This directory contains a module that implements the "Metaphone" code as
a PostgreSQL user-defined function. The Metaphone system is a method of matching similar sounding names (or any words) to the same code. Metaphone was invented by Lawrence Philips as an improvement to the popular name-hashing routine, Soundex. This metaphone code is from Michael Kuhn, and is detailed at http://aspell.sourceforge.net/metaphone/metaphone-kuhn.txt Joel Burton
This commit is contained in:
Bruce Momjian
parent
2c6373fafa
commit
8ace5114dc
|
|
@ -1,4 +1,4 @@
|
|||
# $Header: /cvsroot/pgsql/contrib/Makefile,v 1.18 2001/03/14 00:57:43 tgl Exp $
|
||||
# $Header: /cvsroot/pgsql/contrib/Makefile,v 1.19 2001/05/09 23:00:44 momjian Exp $
|
||||
|
||||
subdir = contrib
|
||||
top_builddir = ..
|
||||
|
|
@ -15,6 +15,7 @@ WANTED_DIRS = \
|
|||
lo \
|
||||
mSQL-interface \
|
||||
mac \
|
||||
metaphone \
|
||||
miscutil \
|
||||
noupdate \
|
||||
oid2name \
|
||||
|
|
|
|||
|
|
@ -72,6 +72,10 @@ mac -
|
|||
Support functions for MAC address types
|
||||
by Lawrence E. Rosenman <ler@lerctr.org>
|
||||
|
||||
metaphone -
|
||||
Improved Soundex function
|
||||
by Joel Burton <jburton@scw.org>
|
||||
|
||||
miscutil -
|
||||
PostgreSQL assert checking and various utility functions
|
||||
by Massimo Dal Zotto <dz@cs.unitn.it>
|
||||
|
|
|
|||
|
|
@ -0,0 +1,39 @@
|
|||
#
|
||||
# $Header: /cvsroot/pgsql/contrib/metaphone/Attic/Makefile,v 1.1 2001/05/09 23:00:44 momjian Exp $
|
||||
#
|
||||
|
||||
subdir = contrib/metaphone
|
||||
top_builddir = ../..
|
||||
include $(top_builddir)/src/Makefile.global
|
||||
|
||||
NAME := metaphone
|
||||
SONAME := $(NAME)$(DLSUFFIX)
|
||||
|
||||
override CPPFLAGS += -I$(srcdir)
|
||||
override CFLAGS += $(CFLAGS_SL)
|
||||
|
||||
all: $(SONAME) $(NAME).sql
|
||||
|
||||
$(NAME).sql: $(NAME).sql.in
|
||||
sed 's,@MODULE_FILENAME@,$(libdir)/contrib/$(SONAME),g' $< >$@
|
||||
|
||||
install: all installdirs
|
||||
$(INSTALL_SHLIB) $(SONAME) $(libdir)/contrib
|
||||
$(INSTALL_DATA) $(NAME).sql $(datadir)/contrib
|
||||
$(INSTALL_DATA) README.$(NAME) $(docdir)/contrib
|
||||
|
||||
installdirs:
|
||||
$(mkinstalldirs) $(libdir)/contrib $(datadir)/contrib $(docdir)/contrib
|
||||
|
||||
uninstall:
|
||||
rm -f $(libdir)/contrib/$(SONAME) $(datadir)/contrib/$(NAME).sql $(docdir)/contrib/README.$(NAME)
|
||||
|
||||
clean distclean maintainer-clean:
|
||||
rm -f $(SONAME) $(NAME).o $(NAME).sql
|
||||
|
||||
depend dep:
|
||||
$(CC) -MM -MG $(CFLAGS) *.c > depend
|
||||
|
||||
ifeq (depend,$(wildcard depend))
|
||||
include depend
|
||||
endif
|
||||
|
|
@ -0,0 +1,79 @@
|
|||
This directory contains a module that implements the "Metaphone" code as
|
||||
a PostgreSQL user-defined function. The Metaphone system is a method of
|
||||
matching similar sounding names (or any words) to the same code.
|
||||
|
||||
Metaphone was invented by Lawrence Philips as an improvement to the popular
|
||||
name-hashing routine, Soundex.
|
||||
|
||||
This metaphone code is from Michael Kuhn, and is detailed at
|
||||
http://aspell.sourceforge.net/metaphone/metaphone-kuhn.txt
|
||||
|
||||
Code for this (including this help file!) was liberally borrowed from
|
||||
the soundex() module for PostgreSQL.
|
||||
|
||||
There are two functions:
|
||||
metaphone(text) : returns hash of a name
|
||||
metaphone(text,int) : returns hash (maximum length of int) of name
|
||||
|
||||
---
|
||||
|
||||
To install it, first configure the main source tree, then run make;
|
||||
make install in this directory. Finally, load the function definition
|
||||
with psql:
|
||||
|
||||
psql -f PREFIX/share/contrib/metaphone.sql
|
||||
|
||||
The following are some usage examples:
|
||||
|
||||
SELECT text_metaphone('hello world!');
|
||||
SELECT text_metaphone('hello world!', 4);
|
||||
|
||||
CREATE TABLE s (nm text)\g
|
||||
|
||||
insert into s values ('john')\g
|
||||
insert into s values ('joan')\g
|
||||
insert into s values ('wobbly')\g
|
||||
|
||||
select * from s
|
||||
where text_metaphone(nm) = text_metaphone('john')\g
|
||||
|
||||
select nm from s a, s b
|
||||
where text_metaphone(a.nm) = text_metaphone(b.nm)
|
||||
and a.oid <> b.oid\g
|
||||
|
||||
CREATE FUNCTION text_mp_eq(text, text) RETURNS bool AS
|
||||
'select text_metaphone($1) = text_metaphone($2)'
|
||||
LANGUAGE 'sql'\g
|
||||
|
||||
CREATE FUNCTION text_mp_lt(text,text) RETURNS bool AS
|
||||
'select text_metaphone($1) < text_metaphone($2)'
|
||||
LANGUAGE 'sql'\g
|
||||
|
||||
CREATE FUNCTION text_mp_gt(text,text) RETURNS bool AS
|
||||
'select text_metaphone($1) > text_metaphone($2)'
|
||||
LANGUAGE 'sql';
|
||||
|
||||
CREATE FUNCTION text_mp_le(text,text) RETURNS bool AS
|
||||
'select text_metaphone($1) <= text_metaphone($2)'
|
||||
LANGUAGE 'sql';
|
||||
|
||||
CREATE FUNCTION text_mp_ge(text,text) RETURNS bool AS
|
||||
'select text_metaphone($1) >= text_metaphone($2)'
|
||||
LANGUAGE 'sql';
|
||||
|
||||
CREATE FUNCTION text_mp_ne(text,text) RETURNS bool AS
|
||||
'select text_metaphone($1) <> text_metaphone($2)'
|
||||
LANGUAGE 'sql';
|
||||
|
||||
DROP OPERATOR #= (text,text)\g
|
||||
|
||||
CREATE OPERATOR #= (leftarg=text, rightarg=text, procedure=text_mp_eq,
|
||||
commutator=text_mp_eq)\g
|
||||
|
||||
SELECT *
|
||||
FROM s
|
||||
WHERE text_mp_eq(nm,'pillsbury')\g
|
||||
|
||||
SELECT *
|
||||
from s
|
||||
where s.nm #= 'pillsbury';
|
||||
|
|
@ -0,0 +1,321 @@
|
|||
|
||||
#include "postgres.h"
|
||||
#include "fmgr.h"
|
||||
#include "utils/builtins.h"
|
||||
|
||||
#include <stdio.h>
|
||||
#include <string.h>
|
||||
#include <ctype.h>
|
||||
|
||||
Datum text_metaphone(PG_FUNCTION_ARGS);
|
||||
Datum text_metaphone_length(PG_FUNCTION_ARGS);
|
||||
|
||||
void phonetic(char *name, char *metaph, int metalen);
|
||||
|
||||
#define METAPHONE_LEN 50
|
||||
#undef METAPHONE_TEST
|
||||
|
||||
#define _textin(str) DirectFunctionCall1(textin, CStringGetDatum(str))
|
||||
#define _textout(str) DatumGetPointer(DirectFunctionCall1(textout, PointerGetDatum(str)))
|
||||
|
||||
#define NULLCHAR (char *) 0
|
||||
|
||||
char *VOWELS="AEIOU",
|
||||
*FRONTV="EIY", /* special cases for letters in FRONT of these */
|
||||
*VARSON="CSPTG", /* variable sound--those modified by adding an "h" */
|
||||
*DOUBLE="."; /* let these double letters through */
|
||||
|
||||
char *excpPAIR="AGKPW", /* exceptions "ae-", "gn-", "kn-", "pn-", "wr-" */
|
||||
*nextLTR ="ENNNR";
|
||||
char *chrptr, *chrptr1;
|
||||
|
||||
void phonetic(name,metaph,metalen)
|
||||
char *name, *metaph;
|
||||
int metalen;
|
||||
{
|
||||
|
||||
int ii, jj, silent, hard, Lng, lastChr;
|
||||
|
||||
char curLtr, prevLtr, nextLtr, nextLtr2, nextLtr3;
|
||||
|
||||
int vowelAfter, vowelBefore, frontvAfter;
|
||||
|
||||
char wname[60];
|
||||
char *ename=wname;
|
||||
|
||||
jj = 0;
|
||||
for (ii=0; name[ii] != '\0'; ii++) {
|
||||
if ( isalpha(name[ii]) ) {
|
||||
ename[jj] = toupper(name[ii]);
|
||||
jj++;
|
||||
}
|
||||
}
|
||||
ename[jj] = '\0';
|
||||
|
||||
if (strlen(ename) == 0) return;
|
||||
|
||||
/* if ae, gn, kn, pn, wr then drop the first letter */
|
||||
if ( (chrptr=strchr(excpPAIR,ename[0]) ) != NULLCHAR ) {
|
||||
chrptr1 = nextLTR + (chrptr-excpPAIR);
|
||||
if ( *chrptr1 == ename[1] ) strcpy(ename,&ename[1]);
|
||||
}
|
||||
/* change x to s */
|
||||
if (ename[0] == 'X') ename[0] = 'S';
|
||||
/* get rid of the "h" in "wh" */
|
||||
if ( strncmp(ename,"WH",2) == 0 ) strcpy(&ename[1], &ename[2]);
|
||||
|
||||
Lng = strlen(ename);
|
||||
lastChr = Lng -1; /* index to last character in string makes code easier*/
|
||||
|
||||
/* Remove an S from the end of the string */
|
||||
if ( ename[lastChr] == 'S' ) {
|
||||
ename[lastChr] = '\0';
|
||||
Lng = strlen(ename);
|
||||
lastChr = Lng -1;
|
||||
}
|
||||
|
||||
for (ii=0; ( (strlen(metaph) < metalen) && (ii < Lng) ); ii++) {
|
||||
|
||||
curLtr = ename[ii];
|
||||
|
||||
vowelBefore = FALSE; prevLtr = ' ';
|
||||
if (ii > 0) {
|
||||
prevLtr = ename[ii-1];
|
||||
if ( strchr(VOWELS,prevLtr) != NULLCHAR ) vowelBefore = TRUE;
|
||||
}
|
||||
/* if first letter is a vowel KEEP it */
|
||||
if (ii == 0 && (strchr(VOWELS,curLtr) != NULLCHAR) ) {
|
||||
strncat(metaph,&curLtr,1);
|
||||
continue;
|
||||
}
|
||||
|
||||
vowelAfter = FALSE; frontvAfter = FALSE; nextLtr = ' ';
|
||||
if ( ii < lastChr ) {
|
||||
nextLtr = ename[ii+1];
|
||||
if ( strchr(VOWELS,nextLtr) != NULLCHAR ) vowelAfter = TRUE;
|
||||
if ( strchr(FRONTV,nextLtr) != NULLCHAR ) frontvAfter = TRUE;
|
||||
}
|
||||
/* skip double letters except ones in list */
|
||||
if (curLtr == nextLtr && (strchr(DOUBLE,nextLtr) == NULLCHAR) ) continue;
|
||||
|
||||
nextLtr2 = ' ';
|
||||
if (ii < (lastChr-1) ) nextLtr2 = ename[ii+2];
|
||||
|
||||
nextLtr3 = ' ';
|
||||
if (ii < (lastChr-2) ) nextLtr3 = ename[ii+3];
|
||||
|
||||
switch (curLtr) {
|
||||
|
||||
case 'B': silent = FALSE;
|
||||
if (ii == lastChr && prevLtr == 'M') silent = TRUE;
|
||||
if (! silent) strncat(metaph,&curLtr,1);
|
||||
break;
|
||||
|
||||
/*silent -sci-,-sce-,-scy-; sci-, etc OK*/
|
||||
case 'C': if (! (ii > 1 && prevLtr == 'S' && frontvAfter) ) {
|
||||
|
||||
if ( ii > 0 && nextLtr == 'I' && nextLtr2 == 'A' )
|
||||
strncat(metaph,"X",1);
|
||||
else
|
||||
if (frontvAfter)
|
||||
strncat(metaph,"S",1);
|
||||
else
|
||||
if (ii > 1 && prevLtr == 'S' && nextLtr == 'H')
|
||||
strncat(metaph,"K",1);
|
||||
else
|
||||
if (nextLtr == 'H')
|
||||
if (ii == 0 && (strchr(VOWELS,nextLtr2) == NULLCHAR) )
|
||||
strncat(metaph,"K",1);
|
||||
else
|
||||
strncat(metaph,"X",1);
|
||||
else
|
||||
if (prevLtr == 'C')
|
||||
strncat(metaph,"C",1);
|
||||
else
|
||||
strncat(metaph,"K",1);
|
||||
}
|
||||
break;
|
||||
|
||||
case 'D': if (nextLtr == 'G' && (strchr(FRONTV,nextLtr2) != NULLCHAR))
|
||||
strncat(metaph,"J",1);
|
||||
else
|
||||
strncat(metaph,"T",1);
|
||||
break;
|
||||
|
||||
case 'G': silent=FALSE;
|
||||
/* SILENT -gh- except for -gh and no vowel after h */
|
||||
if ( (ii < (lastChr-1) && nextLtr == 'H')
|
||||
&& (strchr(VOWELS,nextLtr2) == NULLCHAR) )
|
||||
silent=TRUE;
|
||||
|
||||
if ( (ii == (lastChr-3) )
|
||||
&& nextLtr == 'N' && nextLtr2 == 'E' && nextLtr3 == 'D')
|
||||
silent=TRUE;
|
||||
else
|
||||
if ( (ii == (lastChr-1)) && nextLtr == 'N') silent=TRUE;
|
||||
|
||||
if (prevLtr == 'D' && frontvAfter) silent=TRUE;
|
||||
|
||||
if (prevLtr == 'G')
|
||||
hard=TRUE;
|
||||
else
|
||||
hard=FALSE;
|
||||
|
||||
if (!silent) {
|
||||
if (frontvAfter && (! hard) )
|
||||
strncat(metaph,"J",1);
|
||||
else
|
||||
strncat(metaph,"K",1);
|
||||
}
|
||||
break;
|
||||
|
||||
case 'H': silent = FALSE;
|
||||
if ( strchr(VARSON,prevLtr) != NULLCHAR ) silent = TRUE;
|
||||
|
||||
if ( vowelBefore && !vowelAfter) silent = TRUE;
|
||||
|
||||
if (!silent) strncat(metaph,&curLtr,1);
|
||||
break;
|
||||
|
||||
case 'F':
|
||||
case 'J':
|
||||
case 'L':
|
||||
case 'M':
|
||||
case 'N':
|
||||
case 'R': strncat(metaph,&curLtr,1);
|
||||
break;
|
||||
|
||||
case 'K': if (prevLtr != 'C') strncat(metaph,&curLtr,1);
|
||||
break;
|
||||
|
||||
case 'P': if (nextLtr == 'H')
|
||||
strncat(metaph,"F",1);
|
||||
else
|
||||
strncat(metaph,"P",1);
|
||||
break;
|
||||
|
||||
case 'Q': strncat(metaph,"K",1);
|
||||
break;
|
||||
|
||||
case 'S': if (ii > 1 && nextLtr == 'I'
|
||||
&& ( nextLtr2 == 'O' || nextLtr2 == 'A') )
|
||||
strncat(metaph,"X",1);
|
||||
else
|
||||
if (nextLtr == 'H')
|
||||
strncat(metaph,"X",1);
|
||||
else
|
||||
strncat(metaph,"S",1);
|
||||
break;
|
||||
|
||||
case 'T': if (ii > 1 && nextLtr == 'I'
|
||||
&& ( nextLtr2 == 'O' || nextLtr2 == 'A') )
|
||||
strncat(metaph,"X",1);
|
||||
else
|
||||
if (nextLtr == 'H') /* The=0, Tho=T, Withrow=0 */
|
||||
if (ii > 0 || (strchr(VOWELS,nextLtr2) != NULLCHAR) )
|
||||
strncat(metaph,"0",1);
|
||||
else
|
||||
strncat(metaph,"T",1);
|
||||
else
|
||||
if (! (ii < (lastChr-2) && nextLtr == 'C' && nextLtr2 == 'H'))
|
||||
strncat(metaph,"T",1);
|
||||
break;
|
||||
|
||||
case 'V': strncat(metaph,"F",1);
|
||||
break;
|
||||
|
||||
case 'W':
|
||||
case 'Y': if (ii < lastChr && vowelAfter) strncat(metaph,&curLtr,1);
|
||||
break;
|
||||
|
||||
case 'X': strncat(metaph,"KS",2);
|
||||
break;
|
||||
|
||||
case 'Z': strncat(metaph,"S",1);
|
||||
break;
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
/* DON'T DO THIS NOW, REMOVING "S" IN BEGINNING HAS the same effect
|
||||
with plurals, in addition imbedded S's in the Metaphone are included
|
||||
Lng = strlen(metaph);
|
||||
lastChr = Lng -1;
|
||||
if ( metaph[lastChr] == 'S' && Lng >= 3 ) metaph[lastChr] = '\0';
|
||||
*/
|
||||
|
||||
return;
|
||||
}
|
||||
|
||||
|
||||
#ifdef METAPHONE_TEST
|
||||
int
|
||||
main(int argc, char *argv[])
|
||||
{
|
||||
if (argc < 2)
|
||||
{
|
||||
fprintf(stderr, "usage: %s string\n", argv[0]);
|
||||
return 1;
|
||||
}
|
||||
else
|
||||
{
|
||||
char output[51]="";
|
||||
|
||||
phonetic(argv[1], output, 50);
|
||||
printf("metaphone(%s) = %s\n", argv[1], output);
|
||||
return 0;
|
||||
}
|
||||
}
|
||||
|
||||
#endif /* METAPHONE_TEST */
|
||||
|
||||
#ifndef METAPHONE_TEST
|
||||
/*
|
||||
* SQL function: text_metaphone(text) returns text
|
||||
*/
|
||||
PG_FUNCTION_INFO_V1(text_metaphone);
|
||||
|
||||
Datum
|
||||
text_metaphone(PG_FUNCTION_ARGS)
|
||||
{
|
||||
char outstr[51]="";
|
||||
char *arg;
|
||||
|
||||
arg = _textout(PG_GETARG_TEXT_P(0));
|
||||
|
||||
phonetic(arg, outstr, 50);
|
||||
|
||||
PG_RETURN_TEXT_P(_textin(outstr));
|
||||
}
|
||||
|
||||
/*
|
||||
char outstr[51]="";
|
||||
char *arg;
|
||||
int32 metalen;
|
||||
|
||||
arg = _textout(PG_GETARG_TEXT_P(0));
|
||||
metalen = PG_GETARG_INT32(1);
|
||||
|
||||
|
||||
phonetic(arg, outstr, metalen);
|
||||
*/
|
||||
|
||||
PG_FUNCTION_INFO_V1(text_metaphone_length);
|
||||
|
||||
Datum
|
||||
text_metaphone_length(PG_FUNCTION_ARGS)
|
||||
{
|
||||
char outstr[51]="";
|
||||
char *arg;
|
||||
int32 metalen;
|
||||
|
||||
arg = _textout(PG_GETARG_TEXT_P(0));
|
||||
metalen = PG_GETARG_INT32(1);
|
||||
|
||||
phonetic(arg, outstr, metalen);
|
||||
|
||||
PG_RETURN_TEXT_P(_textin(outstr));
|
||||
}
|
||||
|
||||
|
||||
#endif /* not METAPHONE_TEST */
|
||||
|
|
@ -0,0 +1,3 @@
|
|||
CREATE FUNCTION text_soundex(text) RETURNS text
|
||||
AS '@MODULE_FILENAME@', 'text_metaphone' LANGUAGE 'C';
|
||||
|
||||
|
|
@ -56,7 +56,10 @@ LDFLAGS=--shared -Wl,-E -Wl,-soname,$@
|
|||
|
||||
.PHONY: clean
|
||||
|
||||
all: plpython.so
|
||||
all:
|
||||
@echo "Disabled until merged into our Makefile system, bjm 2001-05-09"
|
||||
|
||||
disabled: plpython.so
|
||||
|
||||
plpython.o: plpython.c plpython.h
|
||||
$(CC) $(CFLAGS) $(INCLUDES) -c -o $@ $<
|
||||
|
|
|
|||
Loading…
Reference in New Issue