File: jaro.c

package info (click to toggle)
pg-similarity 1.0-4
  • links: PTS, VCS
  • area: main
  • in suites: bullseye
  • size: 452 kB
  • sloc: ansic: 3,257; sql: 517; makefile: 25; sh: 1
file content (324 lines) | stat: -rw-r--r-- 7,807 bytes parent folder | download | duplicates (4)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
/*----------------------------------------------------------------------------
 *
 * jaro.c
 *
 * Jaro Distance [1] is a similarity measure
 *
 *      1      m      1      m      1     m - t
 * s = --- * ----- + --- * ----- + --- * -------
 *      3     |a|     3     |b|     3       m
 *
 * where m is the number of matching characters [2], t is the number of
 * transpositions [3], |a| is the length of string a and |b| is the length of
 * string b.
 *
 * [2] two characters from a and b are considered matching iif they're not
 * farther than floor(max(|a|, |b|) / 2) - 1.
 *
 * [3] number of transpositions is the number of matchings that are in a
 * different sequence order divided by 2.
 *
 * Jaro-Winkler [4] Distance is a similarity measure
 *
 * It's an improvement over Jaro's original work. It gives more weight if the
 * initial characters are the same. So,
 *
 * w = s + (l * p * (1 - s))
 *
 * where l is the length of common prefix up to 4 characters, p is a scaling
 * factor (Winkler's suggestion is 0.1), and s is the Jaro Distance.
 *
 * For example:
 *
 * x: euler
 * y: heuser
 *
 *      1     4     1     4     1     4 - 0     4      2     1
 * s = --- * --- + --- * --- + --- * ------- = ---- + --- + --- = 0.822...
 *      3     5     3     6     3       4       15     9     3
 *
 *
 * w = 0.822 + (0 * 0.1 * (1 - 0.822)) = 0.822...
 *
 *
 * [1] Jaro, M. A. (1989). "Advances in record linking methodology as applied
 * to the 1985 census of Tampa Florida". Journal of the American Statistical
 * Society 84 (406): 414–20.
 *
 * [4] Winkler, W. E. (2006). "Overview of Record Linkage and Current Research
 * Directions". Research Report Series, RRS.
 * http://www.census.gov/srd/papers/pdf/rrs2006-02.pdf.
 *
 *
 * Copyright (c) 2008-2018, Euler Taveira de Oliveira
 *
 *----------------------------------------------------------------------------
 */

#include "similarity.h"

#include <math.h>

/* GUC variables */
double	pgs_jaro_threshold = 0.7f;
bool	pgs_jaro_is_normalized = true;
double	pgs_jarowinkler_threshold = 0.7f;
bool	pgs_jarowinkler_is_normalized = true;


static double _jaro(char *a, char *b)
{
	int		alen, blen;
	int		i, j, k;

	int		cd;		/* common window distance */
	int		cc = 0;		/* number of common characters */
	int		tr = 0;		/* number of transpositions */
	double	res;
	int		*amatch;	/* matchs in string a; match = 1; unmatch = 0  !! USED? !!*/
	int		*bmatch;	/* matchs in string b; match = 1; unmatch = 0 */
	int		*posa;		/* positions of matched characters in a */
	int		*posb;		/* positions of matched characters in b */

	alen = strlen(a);
	blen = strlen(b);

	elog(DEBUG1, "alen: %d; blen: %d", alen, blen);

	if (alen > PGS_MAX_STR_LEN || blen > PGS_MAX_STR_LEN)
		ereport(ERROR,
				(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
				errmsg("argument exceeds the maximum length of %d bytes",
					PGS_MAX_STR_LEN)));

	/* if one string has zero length then return zero */
	if (alen == 0 || blen == 0)
		return 0.0;

	/*
	 * allocate 2 vectors of integers. each position will be 0 or 1 depending
	 * on the character in that position is found between common window distance.
	 */
	amatch = palloc(sizeof(int) * alen);
	bmatch = palloc(sizeof(int) * blen);

	for (i = 0; i < alen; i++)
		amatch[i] = 0;
	for (j = 0; j < blen; j++)
		bmatch[j] = 0;

	/* common window distance is floor(max(alen, blen) / 2) - 1 */
	cd = (int) floor((alen > blen ? alen : blen) / 2) - 1;
	/* catch case when alen = blen = 1 */
	if (cd < 0)
		cd = 0;

	elog(DEBUG1, "common window distance: %d", cd);

#ifdef PGS_IGNORE_CASE
	elog(DEBUG2, "case-sensitive turns off");
	for (i = 0; i < alen; i++)
		a[i] = tolower(a[i]);
	for (j = 0; j < blen; j++)
		b[j] = tolower(b[j]);
#endif

	for (i = 0; i < alen; i++)
	{
		/*
		 * calculate window test limits. limit inf to 0 and sup to blen 
		 */
		int inf = max2(i - cd, 0);
		int sup = i + cd + 1;
		sup = min2(sup, blen);

		/*
		 * no more common characters 'cause we don't have characters in b
		 * to test with characters in a
		 */
		if (inf >= sup)
			break;

		for (j = inf; j < sup; j++)
		{
			/*
			 * if found some match and it's not matched yet:
			 * (i) flag match characters in a and b
			 * (ii) increment cc
			 */
			if (bmatch[j] != 1 && a[i] == b[j])
			{
				amatch[i] = 1;
				bmatch[j] = 1;
				cc++;

				break;
			}
		}
	}

	elog(DEBUG1, "common characters: %d", cc);

	/* no common characters then return 0 */
	if (cc == 0)
		return 0.0;

	/* allocate vector of positions */
	posa = palloc(sizeof(int) * cc);
	posb = palloc(sizeof(int) * cc);
	
	k = 0;
	for (i = 0; i < alen; i++)
	{
		if (amatch[i] == 1)
		{
			posa[k] = i;
			k++;
		}
	}

	k = 0;
	for (j = 0; j < blen; j++)
	{
		if (bmatch[j] == 1)
		{
			posb[k] = j;
			k++;
		}
	}

	pfree(amatch);
	pfree(bmatch);

	/* counting half-transpositions */
	for (i = 0; i < cc; i++)
		if (a[posa[i]] != b[posb[i]])
			tr++;

	pfree(posa);
	pfree(posb);

	elog(DEBUG1, "half transpositions: %d", tr);

	/* real number of transpositions */
	tr /= 2;

	elog(DEBUG1, "real transpositions: %d", tr);

	res = PGS_JARO_W1 * cc / alen + PGS_JARO_W2 * cc / blen + PGS_JARO_WT * (cc - tr) / cc;

	elog(DEBUG1, "jaro(%s, %s) = %f * %d / %d + %f * %d / %d + %f * (%d - %d) / %d = %f",
			a, b, PGS_JARO_W1, cc, alen, PGS_JARO_W2, cc, blen, PGS_JARO_WT, cc, tr, cc, res);

	return res;
}

PG_FUNCTION_INFO_V1(jaro);

Datum
jaro(PG_FUNCTION_ARGS)
{
	char	*a, *b;
	float8	res;

	a = DatumGetPointer(DirectFunctionCall1(textout, PointerGetDatum(PG_GETARG_TEXT_P(0))));
	b = DatumGetPointer(DirectFunctionCall1(textout, PointerGetDatum(PG_GETARG_TEXT_P(1))));

	res = _jaro(a, b);

	elog(DEBUG1, "is normalized: %d", pgs_jaro_is_normalized);
	elog(DEBUG1, "jaro(%s, %s) = %f", a, b, res);

	/* normalized and unnormalized version are the same */
	PG_RETURN_FLOAT8(res);
}

PG_FUNCTION_INFO_V1(jaro_op);

Datum jaro_op(PG_FUNCTION_ARGS)
{
	float8	res;

	/*
	 * store *_is_normalized value temporarily 'cause
	 * threshold (we're comparing against) is normalized
	 */
	bool	tmp = pgs_jaro_is_normalized;
	pgs_jaro_is_normalized = true;

	res = DatumGetFloat8(DirectFunctionCall2(
					jaro,
					PG_GETARG_DATUM(0),
					PG_GETARG_DATUM(1)));

	/* we're done; back to the previous value */
	pgs_jaro_is_normalized = tmp;

	PG_RETURN_BOOL(res >= pgs_jaro_threshold);
}

PG_FUNCTION_INFO_V1(jarowinkler);

Datum
jarowinkler(PG_FUNCTION_ARGS)
{
	char	*a, *b;
	float8	resj, res;
	int	i;
	int	plen = 0;

	a = DatumGetPointer(DirectFunctionCall1(textout, PointerGetDatum(PG_GETARG_TEXT_P(0))));
	b = DatumGetPointer(DirectFunctionCall1(textout, PointerGetDatum(PG_GETARG_TEXT_P(1))));

	resj = _jaro(a, b);

	res = resj;

	elog(DEBUG1, "jaro(%s, %s) = %f", a, b, resj);

	if (resj > PGS_JARO_BOOST_THRESHOLD)
	{
		for (i = 0; i < strlen(a) && i < strlen(b) && i < PGS_JARO_PREFIX_SIZE; i++)
		{
			if (a[i] == b[i])
				plen++;
			else
				break;
		}

		elog(DEBUG1, "prefix length: %d", plen);

		res += PGS_JARO_SCALING_FACTOR * plen * (1.0 - resj);
	}

	elog(DEBUG1, "is normalized: %d", pgs_jarowinkler_is_normalized);
	elog(DEBUG1, "jarowinkler(%s, %s) = %f + %d * %f * (1.0 - %f) = %f",
			a, b, resj, plen, PGS_JARO_SCALING_FACTOR, resj, res);

	/* normalized and unnormalized version are the same */
	PG_RETURN_FLOAT8(res);
}

PG_FUNCTION_INFO_V1(jarowinkler_op);

Datum jarowinkler_op(PG_FUNCTION_ARGS)
{
	float8	res;

	/*
	 * store *_is_normalized value temporarily 'cause
	 * threshold (we're comparing against) is normalized
	 */
	bool	tmp = pgs_jarowinkler_is_normalized;
	pgs_jarowinkler_is_normalized = true;

	res = DatumGetFloat8(DirectFunctionCall2(
					jarowinkler,
					PG_GETARG_DATUM(0),
					PG_GETARG_DATUM(1)));

	/* we're done; back to the previous value */
	pgs_jarowinkler_is_normalized = tmp;

	PG_RETURN_BOOL(res >= pgs_jarowinkler_threshold);
}