use of edu.stanford.nlp.util.CoreMap in project CoreNLP by stanfordnlp.
the class TokenSequenceMatcherITest method testMultiplePatterns.
public void testMultiplePatterns() throws IOException {
TokenSequencePattern p1 = TokenSequencePattern.compile("(?$dt \"atropine\") []{0,15} " + "(?$se \"social\" \"avoidant\" \"behaviour\"|\"dysuria\"|\"hyperglycaemia\"| \"mental\" \"disorder\"|\"vertigo\"|\"flutter\"| \"chest\" \"pain\"| \"elevated\" \"blood\" \"pressure\"|\"mania\"| \"rash\" \"erythematous\"|\"manic\"| \"papular\" \"rash\"|\"death\"| \"atrial\" \"arrhythmia\"| \"dry\" \"eyes\"| \"loss\" \"of\" \"libido\"| \"rash\" \"papular\"|\"hypersensitivity\"| \"blood\" \"pressure\" \"increased\"|\"dyspepsia\"| \"accommodation\" \"disorder\"| \"reflexes\" \"increased\"|\"lesions\"|\"asthenia\"| \"gastrointestinal\" \"pain\"|\"excitement\"| \"breast\" \"feeding\"|\"hypokalaemia\"| \"cerebellar\" \"syndrome\"|\"nervousness\"| \"pulmonary\" \"oedema\"| \"inspiratory\" \"stridor\"| \"taste\" \"altered\"|\"paranoia\"| \"psychotic\" \"disorder\"| \"open\" \"angle\" \"glaucoma\"|\"photophobia\"| \"dry\" \"eye\"|\"osteoarthritis\"| \"keratoconjunctivitis\" \"sicca\"| \"haemoglobin\" \"increased\"| \"ventricular\" \"extrasystoles\"|\"hallucinations\"|\"conjunctivitis\"|\"paralysis\"| \"qrs\" \"complex\"|\"anxiety\"| \"conjunctival\" \"disorder\"|\"coma\"|\"strabismus\"|\"thirst\"|\"para\"| \"sicca\" \"syndrome\"| \"atrioventricular\" \"dissociation\"|\"desquamation\"|\"crusting\"| \"abdominal\" \"distension\"|\"blindness\"|\"hypotension\"|\"dermatitis\"| \"sinus\" \"tachycardia\"| \"abdominal\" \"distention\"| \"lacrimation\" \"decreased\"|\"sicca\"| \"paralytic\" \"ileus\"| \"urinary\" \"hesitation\"|\"withdrawn\"| \"erectile\" \"dysfunction\"|\"keratoconjunctivitis\"|\"anaphylaxis\"| \"psychiatric\" \"disorders\"| \"altered\" \"taste\"|\"somnolence\"|\"extrasystoles\"|\"ageusia\"| \"intraocular\" \"pressure\" \"increased\"| \"left\" \"ventricular\" \"failure\"|\"impotence\"|\"drowsiness\"|\"conjunctiva\"| \"delayed\" \"gastric\" \"emptying\"| \"gastrointestinal\" \"sounds\" \"abnormal\"| \"qt\" \"prolonged\"| \"supraventricular\" \"tachycardia\"|\"weakness\"|\"hypertonia\"| \"confusional\" \"state\"|\"anhidrosis\"|\"myopia\"|\"dyspnoea\"| \"speech\" \"impairment\" \"nos\"| \"rash\" \"maculo\" \"papular\"|\"petechiae\"|\"tachypnea\"| \"acute\" \"angle\" \"closure\" \"glaucoma\"| \"gastrooesophageal\" \"reflux\" \"disease\"|\"hypokalemia\"| \"left\" \"heart\" \"failure\"| \"myocardial\" \"infarction\"| \"site\" \"reaction\"| \"ventricular\" \"fibrillation\"|\"fibrillation\"| \"maculopapular\" \"rash\"| \"impaired\" \"gastric\" \"emptying\"|\"amnesia\"| \"labored\" \"respirations\"| \"decreased\" \"lacrimation\"|\"mydriasis\"|\"headache\"| \"dry\" \"mouth\"|\"scab\"| \"cardiac\" \"syncope\"| \"visual\" \"acuity\" \"reduced\"|\"tension\"| \"blurred\" \"vision\"| \"bloated\" \"feeling\"| \"labored\" \"breathing\"| \"stridor\" \"inspiratory\"| \"skin\" \"exfoliation\"| \"memory\" \"loss\"|\"syncope\"| \"rash\" \"scarlatiniform\"|\"hyperpyrexia\"| \"cardiac\" \"flutter\"|\"heartburn\"| \"bowel\" \"sounds\" \"decreased\"|\"blepharitis\"|\"tachycardia\"| \"excessive\" \"thirst\"|\"confusion\"| \"rash\" \"macular\"| \"taste\" \"loss\"| \"respiratory\" \"failure\"|\"hesitancy\"|\"dysmetria\"|\"disorientation\"| \"decreased\" \"hemoglobin\"| \"atrial\" \"fibrillation\"| \"urinary\" \"retention\"| \"dry\" \"skin\"|\"dehydration\"|\"hyponatraemia\"|\"dysgeusia\"|\"disorder\"| \"increased\" \"intraocular\" \"pressure\"| \"speech\" \"disorder\"| \"feeling\" \"abnormal\"|\"pain\"| \"anaphylactic\" \"shock\"|\"hallucination\"| \"abdominal\" \"pain\"| \"junctional\" \"tachycardia\"| \"bun\" \"increased\"| \"ventricular\" \"flutter\"| \"scarlatiniform\" \"rash\"|\"agitation\"| \"feeling\" \"hot\"|\"hyponatremia\"| \"decreased\" \"bowel\" \"sounds\"|\"cyanosis\"|\"dysarthria\"| \"heat\" \"intolerance\"|\"hyperglycemia\"|\"reflux\"| \"angle\" \"closure\" \"glaucoma\"| \"electrocardiogram\" \"qt\" \"prolonged\"| \"vision\" \"blurred\"| \"blood\" \"urea\" \"increased\"|\"dizziness\"|\"arrhythmia\"|\"erythema\"|\"vomiting\"| \"difficulty\" \"in\" \"micturition\"|\"infarction\"|\"laryngospasm\"|\"hypoglycaemia\"|\"hypoglycemia\"| \"elevated\" \"hemoglobin\"| \"skin\" \"warm\"| \"ventricular\" \"arrhythmia\"|\"dissociation\"| \"warm\" \"skin\"| \"follicular\" \"conjunctivitis\"|\"urticaria\"|\"fatigue\"| \"cardiac\" \"fibrillation\"| \"decreased\" \"sweating\"| \"decreased\" \"visual\" \"acuity\"|\"lethargy\"| \"acute\" \"angle\" \"closure\" \"glaucoma\"| \"nodal\" \"rhythm\"|\"borborygmi\"|\"hyperreflexia\"| \"respiratory\" \"depression\"|\"diarrhea\"|\"leukocytosis\"| \"speech\" \"disturbance\"|\"ataxia\"|\"cycloplegia\"|\"tachypnoea\"|\"eczema\"| \"supraventricular\" \"extrasystoles\"|\"ileus\"| \"cardiac\" \"arrest\"| \"ventricular\" \"tachycardia\"|\"laryngitis\"|\"delirium\"|\"lactation\"|\"glaucoma\"|\"obstruction\"|\"hypohidrosis\"|\"parity\"|\"palpitations\"| \"temperature\" \"intolerance\"|\"constipation\"|\"cyclophoria\"| \"acute\" \"coronary\" \"syndrome\"| \"arrhythmia\" \"supraventricular\"|\"arrest\"|\"lesion\"|\"nausea\"| \"sweating\" \"decreased\"|\"keratitis\"|\"dyskinesia\"| \"pulmonary\" \"function\" \"test\" \"decreased\"|\"stridor\"|\"swelling\"|\"dysphagia\"| \"haemoglobin\" \"decreased\"|\"diarrhoea\"| \"ileus\" \"paralytic\"|\"clonus\"|\"insomnia\"| \"electrocardiogram\" \"qrs\" \"complex\"| \"nasal\" \"congestion\"| \"nasal\" \"dryness\"|\"sweating\"|\"rash\"| \"nodal\" \"arrhythmia\"|\"irritability\"|\"hyperhidrosis\"| \"ventricular\" \"failure\")");
TokenSequencePattern p2 = TokenSequencePattern.compile("(?$dt \"disease\") []{0,15} " + "(?$se \"social\" \"avoidant\" \"behaviour\"|\"dysuria\"|\"hyperglycaemia\"| \"mental\" \"disorder\"|\"vertigo\"|\"flutter\"| \"chest\" \"pain\"| \"elevated\" \"blood\" \"pressure\"|\"mania\"| \"rash\" \"erythematous\"|\"manic\"| \"papular\" \"rash\"|\"death\"| \"atrial\" \"arrhythmia\"| \"dry\" \"eyes\"| \"loss\" \"of\" \"libido\"| \"rash\" \"papular\"|\"hypersensitivity\"| \"blood\" \"pressure\" \"increased\"|\"dyspepsia\"| \"accommodation\" \"disorder\"| \"reflexes\" \"increased\"|\"lesions\"|\"asthenia\"| \"gastrointestinal\" \"pain\"|\"excitement\"| \"breast\" \"feeding\"|\"hypokalaemia\"| \"cerebellar\" \"syndrome\"|\"nervousness\"| \"pulmonary\" \"oedema\"| \"inspiratory\" \"stridor\"| \"taste\" \"altered\"|\"paranoia\"| \"psychotic\" \"disorder\"| \"open\" \"angle\" \"glaucoma\"|\"photophobia\"| \"dry\" \"eye\"|\"osteoarthritis\"| \"keratoconjunctivitis\" \"sicca\"| \"haemoglobin\" \"increased\"| \"ventricular\" \"extrasystoles\"|\"hallucinations\"|\"conjunctivitis\"|\"paralysis\"| \"qrs\" \"complex\"|\"anxiety\"| \"conjunctival\" \"disorder\"|\"coma\"|\"strabismus\"|\"thirst\"|\"para\"| \"sicca\" \"syndrome\"| \"atrioventricular\" \"dissociation\"|\"desquamation\"|\"crusting\"| \"abdominal\" \"distension\"|\"blindness\"|\"hypotension\"|\"dermatitis\"| \"sinus\" \"tachycardia\"| \"abdominal\" \"distention\"| \"lacrimation\" \"decreased\"|\"sicca\"| \"paralytic\" \"ileus\"| \"urinary\" \"hesitation\"|\"withdrawn\"| \"erectile\" \"dysfunction\"|\"keratoconjunctivitis\"|\"anaphylaxis\"| \"psychiatric\" \"disorders\"| \"altered\" \"taste\"|\"somnolence\"|\"extrasystoles\"|\"ageusia\"| \"intraocular\" \"pressure\" \"increased\"| \"left\" \"ventricular\" \"failure\"|\"impotence\"|\"drowsiness\"|\"conjunctiva\"| \"delayed\" \"gastric\" \"emptying\"| \"gastrointestinal\" \"sounds\" \"abnormal\"| \"qt\" \"prolonged\"| \"supraventricular\" \"tachycardia\"|\"weakness\"|\"hypertonia\"| \"confusional\" \"state\"|\"anhidrosis\"|\"myopia\"|\"dyspnoea\"| \"speech\" \"impairment\" \"nos\"| \"rash\" \"maculo\" \"papular\"|\"petechiae\"|\"tachypnea\"| \"acute\" \"angle\" \"closure\" \"glaucoma\"| \"gastrooesophageal\" \"reflux\" \"disease\"|\"hypokalemia\"| \"left\" \"heart\" \"failure\"| \"myocardial\" \"infarction\"| \"site\" \"reaction\"| \"ventricular\" \"fibrillation\"|\"fibrillation\"| \"maculopapular\" \"rash\"| \"impaired\" \"gastric\" \"emptying\"|\"amnesia\"| \"labored\" \"respirations\"| \"decreased\" \"lacrimation\"|\"mydriasis\"|\"headache\"| \"dry\" \"mouth\"|\"scab\"| \"cardiac\" \"syncope\"| \"visual\" \"acuity\" \"reduced\"|\"tension\"| \"blurred\" \"vision\"| \"bloated\" \"feeling\"| \"labored\" \"breathing\"| \"stridor\" \"inspiratory\"| \"skin\" \"exfoliation\"| \"memory\" \"loss\"|\"syncope\"| \"rash\" \"scarlatiniform\"|\"hyperpyrexia\"| \"cardiac\" \"flutter\"|\"heartburn\"| \"bowel\" \"sounds\" \"decreased\"|\"blepharitis\"|\"tachycardia\"| \"excessive\" \"thirst\"|\"confusion\"| \"rash\" \"macular\"| \"taste\" \"loss\"| \"respiratory\" \"failure\"|\"hesitancy\"|\"dysmetria\"|\"disorientation\"| \"decreased\" \"hemoglobin\"| \"atrial\" \"fibrillation\"| \"urinary\" \"retention\"| \"dry\" \"skin\"|\"dehydration\"|\"hyponatraemia\"|\"dysgeusia\"|\"disorder\"| \"increased\" \"intraocular\" \"pressure\"| \"speech\" \"disorder\"| \"feeling\" \"abnormal\"|\"pain\"| \"anaphylactic\" \"shock\"|\"hallucination\"| \"abdominal\" \"pain\"| \"junctional\" \"tachycardia\"| \"bun\" \"increased\"| \"ventricular\" \"flutter\"| \"scarlatiniform\" \"rash\"|\"agitation\"| \"feeling\" \"hot\"|\"hyponatremia\"| \"decreased\" \"bowel\" \"sounds\"|\"cyanosis\"|\"dysarthria\"| \"heat\" \"intolerance\"|\"hyperglycemia\"|\"reflux\"| \"angle\" \"closure\" \"glaucoma\"| \"electrocardiogram\" \"qt\" \"prolonged\"| \"vision\" \"blurred\"| \"blood\" \"urea\" \"increased\"|\"dizziness\"|\"arrhythmia\"|\"erythema\"|\"vomiting\"| \"difficulty\" \"in\" \"micturition\"|\"infarction\"|\"laryngospasm\"|\"hypoglycaemia\"|\"hypoglycemia\"| \"elevated\" \"hemoglobin\"| \"skin\" \"warm\"| \"ventricular\" \"arrhythmia\"|\"dissociation\"| \"warm\" \"skin\"| \"follicular\" \"conjunctivitis\"|\"urticaria\"|\"fatigue\"| \"cardiac\" \"fibrillation\"| \"decreased\" \"sweating\"| \"decreased\" \"visual\" \"acuity\"|\"lethargy\"| \"acute\" \"angle\" \"closure\" \"glaucoma\"| \"nodal\" \"rhythm\"|\"borborygmi\"|\"hyperreflexia\"| \"respiratory\" \"depression\"|\"diarrhea\"|\"leukocytosis\"| \"speech\" \"disturbance\"|\"ataxia\"|\"cycloplegia\"|\"tachypnoea\"|\"eczema\"| \"supraventricular\" \"extrasystoles\"|\"ileus\"| \"cardiac\" \"arrest\"| \"ventricular\" \"tachycardia\"|\"laryngitis\"|\"delirium\"|\"lactation\"|\"glaucoma\"|\"obstruction\"|\"hypohidrosis\"|\"parity\"|\"palpitations\"| \"temperature\" \"intolerance\"|\"constipation\"|\"cyclophoria\"| \"acute\" \"coronary\" \"syndrome\"| \"arrhythmia\" \"supraventricular\"|\"arrest\"|\"lesion\"|\"nausea\"| \"sweating\" \"decreased\"|\"keratitis\"|\"dyskinesia\"| \"pulmonary\" \"function\" \"test\" \"decreased\"|\"stridor\"|\"swelling\"|\"dysphagia\"| \"haemoglobin\" \"decreased\"|\"diarrhoea\"| \"ileus\" \"paralytic\"|\"clonus\"|\"insomnia\"| \"electrocardiogram\" \"qrs\" \"complex\"| \"nasal\" \"congestion\"| \"nasal\" \"dryness\"|\"sweating\"|\"rash\"| \"nodal\" \"arrhythmia\"|\"irritability\"|\"hyperhidrosis\"| \"ventricular\" \"failure\")");
CoreMap doc = createDocument("atropine we need to have many many words here but we don't sweating");
MultiPatternMatcher<CoreMap> multiPatternMatcher = TokenSequencePattern.getMultiPatternMatcher(p1, p2);
List<String> expected = new ArrayList<String>();
expected.add("atropine we need to have many many words here but we don't sweating");
Iterator<String> expectedIter = expected.iterator();
Iterable<SequenceMatchResult<CoreMap>> matches = multiPatternMatcher.findAllNonOverlappingMatchesPerPattern(doc.get(CoreAnnotations.TokensAnnotation.class));
for (SequenceMatchResult<CoreMap> match : matches) {
assertEquals(expectedIter.next(), match.group());
}
assertFalse(expectedIter.hasNext());
}
use of edu.stanford.nlp.util.CoreMap in project CoreNLP by stanfordnlp.
the class TokenSequenceMatcherITest method testTokenSequenceMatcherPosNNP.
public void testTokenSequenceMatcherPosNNP() throws IOException {
CoreMap doc = createDocument(testText1);
// Test sequence with groups
TokenSequencePattern p = TokenSequencePattern.compile("[ { tag:\"NNP\" } ]+");
TokenSequenceMatcher m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
boolean match = m.find();
assertTrue(match);
assertEquals(0, m.groupCount());
assertEquals("Mellitus", m.group());
p = TokenSequencePattern.compile("[ { tag:\"NNP\" } ] [ /is|was/ ] []*? [ { tag:\"NNP\" } ]+ ");
m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
match = m.find();
assertTrue(match);
assertEquals(0, m.groupCount());
assertEquals("Mellitus was the first Bishop", m.group());
TokenSequencePattern nnpPattern = TokenSequencePattern.compile("[ { tag:\"NNP\" } ]");
Env env = TokenSequencePattern.getNewEnv();
env.bind("$NNP", nnpPattern);
p = TokenSequencePattern.compile(env, " $NNP [ /is|was/ ] []*? $NNP+ [ \"of\" ] $NNP+ ");
m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
match = m.find();
assertTrue(match);
assertEquals(0, m.groupCount());
assertEquals("Mellitus was the first Bishop of London", m.group());
p = TokenSequencePattern.compile(env, " ($NNP) /is|was/ []*? ($NNP)+ \"of\" ($NNP)+ ");
m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
match = m.find();
assertTrue(match);
assertEquals(3, m.groupCount());
assertEquals("Mellitus was the first Bishop of London", m.group());
assertEquals("Mellitus", m.group(1));
assertEquals("Bishop", m.group(2));
assertEquals("London", m.group(3));
nnpPattern = TokenSequencePattern.compile(" ( [ { tag:\"NNP\" } ] )");
env.bind("$NNP", nnpPattern);
p = TokenSequencePattern.compile(env, " $NNP /is|was/ []*? $NNP+ \"of\" $NNP+ ");
m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
match = m.find();
assertTrue(match);
assertEquals(3, m.groupCount());
assertEquals("Mellitus was the first Bishop of London", m.group());
assertEquals("Mellitus", m.group(1));
assertEquals("Bishop", m.group(2));
assertEquals("London", m.group(3));
// Same as above but without extra "{}"
nnpPattern = TokenSequencePattern.compile(" ( [ tag:\"NNP\" ] )");
env.bind("$NNP", nnpPattern);
p = TokenSequencePattern.compile(env, " $NNP /is|was/ []*? $NNP+ \"of\" $NNP+ ");
m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
match = m.find();
assertTrue(match);
assertEquals(3, m.groupCount());
assertEquals("Mellitus was the first Bishop of London", m.group());
assertEquals("Mellitus", m.group(1));
assertEquals("Bishop", m.group(2));
assertEquals("London", m.group(3));
// Same as above but using "pos"
nnpPattern = TokenSequencePattern.compile(" ( [ pos:\"NNP\" ] )");
env.bind("$NNP", nnpPattern);
p = TokenSequencePattern.compile(env, " $NNP /is|was/ []*? $NNP+ \"of\" $NNP+ ");
m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
match = m.find();
assertTrue(match);
assertEquals(3, m.groupCount());
assertEquals("Mellitus was the first Bishop of London", m.group());
assertEquals("Mellitus", m.group(1));
assertEquals("Bishop", m.group(2));
assertEquals("London", m.group(3));
}
use of edu.stanford.nlp.util.CoreMap in project CoreNLP by stanfordnlp.
the class TokenSequenceMatcherITest method testTokenSequenceMatcher8.
public void testTokenSequenceMatcher8() throws IOException {
CoreMap doc = createDocument(testText1);
// Test sequence with groups
TokenSequencePattern p = TokenSequencePattern.compile("[ /[A-Za-z]+/ ]*");
TokenSequenceMatcher m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
boolean match = m.find();
assertTrue(match);
assertEquals(0, m.groupCount());
assertEquals("Mellitus was the first Bishop of London", m.group());
match = m.find();
assertTrue(match);
assertEquals(0, m.groupCount());
assertEquals("the third Archbishop of Canterbury", m.group());
p = TokenSequencePattern.compile("[ /[A-Za-z]+/ ]* [\"Mellitus\"] [ \"was\"]");
m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
match = m.find();
assertTrue(match);
assertEquals(0, m.groupCount());
assertEquals("Mellitus was", m.group());
match = m.find();
assertFalse(match);
p = TokenSequencePattern.compile("[ /[A-Za-z]+/ ]+ [\"Mellitus\"] [ \"was\"]");
m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
match = m.find();
assertFalse(match);
}
use of edu.stanford.nlp.util.CoreMap in project CoreNLP by stanfordnlp.
the class TokenSequenceMatcherITest method testTokenSequenceMatchesWildcard.
public void testTokenSequenceMatchesWildcard() throws IOException {
CoreMap doc = createDocument("word1 word2");
// Test sequence with groups
TokenSequencePattern p = TokenSequencePattern.compile("[]{2}|[]");
TokenSequenceMatcher m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
boolean matches = m.matches();
assertTrue(matches);
// Reverse order
p = TokenSequencePattern.compile("[]|[]{2}");
m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
matches = m.matches();
assertTrue(matches);
// Using {1,2}
p = TokenSequencePattern.compile("[]{1,2}");
m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
matches = m.matches();
assertTrue(matches);
}
use of edu.stanford.nlp.util.CoreMap in project CoreNLP by stanfordnlp.
the class TokenSequenceMatcherITest method testTokenSequenceMatcherValue.
public void testTokenSequenceMatcherValue() throws IOException {
CoreMap doc = createDocument(testText);
// Test simple sequence with value
TokenSequencePattern p = TokenSequencePattern.compile(getOrPatternExpr(new Pair<String, Object>("one", 1), new Pair<String, Object>("two", null), new Pair<String, Object>("fifty", 50)));
TokenSequenceMatcher m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
boolean match = m.find();
assertTrue(match);
assertEquals("one", m.group());
assertEquals(1, m.groupValue());
match = m.find();
assertTrue(match);
assertEquals("two", m.group());
assertNull(m.groupValue());
match = m.find();
assertTrue(match);
assertEquals("fifty", m.group());
assertEquals(50, m.groupValue());
match = m.find();
assertFalse(match);
}
Aggregations