(thisGroup, nextGroup, sentence, valence.get(ValenceDicts.NounsValence))) {
- List<Token> heads = thisGroup.getSemanticHeadTokens();
+ if (group.getType().startsWith("NG") && nextGroup != null &&
+ nextnextGroup != null && nextnextnextGroup != null &&
+ quatroCompatibility(group, nextGroup, nextnextGroup,
+ nextnextnextGroup, valence.get(ValenceDicts.NounsValence))) {
+ List<Token> heads = group.getSemanticHeadTokens();
List<Token> segments = new ArrayList<Token>();
- segments.addAll(thisGroup.getTokens());
+ segments.addAll(group.getTokens());
segments.addAll(nextGroup.getTokens());
+ segments.addAll(nextnextGroup.getTokens());
+ segments.addAll(nextnextnextGroup.getTokens());
sentence.addMention(new Mention(segments, heads));
- } else if (thisGroup.getType().startsWith("NG") && nextGroup != null &&
- nextGroup.getType().startsWith("NG") &&
- NGNGValenceCompatibility(thisGroup, nextGroup, sentence, valence.get(ValenceDicts.NounsValence))
- ) {
- List<Token> heads = thisGroup.getSemanticHeadTokens();
+ } else if (group.getType().startsWith("NG") && nextGroup != null &&
+ nextnextGroup != null && tripleCompatibility(group, nextGroup, nextnextGroup, valence.get(ValenceDicts.NounsValence))) {
+ List<Token> heads = group.getSemanticHeadTokens();
List<Token> segments = new ArrayList<Token>();
- segments.addAll(thisGroup.getTokens());
+ segments.addAll(group.getTokens());
segments.addAll(nextGroup.getTokens());
+ segments.addAll(nextnextGroup.getTokens());
sentence.addMention(new Mention(segments, heads));
- } /*else if (thisGroup.getType().startsWith("NG") && nextNG != null && nearPrepNG == null &&
- NGcomplexPrepNGValenceCompatibility(thisGroup, nextNG, sentence, valence.get(ValenceDicts.NounsValence))) {
- List<Token> heads = thisGroup.getSemanticHeadTokens();
-
+ } else if (group.getType().startsWith("NG") && nextGroup != null &&
+ groupsValenceCompatibility(group, nextGroup, sentence, valence.get(ValenceDicts.NounsValence))
+ ) {
+ List<Token> heads = group.getSemanticHeadTokens();
List<Token> segments = new ArrayList<Token>();
- segments.addAll(thisGroup.getTokens());
-
- int prepStart = thisGroup.getSentencePositionEnd() + 1;
- int prepEnd = nextNG.getSentencePositionStart() - 1;
- ArrayList<Token> prepSegments = sentence.getSegmentsInsideSpan(prepStart, prepEnd);
- segments.addAll(prepSegments);
-
- segments.addAll(nextNG.getTokens());
+ segments.addAll(group.getTokens());
+ segments.addAll(nextGroup.getTokens());
sentence.addMention(new Mention(segments, heads));
- }*/
- //else if // NG + im./pt. NG
- // daty nie sa oznaczane np 21 kwietnia itd. , to co z ta gramatyka
- // "instytut naukowy w montrealu" czemu sie nie laczy? ==> NE(orgName) + w(prep) + NE(placeName)
- else if (thisGroup.getType().startsWith("NG")) {
- List<Token> segments = thisGroup.getTokens();
- List<Token> heads = thisGroup.getSemanticHeadTokens();
-
- sentence.addMention(new Mention(segments, heads));
- }
- }
-
- // oryginalna wersja
- /*for (SyntacticGroup group : sentence.getGroups()) {
- if (group.getType().startsWith("NG")) {
+ } else if (group.getType().startsWith("NG")) {
List<Token> segments = group.getTokens();
List<Token> heads = group.getSemanticHeadTokens();
sentence.addMention(new Mention(segments, heads));
}
- }*/
- }
-
- private static boolean followingWordIsInf(SyntacticGroup group,
- Sentence sentence) {
- int followingTokenPosition = group.getSentencePositionEnd() + 1;
- for (SyntacticWord word : sentence.getSyntacticWords()) {
- int firstWordPosition = word.getSentencePositionStart();
- if (followingTokenPosition == firstWordPosition &&
- (word.getCtag().equals("Inf"))) {
- return true;
- }
- }
-
- return false;
- }
-
- private static SyntacticGroup getFollowingPrepNGs(int sentencePosition,
- Sentence sentence) {
- SyntacticGroup largestGroup = null;
- int nextTokenPosition = sentencePosition + 1;
- for (SyntacticGroup group : sentence.getGroups()) {
- if (group.getType().startsWith("PrepNG") &&
- group.getSentencePositionStart() == nextTokenPosition) {
- if (largestGroup == null ||
- largestGroup.getTokens().size() < group.getTokens().size()) {
- largestGroup = group;
- }
- }
- }
- return largestGroup;
- }
-
- private static boolean isPartOfPrepNG(SyntacticGroup NGGroup,
- Sentence sentence) {
- int NGGroupStart = NGGroup.getSentencePositionStart();
- int NGGroupEnd = NGGroup.getSentencePositionEnd();
- for (SyntacticGroup group : sentence.getGroups()) {
- if (group.getType().startsWith("PrepNG") &&
- group.getSentencePositionStart() <= NGGroupStart &&
- group.getSentencePositionEnd() >= NGGroupEnd) {
- return true;
- }
- }
- return false;
- }
-
- private static boolean precedingWordIsVerb(SyntacticGroup group,
- Sentence sentence) {
- int precedingTokenPosition = group.getSentencePositionStart() - 1;
- if(isPartOfPrepNG(group, sentence)) {
- SyntacticGroup parentGroup = getParentPrepNG(group, sentence);
- precedingTokenPosition = parentGroup.getSentencePositionStart() - 1;
- }
-
- for (SyntacticWord word : sentence.getSyntacticWords()) {
- int lastWordPosition = word.getSentencePositionEnd();
- if (precedingTokenPosition == lastWordPosition &&
- (word.getCtag().equals("Verbfin") || word.getCtag().equals("Inf"))) {
- return true;
- }
- }
- return false;
- }
-
- // czy sie w lemacie bedzie zgodne miedzy walentym a spejdem?
- // czy prep moze sie skladac z wiecej niz jednego segmentu?
- // dopasowywac refla i recip do sie spejdowego
- private static boolean precedingVerbModifyPrepNG(SyntacticGroup NGGroup,
- SyntacticGroup PrepNGGroup, Sentence sentence,
- Map<String,ArrayList<String>> walentyMapping) {
- int precedingTokenPosition = NGGroup.getSentencePositionStart() - 1;
- for (SyntacticWord word : sentence.getSyntacticWords()) {
- int lastWordPosition = word.getSentencePositionEnd();
- if (precedingTokenPosition == lastWordPosition &&
- (word.getCtag().equals("Verbfin") || word.getCtag().equals("Inf"))) {
- String verb = word.getBase();
- if (!walentyMapping.containsKey(verb)) {
- return true;
- } else {
- SyntacticWord prepWord = PrepNGGroup.getFirstWord();
-
- if (prepWord.getTokens().size() == 1) {
- Token prep = prepWord.getTokens().get(0);
- String prepBase = prep.getBase();
- // sprawdzic czy glowa moze miec wiele tokenow
- String prepCase = PrepNGGroup.getSemanticHeadTokens().get(0).getCase();
- ArrayList<String> prepnps = getPrepnps(prepBase, prepCase);
-
- ArrayList<String> schemata = walentyMapping.get(verb);
- for (String schema : schemata) {
- for (String prepnp : prepnps) {
- if (schema.contains(prepnp)) {
- return true;
- }
- }
- }
- } else if (prepWord.getTokens().size() > 1) {
- String prepOrth = prepWord.getOrth().toLowerCase();
- String comprepnp = String.format("comprepnp(%s)", prepOrth);
- ArrayList<String> schemata = walentyMapping.get(verb);
- for (String schema : schemata) {
- if (schema.contains(comprepnp)) {
- return true;
- }
- }
-
- }
-
-
- }
- }
}
- return false;
- }
-
- private static boolean precedingVerbModifyNG(SyntacticGroup NGGroup,
- SyntacticGroup PrepNGGroup, Sentence sentence,
- Map<String,ArrayList<String>> walentyMapping) {
- int precedingTokenPosition = NGGroup.getSentencePositionStart() - 1;
- if(isPartOfPrepNG(NGGroup, sentence)) {
- SyntacticGroup parentNGGroup = getParentPrepNG(NGGroup, sentence);
- precedingTokenPosition = parentNGGroup.getSentencePositionStart() - 1;
- }
- for (SyntacticWord word : sentence.getSyntacticWords()) {
- int lastWordPosition = word.getSentencePositionEnd();
- if (precedingTokenPosition == lastWordPosition &&
- (word.getCtag().equals("Verbfin") || word.getCtag().equals("Inf"))) {
- if (verbModifyNG(word, NGGroup, PrepNGGroup, sentence, walentyMapping)) {
- return true;
- }
- if (!walentyMapping.containsKey(word.getBase())) {
- return true;
- }
-
- }
- }
- return false;
- }
-
- private static boolean verbModifyNG(SyntacticWord verb, SyntacticGroup NGGroup,
- SyntacticGroup PrepNGGroup, Sentence sentence,
- Map<String,ArrayList<String>> walentyMapping) {
- String verbBase = verb.getBase();
- if (!walentyMapping.containsKey(verbBase)) {
- return true;
- } else {
- ArrayList<String> schemata = walentyMapping.get(verbBase);
-
- // PrepNG + PrepNG
- if (isPartOfPrepNG(NGGroup, sentence)) {
- SyntacticGroup NGParentGroup = getParentPrepNG(NGGroup, sentence);
- ArrayList<String> prepNG1Realizations = NGParentGroup.getWalentyRealizations();
- ArrayList<String> prepNG2Realizations = PrepNGGroup.getWalentyRealizations();
- for (String schema : schemata) {
- if (isProperSchema(schema, prepNG1Realizations, prepNG2Realizations)) {
- return true;
- }
- }
- }
-
- // NG + PrepNG
- else {
- ArrayList<String> NGRealizations = NGGroup.getWalentyRealizations();
- ArrayList<String> prepNGRealizations = PrepNGGroup.getWalentyRealizations();
- for (String schema : schemata) {
- if (isProperSchema(schema, NGRealizations, prepNGRealizations)) {
- return true;
- }
- }
- }
- }
- return false;
}
private static boolean isProperSchema(String schema, ArrayList<String> group1Types,
ArrayList<String> group2Types) {
for (String group1Type : group1Types) {
- if (schema.contains(group1Type)) {
+ if (schemaContains(schema, group1Type)) {
for (String group2Type : group2Types) {
- if (schema.contains(group2Type)) {
+ if (schemaContains(schema, group2Type)) {
return true;
}
}
@@ -404,103 +169,71 @@ public class Detector {
return false;
}
- private static SyntacticGroup getParentPrepNG(SyntacticGroup NGGroup,
- Sentence sentence) {
- SyntacticGroup parentPrepNG = null;
- int NGGroupStart = NGGroup.getSentencePositionStart();
- int NGGroupEnd = NGGroup.getSentencePositionEnd();
- for (SyntacticGroup group : sentence.getGroups()) {
- if (group.getType().startsWith("PrepNG") &&
- group.getSentencePositionStart() <= NGGroupStart &&
- group.getSentencePositionEnd() >= NGGroupEnd) {
- if (parentPrepNG == null || group.getTokens().size() > parentPrepNG.getTokens().size()) {
- parentPrepNG = group;
- }
- }
- }
- return parentPrepNG;
- }
-
- private static boolean NGPrepNGValenceCompatibility(SyntacticGroup NGGroup,
- SyntacticGroup PrepNGGroup, Sentence sentence,
+ private static boolean groupsValenceCompatibility(SyntacticGroup NG1,
+ SyntacticGroup NG2, Sentence sentence,
Map<String,ArrayList<String>> walentyMapping) {
- Token NGHead = NGGroup.getSemanticHeadTokens().get(0);
+ Token NG1Head = NG1.getSemanticHeadTokens().get(0);
- String NGHeadBase = NGHead.getBase();
+ String NGHeadBase = NG1Head.getBase();
if (!walentyMapping.containsKey(NGHeadBase)) {
return false;
} else {
- SyntacticWord prepWord = PrepNGGroup.getFirstWord();
+ ArrayList<String> NG2realizations = NG2.getWalentyRealizations();
- if (prepWord.getTokens().size() == 1) {
- Token prep = prepWord.getTokens().get(0);
- String prepBase = prep.getBase();
- String prepCase = PrepNGGroup.getSemanticHeadTokens().get(0).getCase();
- String prepnp = String.format("prepnp(%s,%s)", prepBase, prepCase);
- ArrayList<String> schemata = walentyMapping.get(NGHeadBase);
- for (String schema : schemata) {
- if (schemaContains(schema, prepnp)) {
- return true;
- }
- }
- } else if (prepWord.getTokens().size() > 1) {
- String prepOrth = prepWord.getOrth().toLowerCase();
- String comprepnp = String.format("comprepnp(%s)", prepOrth);
- ArrayList<String> schemata = walentyMapping.get(NGHeadBase);
+ ArrayList<String> schemata = walentyMapping.get(NGHeadBase);
+ for (String real : NG2realizations) {
for (String schema : schemata) {
- if (schemaContains(schema, comprepnp)) {
+ if (schemaContains(schema, real)) {
return true;
}
}
-
}
-
}
return false;
}
- private static boolean NGNGValenceCompatibility(SyntacticGroup NG1,
- SyntacticGroup NG2, Sentence sentence,
+ private static boolean tripleCompatibility(SyntacticGroup group1,
+ SyntacticGroup group2, SyntacticGroup group3,
Map<String,ArrayList<String>> walentyMapping) {
- Token NG1Head = NG1.getSemanticHeadTokens().get(0);
+ Token group1Head = group1.getSemanticHeadTokens().get(0);
- String NGHeadBase = NG1Head.getBase();
+ String group1HeadBase = group1Head.getBase();
- if (!walentyMapping.containsKey(NGHeadBase)) {
+ if (!walentyMapping.containsKey(group1HeadBase)) {
return false;
} else {
- ArrayList<String> NG2realizations = NG2.getWalentyRealizations();
+ ArrayList<String> group2realizations = group2.getWalentyRealizations();
+ ArrayList<String> group3realizations = group3.getWalentyRealizations();
- ArrayList<String> schemata = walentyMapping.get(NGHeadBase);
- for (String real : NG2realizations) {
- for (String schema : schemata) {
- if (schemaContains(schema, real)) {
- return true;
- }
+ ArrayList<String> schemata = walentyMapping.get(group1HeadBase);
+ for (String schema : schemata) {
+ if (isProperSchema(schema, group2realizations, group3realizations)) {
+ return true;
}
}
}
return false;
}
- private static boolean NGcomplexPrepNGValenceCompatibility(SyntacticGroup NGGroup1,
- SyntacticGroup NGGroup2, Sentence sentence,
+ private static boolean quatroCompatibility(SyntacticGroup group1,
+ SyntacticGroup group2, SyntacticGroup group3, SyntacticGroup group4,
Map<String,ArrayList<String>> walentyMapping) {
-
- Token NGHead = NGGroup1.getSemanticHeadTokens().get(0);
- String NGHeadBase = NGHead.getBase();
+ Token group1Head = group1.getSemanticHeadTokens().get(0);
+
+ String group1HeadBase = group1Head.getBase();
- if (!walentyMapping.containsKey(NGHeadBase)) {
+ if (!walentyMapping.containsKey(group1HeadBase)) {
return false;
} else {
- int prepStart = NGGroup1.getSentencePositionEnd() + 1;
- int prepEnd = NGGroup2.getSentencePositionStart() - 1;
- String complexPrep = sentence.getTextInsideSpan(prepStart, prepEnd);
- String comprepnp = String.format("comprepnp(%s)", complexPrep);
- ArrayList<String> schemata = walentyMapping.get(NGHeadBase);
+ ArrayList<String> group2realizations = group2.getWalentyRealizations();
+ ArrayList<String> group3realizations = group3.getWalentyRealizations();
+ ArrayList<String> group4realizations = group4.getWalentyRealizations();
+
+ ArrayList<String> schemata = walentyMapping.get(group1HeadBase);
for (String schema : schemata) {
- if (schemaContains(schema, comprepnp)) {
+ if (isTripleProperSchema(schema, group2realizations, group3realizations,
+ group4realizations)) {
return true;
}
}
@@ -508,67 +241,119 @@ public class Detector {
return false;
}
- private static boolean schemaContains(String schema, String phraseType) {
- for (String position : schema.split("\\s\\+\\s")) {
- position = position.trim();
- position = position.substring(1, position.length()-1);
- for (String phrT : position.split(";")) {
- if (phrT.equals(phraseType)) {
- return true;
+ private static boolean isTripleProperSchema(String schema, ArrayList<String> group1Types,
+ ArrayList<String> group2Types, ArrayList<String> group3Types) {
+ for (String group1Type : group1Types) {
+ if (schemaContains(schema, group1Type)) {
+ for (String group2Type : group2Types) {
+ if (schemaContains(schema, group2Type)) {
+ for (String group3Type : group3Types) {
+ if (schemaContains(schema, group3Type)) {
+ return true;
+ }
+ }
+ }
}
}
}
return false;
}
- private static boolean schemaContainsType(String schema, String type) {
- // to lepiej dziala dla rzeczownikow
- for (String position : schema.split("\\s\\+\\s")) {
- position = position.trim();
- position = position.substring(1, position.length()-1);
- for (String phrT : position.split(";")) {
-
- if (phrT.startsWith(type+"(")) {
- return true;
+ /*private static boolean isTripleProperSchema(String schema, ArrayList<String> group1Types,
+ ArrayList<String> group2Types, ArrayList<String> group3Types) {
+
+ ArrayList<String> group1MPositions = getMatchingPositions(schema, group1Types);
+ ArrayList<String> group2MPositions = getMatchingPositions(schema, group2Types);
+ ArrayList<String> group3MPositions = getMatchingPositions(schema, group3Types);
+
+
+
+ ArrayList<String> group1MPositionsCopy = new ArrayList<String>();
+ ArrayList<String> group2MPositionsCopy = getMatchingPositions(schema, group2Types);
+ ArrayList<String> group3MPositionsCopy = getMatchingPositions(schema, group3Types);
+
+
+ if (group1MPositions.isEmpty() || group2MPositions.isEmpty() || group3MPositions.isEmpty()) {
+ return false;
+ }
+
+ boolean group1ok = false;
+ boolean group2ok = false;
+ boolean group3ok = false;
+
+ for (String pos : group1MPositions) {
+
+ }
+
+ ArrayList<String>
+
+ if (union(group1MPositions, group2MPositions).size() > group1MPositions.size() &&
+ )
+
+
+ for (String group1Type : group1Types) {
+ if (schemaContains(schema, group1Type)) {
+ for (String group2Type : group2Types) {
+ if (schemaContains(schema, group2Type)) {
+ for (String group3Type : group3Types) {
+ if (schemaContains(schema, group3Type)) {
+ return true;
+ }
+ }
+ }
}
}
}
return false;
+ }*/
+
+ public static List<String> union(List<String> list1, List<String> list2) {
+ HashSet<String> set = new HashSet<String>();
+
+ set.addAll(list1);
+ set.addAll(list2);
+
+ return new ArrayList<String>(set);
}
+ public static List<String> tripleUnion(List<String> list1, List<String> list2,
+ List<String> list3) {
+ HashSet<String> set = new HashSet<String>();
+
+ set.addAll(list1);
+ set.addAll(list2);
+ set.addAll(list3);
+
+ return new ArrayList<String>(set);
+ }
- // compar ??
- private static ArrayList<String> getPrepnps(String prepBase, String prepCase) {
- ArrayList<String> prepnps = new ArrayList<String>();
- prepnps.add(String.format("prepnp(%s,%s)", prepBase, prepCase));
- if (prepCase.equals("nom") || prepCase.equals("gen") || prepCase.equals("acc")) {
- prepnps.add(String.format("prepnp(%s,str)", prepBase));
- }
- if (prepCase.equals("gen") || prepCase.equals("acc")) {
- prepnps.add(String.format("prepnp(%s,part)", prepBase));
+ private static ArrayList<String> getMatchingPositions(String schema, ArrayList<String> phraseRealizations) {
+ ArrayList<String> positions = new ArrayList<String>();
+ for (String position : schema.split("\\s\\+\\s")) {
+ position = position.trim();
+ position = position.substring(1, position.length()-1);
+ for (String phrT : position.split(";")) {
+ if (phraseRealizations.contains(phrT.trim())) {
+ positions.add(position);
+ break;
+ }
+ }
}
- return prepnps;
+ return positions;
}
- // eliminuje "od wsi do wsi"
- private static boolean sameSemanticHeads(SyntacticGroup group1,
- SyntacticGroup group2) {
-
- List<Token> group1HeadTokens = group1.getSemanticHeadTokens();
- List<Token> group2HeadTokens = group2.getSemanticHeadTokens();
- if (group1HeadTokens.size() != group2HeadTokens.size()) {
- return false;
- }
-
- for (int i=0; i < group1HeadTokens.size(); i++) {
- if (!group1HeadTokens.get(i).getBase().equals(group2HeadTokens.get(i).getBase())) {
- return false;
+ private static boolean schemaContains(String schema, String phraseType) {
+ for (String position : schema.split("\\s\\+\\s")) {
+ position = position.trim();
+ position = position.substring(1, position.length()-1);
+ for (String phrT : position.split(";")) {
+ if (phrT.equals(phraseType)) {
+ return true;
+ }
}
}
-
- return true;
+ return false;
}
-
/**
* Wyszukuję i oznaczam wszystkie NER
diff --git a/src/main/java/pl/waw/ipipan/zil/core/md/entities/Mention.java b/src/main/java/pl/waw/ipipan/zil/core/md/entities/Mention.java
index 2fe1e86..120d94b 100644
--- a/src/main/java/pl/waw/ipipan/zil/core/md/entities/Mention.java
+++ b/src/main/java/pl/waw/ipipan/zil/core/md/entities/Mention.java
@@ -1,11 +1,13 @@
package pl.waw.ipipan.zil.core.md.entities;
import java.util.ArrayList;
-import java.util.Arrays;
import java.util.List;
+import pl.waw.ipipan.zil.core.md.detection.Constants;
+
/**
* @author Mateusz Kopec
+ * Modified 2017 by Bartlomiej Niton
*
*/
public class Mention implements Comparable<Mention> {
@@ -205,77 +207,22 @@ public class Mention implements Comparable<Mention> {
return isZeroSubject;
}
- public int getSentencePositionStart() {
+ public int getSentenceStartPosition() {
Token startToken = this.getFirstSegment();
return startToken.getSentencePosition();
}
- public int getSentencePositionEnd() {
+ public int getSentenceEndPosition() {
Token endToken = this.getLastSegment();
return endToken.getSentencePosition();
}
-
- public boolean isPartOfQub() {
- if (this.segments.size() == 1) {
- Sentence sentence = this.segments.get(0).getSentence();
- for (SyntacticWord word : sentence.getSyntacticWords()) {
- if (word.getTokens().contains(this.segments.get(0)) &&
- word.getCtag().equals("Qub")) {
- return true;
- }
- }
- }
- return false;
- }
-
- public boolean isPartOfPrep() {
- if (this.segments.size() == 1) {
- Sentence sentence = this.segments.get(0).getSentence();
- for (SyntacticWord word : sentence.getSyntacticWords()) {
- if (word.getTokens().contains(this.segments.get(0)) &&
- word.getCtag().equals("Prep")) {
- return true;
- }
- }
- }
- return false;
- }
-
- private final List<String> FRAZEOS = Arrays.asList("Prep", "Qub", "Adv", "Interj",
- "Adj", "Conj", "Comp");
public boolean isPartOfFrazeo() {
if (this.segments.size() == 1) {
Sentence sentence = this.segments.get(0).getSentence();
for (SyntacticWord word : sentence.getSyntacticWords()) {
if (word.getTokens().contains(this.segments.get(0)) &&
- FRAZEOS.contains(word.getCtag())) {
- return true;
- }
- }
- }
- return false;
- }
-
- public boolean isPartOfComplexPrep(ArrayList<String> complexPreps) {
- if (this.segments.size() == 1) {
- Sentence sentence = this.segments.get(0).getSentence();
- if (this.getSentencePositionStart() - 1 >= 0) {
- String prep = sentence.get(this.getSentencePositionStart() - 1).getOrth();
- String noun = sentence.get(this.getSentencePositionStart()).getOrth();
- String possiblePrep = String.format("%s %s", prep, noun);
- if (complexPreps.contains(possiblePrep)) {
- return true;
- }
- }
-
- if (this.getSentencePositionStart() - 1 >= 0 &&
- this.getSentencePositionStart() + 1 < sentence.size()) {
- String prep1 = sentence.get(this.getSentencePositionStart() - 1).getOrth();
- String noun = sentence.get(this.getSentencePositionStart()).getOrth();
- String prep2 = sentence.get(this.getSentencePositionStart() + 1).getOrth();
- String possiblePrep = String.format("%s %s %s", prep1, noun, prep2);
- if (complexPreps.contains(possiblePrep)) {
+ Constants.FRAZEO_CTAGS.contains(word.getCtag())) {
return true;
}
}
diff --git a/src/main/java/pl/waw/ipipan/zil/core/md/entities/Sentence.java b/src/main/java/pl/waw/ipipan/zil/core/md/entities/Sentence.java
index 476cdf9..558a71c 100644
--- a/src/main/java/pl/waw/ipipan/zil/core/md/entities/Sentence.java
+++ b/src/main/java/pl/waw/ipipan/zil/core/md/entities/Sentence.java
@@ -110,35 +110,6 @@ public class Sentence extends ArrayList<Token> {
namedEntities.add(namedEntity);
}
- public ArrayList<SyntacticGroup> getGroupsInsideSpan(int start, int end) {
- ArrayList<SyntacticGroup> groupsAtSpan = new ArrayList<SyntacticGroup>();
- for (SyntacticGroup group : this.syntacticGroups) {
- if (group.getSentencePositionStart() >= start &&
- group.getSentencePositionEnd() <= end) {
- if (!(group.getSentencePositionStart() == start &&
- group.getSentencePositionEnd() == end)) {
- groupsAtSpan.add(group);
- }
- }
- }
- return groupsAtSpan;
- }
-
- public ArrayList<SyntacticGroup> getLargestNotIntersectingGroupsInsideSpan(int start, int end) {
- ArrayList<SyntacticGroup> groupsAtSpan = new ArrayList<SyntacticGroup>();
- for (SyntacticGroup group : this.syntacticGroups) {
-
- if (group.getSentencePositionStart() >= start &&
- group.getSentencePositionEnd() <= end) {
- if (!(group.getSentencePositionStart() == start &&
- group.getSentencePositionEnd() == end)) {
- groupsAtSpan.add(group);
- }
- }
- }
- return groupsAtSpan;
- }
-
public SyntacticGroup getFirstGroup(int start, int end) {
SyntacticGroup largestGroup = null;
int step = start;
@@ -152,8 +123,8 @@ public class Sentence extends ArrayList<Token> {
private SyntacticGroup getLargestGroupOnStartPoint(int start, int end) {
SyntacticGroup largestGroup = null;
for (SyntacticGroup group : this.getGroups()) {
- int groupStart = group.getSentencePositionStart();
- int groupEnd = group.getSentencePositionEnd();
+ int groupStart = group.getSentenceStartPosition();
+ int groupEnd = group.getSentenceEndPosition();
if (groupStart == start && groupEnd <= end &&
!(groupStart == start && groupEnd == end) &&
(largestGroup == null ||
@@ -177,8 +148,8 @@ public class Sentence extends ArrayList<Token> {
private SyntacticGroup getLargestGroupOnEndPoint(int start, int end) {
SyntacticGroup largestGroup = null;
for (SyntacticGroup group : this.getGroups()) {
- int groupStart = group.getSentencePositionStart();
- int groupEnd = group.getSentencePositionEnd();
+ int groupStart = group.getSentenceStartPosition();
+ int groupEnd = group.getSentenceEndPosition();
if (groupEnd == end && groupStart >= start &&
!(groupStart == start && groupEnd == end) &&
(largestGroup == null ||
@@ -189,38 +160,4 @@ public class Sentence extends ArrayList<Token> {
return largestGroup;
}
- public ArrayList<Mention> getMentionsInsideSpan(int start, int end) {
- ArrayList<Mention> mentionsAtSpan = new ArrayList<Mention>();
- for (Mention mention : this.mentions) {
- if (mention.getSentencePositionStart() >= start &&
- mention.getSentencePositionEnd() <= end) {
- mentionsAtSpan.add(mention);
- }
- }
- return mentionsAtSpan;
- }
-
- public String getTextInsideSpan(int start, int end) {
- String text = "";
- int step = start;
- while (step <= end) {
- if (step != start) {
- text += " ";
- }
- text += this.get(step).getOrth();
- step++;
- }
- return text;
- }
-
- public ArrayList<Token> getSegmentsInsideSpan(int start, int end) {
- ArrayList<Token> tokensAtSpan = new ArrayList<Token>();
- int step = start;
- while (step <= end) {
- tokensAtSpan.add(this.get(step));
- step++;
- }
- return tokensAtSpan;
- }
-
}
diff --git a/src/main/java/pl/waw/ipipan/zil/core/md/entities/SyntacticGroup.java b/src/main/java/pl/waw/ipipan/zil/core/md/entities/SyntacticGroup.java
index ed6f234..b10fa3d 100644
--- a/src/main/java/pl/waw/ipipan/zil/core/md/entities/SyntacticGroup.java
+++ b/src/main/java/pl/waw/ipipan/zil/core/md/entities/SyntacticGroup.java
@@ -55,34 +55,19 @@ public class SyntacticGroup implements Comparable<SyntacticGroup> {
return getType().compareTo(o.getType());
}
- public int getSentencePositionStart() {
+ public int getSentenceStartPosition() {
Token startToken = tokens.get(0);
return startToken.getSentencePosition();
}
- public int getSentencePositionEnd() {
+ public int getSentenceEndPosition() {
Token endToken = tokens.get(tokens.size()-1);
return endToken.getSentencePosition();
}
-
- public SyntacticWord getFirstWord() {
- SyntacticWord firstWord = null;
- Token startToken = tokens.get(0);
- Sentence sentence = startToken.getSentence();
- for (SyntacticWord word : sentence.getSyntacticWords()) {
- if(startToken.compareTo(word.getTokens().get(0)) == 0 &&
- (firstWord == null || firstWord.getTokens().size() < word.getTokens().size())) {
- firstWord = word;
- }
- }
- return firstWord;
- }
-
- // NG and PrepNG only now
public ArrayList<String> getWalentyRealizations() {
ArrayList<String> realizations = new ArrayList<String>();
- if (this.type.startsWith("PrepNG")) {
+ if (this.type.equals("PrepNG")) {
SyntacticWord prepWord = this.getFirstWord();
if (prepWord.getTokens().size() == 1) {
@@ -105,7 +90,19 @@ public class SyntacticGroup implements Comparable<SyntacticGroup> {
return realizations;
}
- // compar ??
+ public SyntacticWord getFirstWord() {
+ SyntacticWord firstWord = null;
+ Token startToken = tokens.get(0);
+ Sentence sentence = startToken.getSentence();
+ for (SyntacticWord word : sentence.getSyntacticWords()) {
+ if(startToken.compareTo(word.getTokens().get(0)) == 0 &&
+ (firstWord == null || firstWord.getTokens().size() < word.getTokens().size())) {
+ firstWord = word;
+ }
+ }
+ return firstWord;
+ }
+
private ArrayList<String> getPrepnps(String prepBase, String prepCase) {
ArrayList<String> prepnps = new ArrayList<String>();
prepnps.add(String.format("prepnp(%s,%s)", prepBase, prepCase));
@@ -130,44 +127,13 @@ public class SyntacticGroup implements Comparable<SyntacticGroup> {
return nps;
}
- public boolean precedingWordIsVerb() {
- Sentence sentence = this.tokens.get(0).getSentence();
- int precedingTokenPosition = this.getSentencePositionStart() - 1;
- for (SyntacticWord word : sentence.getSyntacticWords()) {
- int lastWordPosition = word.getSentencePositionEnd();
- if (precedingTokenPosition == lastWordPosition &&
- (word.getCtag().equals("Verbfin") || word.getCtag().equals("Inf"))) {
- return true;
- }
- }
- return false;
- }
-
- public SyntacticGroup getNextNG() {
- Sentence sentence = this.tokens.get(0).getSentence();
- int thisGroupEnd = this.getSentencePositionEnd();
- int sentenceLength = sentence.size();
-
- SyntacticGroup nextNG = null;
- for (int step = thisGroupEnd; step < sentenceLength; step++) {
- nextNG = sentence.getFirstGroup(step, sentenceLength);
- if (nextNG != null && nextNG.type.startsWith("NG") &&
- this.getSentencePositionEnd() < nextNG.getSentencePositionStart()) {
- break;
- } else {
- nextNG = null;
- }
- }
- return nextNG;
- }
-
public SyntacticGroup getFollowingGroup() {
SyntacticGroup largestGroup = null;
Sentence sentence = this.tokens.get(0).getSentence();
- int nextTokenPosition = this.getSentencePositionEnd() + 1;
+ int nextTokenPosition = this.getSentenceEndPosition() + 1;
for (SyntacticGroup group : sentence.getGroups()) {
- if ((group.getType().startsWith("PrepNG") || group.getType().startsWith("NG")) &&
- group.getSentencePositionStart() == nextTokenPosition) {
+ if ((group.getType().equals("PrepNG") || group.getType().startsWith("NG")) &&
+ group.getSentenceStartPosition() == nextTokenPosition) {
if (largestGroup == null ||
largestGroup.getTokens().size() < group.getTokens().size()) {
largestGroup = group;
@@ -178,14 +144,14 @@ public class SyntacticGroup implements Comparable<SyntacticGroup> {
}
public SyntacticWord getPrecedingVerb() {
- int precedingTokenPosition = this.getSentencePositionStart() - 1;
+ int precedingTokenPosition = this.getSentenceStartPosition() - 1;
Sentence sentence = this.tokens.get(0).getSentence();
if(this.isPartOfPrepNG()) {
SyntacticGroup parentNGGroup = this.getParentPrepNG();
- precedingTokenPosition = parentNGGroup.getSentencePositionStart() - 1;
+ precedingTokenPosition = parentNGGroup.getSentenceStartPosition() - 1;
}
for (SyntacticWord word : sentence.getSyntacticWords()) {
- int lastWordPosition = word.getSentencePositionEnd();
+ int lastWordPosition = word.getSentenceEndPosition();
if (precedingTokenPosition == lastWordPosition &&
(word.getCtag().equals("Verbfin") || word.getCtag().equals("Inf"))) {
return word;
@@ -195,13 +161,13 @@ public class SyntacticGroup implements Comparable<SyntacticGroup> {
}
private boolean isPartOfPrepNG() {
- int NGGroupStart = this.getSentencePositionStart();
- int NGGroupEnd = this.getSentencePositionEnd();
+ int NGGroupStart = this.getSentenceStartPosition();
+ int NGGroupEnd = this.getSentenceEndPosition();
Sentence sentence = this.tokens.get(0).getSentence();
for (SyntacticGroup group : sentence.getGroups()) {
- if (group.getType().startsWith("PrepNG") &&
- group.getSentencePositionStart() <= NGGroupStart &&
- group.getSentencePositionEnd() >= NGGroupEnd) {
+ if (group.getType().equals("PrepNG") &&
+ group.getSentenceStartPosition() <= NGGroupStart &&
+ group.getSentenceEndPosition() >= NGGroupEnd) {
return true;
}
}
@@ -210,13 +176,13 @@ public class SyntacticGroup implements Comparable<SyntacticGroup> {
private SyntacticGroup getParentPrepNG() {
SyntacticGroup parentPrepNG = null;
- int NGGroupStart = this.getSentencePositionStart();
- int NGGroupEnd = this.getSentencePositionEnd();
+ int NGGroupStart = this.getSentenceStartPosition();
+ int NGGroupEnd = this.getSentenceEndPosition();
Sentence sentence = this.tokens.get(0).getSentence();
for (SyntacticGroup group : sentence.getGroups()) {
- if (group.getType().startsWith("PrepNG") &&
- group.getSentencePositionStart() <= NGGroupStart &&
- group.getSentencePositionEnd() >= NGGroupEnd) {
+ if (group.getType().equals("PrepNG") &&
+ group.getSentenceStartPosition() <= NGGroupStart &&
+ group.getSentenceEndPosition() >= NGGroupEnd) {
if (parentPrepNG == null || group.getTokens().size() > parentPrepNG.getTokens().size()) {
parentPrepNG = group;
}
diff --git a/src/main/java/pl/waw/ipipan/zil/core/md/entities/SyntacticWord.java b/src/main/java/pl/waw/ipipan/zil/core/md/entities/SyntacticWord.java
index aa80dec..14178b6 100644
--- a/src/main/java/pl/waw/ipipan/zil/core/md/entities/SyntacticWord.java
+++ b/src/main/java/pl/waw/ipipan/zil/core/md/entities/SyntacticWord.java
@@ -4,6 +4,8 @@ import java.util.ArrayList;
import java.util.Iterator;
import java.util.List;
+import pl.waw.ipipan.zil.core.md.detection.Constants;
+
public class SyntacticWord implements Comparable<SyntacticWord> {
private String base;
@@ -22,6 +24,14 @@ public class SyntacticWord implements Comparable<SyntacticWord> {
public String getCtag() {
return ctag;
}
+
+ public String getBase() {
+ return base;
+ }
+
+ public String getOrth() {
+ return orth;
+ }
public List<Token> getTokens() {
return tokens;
@@ -45,33 +55,18 @@ public class SyntacticWord implements Comparable<SyntacticWord> {
return getCtag().compareTo(o.getCtag());
}
- public int getSentencePositionStart() {
+ public int getSentenceStartPosition() {
Token startToken = tokens.get(0);
return startToken.getSentencePosition();
}
- public int getSentencePositionEnd() {
+ public int getSentenceEndPosition() {
Token endToken = tokens.get(tokens.size()-1);
return endToken.getSentencePosition();
}
- public String getBase() {
- return this.base;
- }
-
- public String getOrth() {
- return this.orth;
- }
-
public boolean isVerb() {
- if (this.ctag.equals("Verbfin") || this.ctag.equals("Inf")) {
- return true;
- }
- return false;
- }
-
- public boolean isInterp() {
- if (this.ctag.equals("Interp")) {
+ if (Constants.VERB_CTAGS.contains(this.ctag)) {
return true;
}
return false;
diff --git a/src/main/java/pl/waw/ipipan/zil/core/md/io/tei/TeiLoader.java b/src/main/java/pl/waw/ipipan/zil/core/md/io/tei/TeiLoader.java
index 255a056..e3216a5 100644
--- a/src/main/java/pl/waw/ipipan/zil/core/md/io/tei/TeiLoader.java
+++ b/src/main/java/pl/waw/ipipan/zil/core/md/io/tei/TeiLoader.java
@@ -70,7 +70,6 @@ public class TeiLoader {
for (TEIMorph mo : m.getHeadMorphs())
headTokens.add(teiMorph2Segment.get(mo));
s.addMention(new Mention(tokens, headTokens, m.isZeroSubject()));
- System.out.println(tokens.toString());
}
private static void loadSyntacticGroup(Sentence s, TEIGroup g,
--
libgit2 0.22.2