POLICY_NOUN_PHRASES

Provides the ability to extract the noun phrases along with part-of-speech information for each word in each noun phrase from a given document.

For example, consider the following sentence:

“The mayor of Chicago is giving a brief press conference.”

The noun phrases for this input are “mayor of Chicago” and “brief press conference.” The subgroups in the input text are not returned. For instance, in the above example, subgroups such as “mayor, Chicago, brief press, press conference, press, conference” are not returned.

All AUTO_LEXER languages are supported for POLICY_NOUN_PHRASES and POLICY_PART_OF_SPEECHPOLICY_NOUN_PHRASES.

Syntax

ctx_doc.policy_noun_phrases (
   policy_name    in varchar2,
   document       in varchar2 | CLOB,
   restab         in out nocopy noun_phrase_tab,
   language       in varchar2  default NULL,
   format         in varchar2  default NULL,
   charset        in varchar2  default NULL
);

policy_name

Specify the policy name created with CTX_DDL.CREATE_POLICY.

document

A body of text for which the languages are to be extracted. The text is assumed to be plain text with UTF-8 character encoding.

restab

Specify the name of the CLOB locator.

language

Specify the language. See the list of supported languages in this section. If this parameter is null, the language will be automatically detected. There is a cost associated with language detection.

format

The format of the input text.

charset

The character set of the input text.

Abbreviations for Use with POLICY_NOUN_PHRASES and POLICY_PART_OF_SPEECH

This is a list of abbreviations that you can use in queries for POLICY_NOUN_PHRASES and POLICY_PART_OF_SPEECH. The examples use these abbreviations.

Table 1 Part of Speech Abbreviations

Abbreviation	Part of Speech
N	noun
propN	nounProper
V	verb
Adj	adjective
Adv	adverb
Prep	preposition
Part	particle
Punct	punct
Pro	pronoun
Wh	interrog
Det	determiner
Conj	conjunction
Card	numCardinal
Ord	numOrdinal
Suf	suffix
Pre	prefix
Acr	nounAcronym
Poss	poss
Unk	unknown

Example for POLICY__NOUN_PHRASES

The example in this section uses the abbreviations shown in the preceding table.

set serverout on
create or replace function toString(b boolean) return varchar2 is
    begin
      if (b) then
        return 'TRUE';
      end if;
      return 'FALSE';
 end;
 /

declare
 the_nps ctx_doc.noun_phrase_tab;
begin
  ctx_ddl.create_preference('rvlex', 'AUTO_LEXER');
  ctx_ddl.set_attribute('rvlex','mixed_case','YES');
  ctx_ddl.set_attribute('rvlex','timeout',0);

  ctx_ddl.create_policy(policy_name => 'rv_policy_21',lexer => 'rvlex');

 ctx_doc.policy_noun_phrases('rv_policy_21','The mayor of Chicago is giving a
 brief press conference',the_nps);
 dbms_output.put_line(the_nps.count);

 for i in 1..the_nps.count loop
      if (the_nps(i).is_phrase_start) then
        if (i>1) then
          dbms_output.put(']');
          dbms_output.new_line;
        end if;
        dbms_output.put('Phrase{term,POS,is_in_lex,offset,len,is_phrase_
        start}:[');
      else
        dbms_output.put(',');
      end if;
      dbms_output.put('{' || the_nps(i).term || ',' || the_nps(i).pos_tag || ','
      || toString(the_nps(i).is_in_lexicon) || ',' || the_nps(i).offset
      || ',' || the_nps(i).length || ',' || toString(the_nps(i).is_phrase_start)
      || '}');
      end loop;
      dbms_output.put(']');
      dbms_output.new_line;
end;
/

Output for this example:

Phrase{term,POS,is_in_lex,offset,len,is_phrase_start}:
[{The,Det,TRUE,1,3,TRUE},{mayor,N,TRUE,5,5,FALSE},
{of,Prep,TRUE,11,2,FALSE},{Chicago,propN,TRUE,14,7,FALSE}

Phrase{term,POS,is_in_lex,offset,len,is_phrase_start}:
[{a,Det,TRUE,32,1,TRUE},{brief,N,TRUE,34,5,FALSE},
{press,N,TRUE,40,5,FALSE},{conference,N,TRUE,46,10,FALSE}]

Related Topics

• AUTO_LEXER Language Support

• CREATE_POLICY

• POLICY_PART_OF_SPEECH