mirror of
https://github.com/pgvector/pgvector.git
synced 2026-07-02 10:40:57 +08:00
Added intvec type
This commit is contained in:
Andrew Kane
@@ -57,7 +57,8 @@
|
||||
|
||||
typedef enum HnswType
|
||||
{
|
||||
HNSW_TYPE_VECTOR
|
||||
HNSW_TYPE_VECTOR,
|
||||
HNSW_TYPE_INTVEC
|
||||
} HnswType;
|
||||
|
||||
/* Build phases */
|
||||
|
||||
@@ -683,6 +683,9 @@ InitBuildState(HnswBuildState * buildstate, Relation heap, Relation index, Index
|
||||
buildstate->efConstruction = HnswGetEfConstruction(index);
|
||||
buildstate->dimensions = TupleDescAttr(index->rd_att, 0)->atttypmod;
|
||||
|
||||
if (buildstate->type == HNSW_TYPE_INTVEC)
|
||||
maxDimensions *= 4;
|
||||
|
||||
/* Require column to have dimensions to be indexed */
|
||||
if (buildstate->dimensions < 0)
|
||||
elog(ERROR, "column does not have dimensions");
|
||||
|
||||
@@ -1,14 +1,17 @@
|
||||
#include "postgres.h"
|
||||
|
||||
#include <float.h>
|
||||
#include <math.h>
|
||||
|
||||
#include "access/generic_xlog.h"
|
||||
#include "catalog/pg_type.h"
|
||||
#include "hnsw.h"
|
||||
#include "lib/pairingheap.h"
|
||||
#include "storage/bufmgr.h"
|
||||
#include "utils/datum.h"
|
||||
#include "utils/memdebug.h"
|
||||
#include "utils/rel.h"
|
||||
#include "utils/syscache.h"
|
||||
#include "vector.h"
|
||||
|
||||
#if PG_VERSION_NUM >= 130000
|
||||
@@ -155,7 +158,24 @@ HnswOptionalProcInfo(Relation index, uint16 procnum)
|
||||
HnswType
|
||||
HnswGetType(Relation index)
|
||||
{
|
||||
return HNSW_TYPE_VECTOR;
|
||||
Oid typeOid = TupleDescAttr(index->rd_att, 0)->atttypid;
|
||||
HeapTuple tuple;
|
||||
Form_pg_type type;
|
||||
int result;
|
||||
|
||||
tuple = SearchSysCache1(TYPEOID, ObjectIdGetDatum(typeOid));
|
||||
if (!HeapTupleIsValid(tuple))
|
||||
elog(ERROR, "cache lookup failed for type %u", typeOid);
|
||||
|
||||
type = (Form_pg_type) GETSTRUCT(tuple);
|
||||
if (strcmp(NameStr(type->typname), "intvec") == 0)
|
||||
result = HNSW_TYPE_INTVEC;
|
||||
else
|
||||
result = HNSW_TYPE_VECTOR;
|
||||
|
||||
ReleaseSysCache(tuple);
|
||||
|
||||
return result;
|
||||
}
|
||||
|
||||
/*
|
||||
@@ -592,7 +612,14 @@ HnswLoadElement(HnswElement element, float *distance, Datum *q, Relation index,
|
||||
if (DatumGetPointer(*q) == NULL)
|
||||
*distance = 0;
|
||||
else
|
||||
{
|
||||
*distance = (float) DatumGetFloat8(FunctionCall2Coll(procinfo, collation, *q, PointerGetDatum(&etup->data)));
|
||||
|
||||
/* Needed for intvec cosine distance */
|
||||
/* TODO Improve */
|
||||
if (isnan(*distance))
|
||||
*distance = FLT_MAX;
|
||||
}
|
||||
}
|
||||
|
||||
UnlockReleaseBuffer(buf);
|
||||
|
||||
586
src/intvec.c
Normal file
586
src/intvec.c
Normal file
@@ -0,0 +1,586 @@
|
||||
#include "postgres.h"
|
||||
|
||||
#include <limits.h>
|
||||
#include <math.h>
|
||||
|
||||
#include "catalog/pg_type.h"
|
||||
#include "fmgr.h"
|
||||
#include "intvec.h"
|
||||
#include "lib/stringinfo.h"
|
||||
#include "libpq/pqformat.h"
|
||||
#include "utils/array.h"
|
||||
#include "utils/builtins.h"
|
||||
#include "utils/lsyscache.h"
|
||||
|
||||
/*
|
||||
* Ensure same dimensions
|
||||
*/
|
||||
static inline void
|
||||
CheckDims(IntVector * a, IntVector * b)
|
||||
{
|
||||
if (a->dim != b->dim)
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_DATA_EXCEPTION),
|
||||
errmsg("different intvec dimensions %d and %d", a->dim, b->dim)));
|
||||
}
|
||||
|
||||
/*
|
||||
* Ensure expected dimensions
|
||||
*/
|
||||
static inline void
|
||||
CheckExpectedDim(int32 typmod, int dim)
|
||||
{
|
||||
if (typmod != -1 && typmod != dim)
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_DATA_EXCEPTION),
|
||||
errmsg("expected %d dimensions, not %d", typmod, dim)));
|
||||
}
|
||||
|
||||
/*
|
||||
* Ensure valid dimensions
|
||||
*/
|
||||
static inline void
|
||||
CheckDim(int dim)
|
||||
{
|
||||
if (dim < 1)
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_DATA_EXCEPTION),
|
||||
errmsg("intvec must have at least 1 dimension")));
|
||||
|
||||
if (dim > INTVEC_MAX_DIM)
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
|
||||
errmsg("intvec cannot have more than %d dimensions", INTVEC_MAX_DIM)));
|
||||
}
|
||||
|
||||
/*
|
||||
* Ensure element in range
|
||||
*/
|
||||
static inline void
|
||||
CheckElement(long value)
|
||||
{
|
||||
if (value < SCHAR_MIN || value > SCHAR_MAX)
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
|
||||
errmsg("value \"%ld\" is out of range for type intvec", value)));
|
||||
}
|
||||
|
||||
/*
|
||||
* Allocate and initialize a new int vector
|
||||
*/
|
||||
IntVector *
|
||||
InitIntVector(int dim)
|
||||
{
|
||||
IntVector *result;
|
||||
int size;
|
||||
|
||||
size = INTVEC_SIZE(dim);
|
||||
result = (IntVector *) palloc0(size);
|
||||
SET_VARSIZE(result, size);
|
||||
result->dim = dim;
|
||||
|
||||
return result;
|
||||
}
|
||||
|
||||
/*
|
||||
* Check for whitespace, since array_isspace() is static
|
||||
*/
|
||||
static inline bool
|
||||
intvec_isspace(char ch)
|
||||
{
|
||||
if (ch == ' ' ||
|
||||
ch == '\t' ||
|
||||
ch == '\n' ||
|
||||
ch == '\r' ||
|
||||
ch == '\v' ||
|
||||
ch == '\f')
|
||||
return true;
|
||||
return false;
|
||||
}
|
||||
|
||||
/*
|
||||
* Convert textual representation to internal representation
|
||||
*/
|
||||
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_in);
|
||||
Datum
|
||||
intvec_in(PG_FUNCTION_ARGS)
|
||||
{
|
||||
char *lit = PG_GETARG_CSTRING(0);
|
||||
int32 typmod = PG_GETARG_INT32(2);
|
||||
int8 x[INTVEC_MAX_DIM];
|
||||
int dim = 0;
|
||||
char *pt;
|
||||
char *stringEnd;
|
||||
IntVector *result;
|
||||
char *litcopy = pstrdup(lit);
|
||||
char *str = litcopy;
|
||||
|
||||
while (intvec_isspace(*str))
|
||||
str++;
|
||||
|
||||
if (*str != '[')
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
|
||||
errmsg("malformed intvec literal: \"%s\"", lit),
|
||||
errdetail("Vector contents must start with \"[\".")));
|
||||
|
||||
str++;
|
||||
pt = strtok(str, ",");
|
||||
stringEnd = pt;
|
||||
|
||||
while (pt != NULL && *stringEnd != ']')
|
||||
{
|
||||
long l;
|
||||
|
||||
if (dim == INTVEC_MAX_DIM)
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
|
||||
errmsg("intvec cannot have more than %d dimensions", INTVEC_MAX_DIM)));
|
||||
|
||||
while (intvec_isspace(*pt))
|
||||
pt++;
|
||||
|
||||
/* Check for empty string like float4in */
|
||||
if (*pt == '\0')
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
|
||||
errmsg("invalid input syntax for type intvec: \"%s\"", lit)));
|
||||
|
||||
/* Use similar logic as int2vectorin */
|
||||
errno = 0;
|
||||
l = strtol(pt, &stringEnd, 10);
|
||||
|
||||
if (stringEnd == pt)
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
|
||||
errmsg("invalid input syntax for type intvec: \"%s\"", lit)));
|
||||
|
||||
if (errno == ERANGE || l < SCHAR_MIN || l > SCHAR_MAX)
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
|
||||
errmsg("value \"%s\" is out of range for type intvec", pt)));
|
||||
|
||||
x[dim++] = l;
|
||||
|
||||
while (intvec_isspace(*stringEnd))
|
||||
stringEnd++;
|
||||
|
||||
if (*stringEnd != '\0' && *stringEnd != ']')
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
|
||||
errmsg("invalid input syntax for type intvec: \"%s\"", lit)));
|
||||
|
||||
pt = strtok(NULL, ",");
|
||||
}
|
||||
|
||||
if (stringEnd == NULL || *stringEnd != ']')
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
|
||||
errmsg("malformed intvec literal: \"%s\"", lit),
|
||||
errdetail("Unexpected end of input.")));
|
||||
|
||||
stringEnd++;
|
||||
|
||||
/* Only whitespace is allowed after the closing brace */
|
||||
while (intvec_isspace(*stringEnd))
|
||||
stringEnd++;
|
||||
|
||||
if (*stringEnd != '\0')
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
|
||||
errmsg("malformed intvec literal: \"%s\"", lit),
|
||||
errdetail("Junk after closing right brace.")));
|
||||
|
||||
/* Ensure no consecutive delimiters since strtok skips */
|
||||
for (pt = lit + 1; *pt != '\0'; pt++)
|
||||
{
|
||||
if (pt[-1] == ',' && *pt == ',')
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
|
||||
errmsg("malformed intvec literal: \"%s\"", lit)));
|
||||
}
|
||||
|
||||
if (dim < 1)
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_DATA_EXCEPTION),
|
||||
errmsg("intvec must have at least 1 dimension")));
|
||||
|
||||
pfree(litcopy);
|
||||
|
||||
CheckExpectedDim(typmod, dim);
|
||||
|
||||
result = InitIntVector(dim);
|
||||
for (int i = 0; i < dim; i++)
|
||||
result->x[i] = x[i];
|
||||
|
||||
PG_RETURN_POINTER(result);
|
||||
}
|
||||
|
||||
/*
|
||||
* Convert internal representation to textual representation
|
||||
*/
|
||||
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_out);
|
||||
Datum
|
||||
intvec_out(PG_FUNCTION_ARGS)
|
||||
{
|
||||
IntVector *vector = PG_GETARG_INTVEC_P(0);
|
||||
int dim = vector->dim;
|
||||
char *buf;
|
||||
char *ptr;
|
||||
int n;
|
||||
|
||||
/*
|
||||
* Need:
|
||||
*
|
||||
* dim * 4 bytes for elements (-128 to 127)
|
||||
*
|
||||
* dim - 1 bytes for separator
|
||||
*
|
||||
* 3 bytes for [, ], and \0
|
||||
*/
|
||||
buf = (char *) palloc(5 * dim + 2);
|
||||
ptr = buf;
|
||||
|
||||
*ptr = '[';
|
||||
ptr++;
|
||||
for (int i = 0; i < dim; i++)
|
||||
{
|
||||
if (i > 0)
|
||||
{
|
||||
*ptr = ',';
|
||||
ptr++;
|
||||
}
|
||||
|
||||
n = pg_ltoa(vector->x[i], ptr);
|
||||
ptr += n;
|
||||
}
|
||||
*ptr = ']';
|
||||
ptr++;
|
||||
*ptr = '\0';
|
||||
|
||||
PG_FREE_IF_COPY(vector, 0);
|
||||
PG_RETURN_CSTRING(buf);
|
||||
}
|
||||
|
||||
/*
|
||||
* Convert type modifier
|
||||
*/
|
||||
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_typmod_in);
|
||||
Datum
|
||||
intvec_typmod_in(PG_FUNCTION_ARGS)
|
||||
{
|
||||
ArrayType *ta = PG_GETARG_ARRAYTYPE_P(0);
|
||||
int32 *tl;
|
||||
int n;
|
||||
|
||||
tl = ArrayGetIntegerTypmods(ta, &n);
|
||||
|
||||
if (n != 1)
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
|
||||
errmsg("invalid type modifier")));
|
||||
|
||||
if (*tl < 1)
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
|
||||
errmsg("dimensions for type intvec must be at least 1")));
|
||||
|
||||
if (*tl > INTVEC_MAX_DIM)
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
|
||||
errmsg("dimensions for type intvec cannot exceed %d", INTVEC_MAX_DIM)));
|
||||
|
||||
PG_RETURN_INT32(*tl);
|
||||
}
|
||||
|
||||
/*
|
||||
* Convert external binary representation to internal representation
|
||||
*/
|
||||
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_recv);
|
||||
Datum
|
||||
intvec_recv(PG_FUNCTION_ARGS)
|
||||
{
|
||||
StringInfo buf = (StringInfo) PG_GETARG_POINTER(0);
|
||||
int32 typmod = PG_GETARG_INT32(2);
|
||||
IntVector *result;
|
||||
int16 dim;
|
||||
int16 unused;
|
||||
|
||||
dim = pq_getmsgint(buf, sizeof(int16));
|
||||
unused = pq_getmsgint(buf, sizeof(int16));
|
||||
|
||||
CheckDim(dim);
|
||||
CheckExpectedDim(typmod, dim);
|
||||
|
||||
if (unused != 0)
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_DATA_EXCEPTION),
|
||||
errmsg("expected unused to be 0, not %d", unused)));
|
||||
|
||||
result = InitIntVector(dim);
|
||||
for (int i = 0; i < dim; i++)
|
||||
result->x[i] = pq_getmsgint(buf, sizeof(int8));
|
||||
|
||||
PG_RETURN_POINTER(result);
|
||||
}
|
||||
|
||||
/*
|
||||
* Convert internal representation to the external binary representation
|
||||
*/
|
||||
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_send);
|
||||
Datum
|
||||
intvec_send(PG_FUNCTION_ARGS)
|
||||
{
|
||||
IntVector *vec = PG_GETARG_INTVEC_P(0);
|
||||
StringInfoData buf;
|
||||
|
||||
pq_begintypsend(&buf);
|
||||
pq_sendint(&buf, vec->dim, sizeof(int16));
|
||||
pq_sendint(&buf, vec->unused, sizeof(int16));
|
||||
for (int i = 0; i < vec->dim; i++)
|
||||
pq_sendint8(&buf, vec->x[i]);
|
||||
|
||||
PG_RETURN_BYTEA_P(pq_endtypsend(&buf));
|
||||
}
|
||||
|
||||
/*
|
||||
* Convert int vector to int vector
|
||||
* This is needed to check the type modifier
|
||||
*/
|
||||
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec);
|
||||
Datum
|
||||
intvec(PG_FUNCTION_ARGS)
|
||||
{
|
||||
IntVector *vec = PG_GETARG_INTVEC_P(0);
|
||||
int32 typmod = PG_GETARG_INT32(1);
|
||||
|
||||
CheckExpectedDim(typmod, vec->dim);
|
||||
|
||||
PG_RETURN_POINTER(vec);
|
||||
}
|
||||
|
||||
/*
|
||||
* Convert array to intvec vector
|
||||
*/
|
||||
PGDLLEXPORT PG_FUNCTION_INFO_V1(array_to_intvec);
|
||||
Datum
|
||||
array_to_intvec(PG_FUNCTION_ARGS)
|
||||
{
|
||||
ArrayType *array = PG_GETARG_ARRAYTYPE_P(0);
|
||||
int32 typmod = PG_GETARG_INT32(1);
|
||||
Vector *result;
|
||||
int16 typlen;
|
||||
bool typbyval;
|
||||
char typalign;
|
||||
Datum *elemsp;
|
||||
int nelemsp;
|
||||
|
||||
if (ARR_NDIM(array) > 1)
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_DATA_EXCEPTION),
|
||||
errmsg("array must be 1-D")));
|
||||
|
||||
if (ARR_HASNULL(array) && array_contains_nulls(array))
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_NULL_VALUE_NOT_ALLOWED),
|
||||
errmsg("array must not contain nulls")));
|
||||
|
||||
get_typlenbyvalalign(ARR_ELEMTYPE(array), &typlen, &typbyval, &typalign);
|
||||
deconstruct_array(array, ARR_ELEMTYPE(array), typlen, typbyval, typalign, &elemsp, NULL, &nelemsp);
|
||||
|
||||
CheckDim(nelemsp);
|
||||
CheckExpectedDim(typmod, nelemsp);
|
||||
|
||||
result = InitVector(nelemsp);
|
||||
|
||||
if (ARR_ELEMTYPE(array) == INT4OID)
|
||||
{
|
||||
for (int i = 0; i < nelemsp; i++)
|
||||
{
|
||||
long l = DatumGetInt32(elemsp[i]);
|
||||
|
||||
CheckElement(l);
|
||||
|
||||
result->x[i] = l;
|
||||
}
|
||||
}
|
||||
else
|
||||
{
|
||||
ereport(ERROR,
|
||||
(errcode(ERRCODE_DATA_EXCEPTION),
|
||||
errmsg("unsupported array type")));
|
||||
}
|
||||
|
||||
/*
|
||||
* Free allocation from deconstruct_array. Do not free individual elements
|
||||
* when pass-by-reference since they point to original array.
|
||||
*/
|
||||
pfree(elemsp);
|
||||
|
||||
PG_RETURN_POINTER(result);
|
||||
}
|
||||
|
||||
/*
|
||||
* Get the L2 distance between int vectors
|
||||
*/
|
||||
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_l2_distance);
|
||||
Datum
|
||||
intvec_l2_distance(PG_FUNCTION_ARGS)
|
||||
{
|
||||
IntVector *a = PG_GETARG_INTVEC_P(0);
|
||||
IntVector *b = PG_GETARG_INTVEC_P(1);
|
||||
int8 *ax = a->x;
|
||||
int8 *bx = b->x;
|
||||
int distance = 0;
|
||||
|
||||
CheckDims(a, b);
|
||||
|
||||
/* Auto-vectorized */
|
||||
for (int i = 0; i < a->dim; i++)
|
||||
{
|
||||
int diff = ax[i] - bx[i];
|
||||
|
||||
distance += diff * diff;
|
||||
}
|
||||
|
||||
PG_RETURN_FLOAT8(sqrt((double) distance));
|
||||
}
|
||||
|
||||
/*
|
||||
* Get the L2 squared distance between int vectors
|
||||
*/
|
||||
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_l2_squared_distance);
|
||||
Datum
|
||||
intvec_l2_squared_distance(PG_FUNCTION_ARGS)
|
||||
{
|
||||
IntVector *a = PG_GETARG_INTVEC_P(0);
|
||||
IntVector *b = PG_GETARG_INTVEC_P(1);
|
||||
int8 *ax = a->x;
|
||||
int8 *bx = b->x;
|
||||
int distance = 0;
|
||||
|
||||
CheckDims(a, b);
|
||||
|
||||
/* Auto-vectorized */
|
||||
for (int i = 0; i < a->dim; i++)
|
||||
{
|
||||
int diff = ax[i] - bx[i];
|
||||
|
||||
distance += diff * diff;
|
||||
}
|
||||
|
||||
PG_RETURN_FLOAT8((double) distance);
|
||||
}
|
||||
|
||||
/*
|
||||
* Get the inner product of two int vectors
|
||||
*/
|
||||
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_inner_product);
|
||||
Datum
|
||||
intvec_inner_product(PG_FUNCTION_ARGS)
|
||||
{
|
||||
IntVector *a = PG_GETARG_INTVEC_P(0);
|
||||
IntVector *b = PG_GETARG_INTVEC_P(1);
|
||||
int8 *ax = a->x;
|
||||
int8 *bx = b->x;
|
||||
int distance = 0;
|
||||
|
||||
CheckDims(a, b);
|
||||
|
||||
/* Auto-vectorized */
|
||||
for (int i = 0; i < a->dim; i++)
|
||||
distance += ax[i] * bx[i];
|
||||
|
||||
PG_RETURN_FLOAT8((double) distance);
|
||||
}
|
||||
|
||||
/*
|
||||
* Get the negative inner product of two int vectors
|
||||
*/
|
||||
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_negative_inner_product);
|
||||
Datum
|
||||
intvec_negative_inner_product(PG_FUNCTION_ARGS)
|
||||
{
|
||||
IntVector *a = PG_GETARG_INTVEC_P(0);
|
||||
IntVector *b = PG_GETARG_INTVEC_P(1);
|
||||
int8 *ax = a->x;
|
||||
int8 *bx = b->x;
|
||||
int distance = 0;
|
||||
|
||||
CheckDims(a, b);
|
||||
|
||||
/* Auto-vectorized */
|
||||
for (int i = 0; i < a->dim; i++)
|
||||
distance += ax[i] * bx[i];
|
||||
|
||||
PG_RETURN_FLOAT8((double) -distance);
|
||||
}
|
||||
|
||||
/*
|
||||
* Get the cosine distance between two int vectors
|
||||
*/
|
||||
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_cosine_distance);
|
||||
Datum
|
||||
intvec_cosine_distance(PG_FUNCTION_ARGS)
|
||||
{
|
||||
IntVector *a = PG_GETARG_INTVEC_P(0);
|
||||
IntVector *b = PG_GETARG_INTVEC_P(1);
|
||||
int8 *ax = a->x;
|
||||
int8 *bx = b->x;
|
||||
int distance = 0;
|
||||
int norma = 0;
|
||||
int normb = 0;
|
||||
double similarity;
|
||||
|
||||
CheckDims(a, b);
|
||||
|
||||
/* Auto-vectorized */
|
||||
for (int i = 0; i < a->dim; i++)
|
||||
{
|
||||
int8 axi = ax[i];
|
||||
int8 bxi = bx[i];
|
||||
|
||||
distance += axi * bxi;
|
||||
norma += axi * axi;
|
||||
normb += bxi * bxi;
|
||||
}
|
||||
|
||||
/* Use sqrt(a * b) over sqrt(a) * sqrt(b) */
|
||||
similarity = (double) distance / sqrt((double) norma * (double) normb);
|
||||
|
||||
#ifdef _MSC_VER
|
||||
/* /fp:fast may not propagate NaN */
|
||||
if (isnan(similarity))
|
||||
PG_RETURN_FLOAT8(NAN);
|
||||
#endif
|
||||
|
||||
/* Keep in range */
|
||||
if (similarity > 1)
|
||||
similarity = 1;
|
||||
else if (similarity < -1)
|
||||
similarity = -1;
|
||||
|
||||
PG_RETURN_FLOAT8(1 - similarity);
|
||||
}
|
||||
|
||||
/*
|
||||
* Get the L1 distance between two int vectors
|
||||
*/
|
||||
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_l1_distance);
|
||||
Datum
|
||||
intvec_l1_distance(PG_FUNCTION_ARGS)
|
||||
{
|
||||
IntVector *a = PG_GETARG_INTVEC_P(0);
|
||||
IntVector *b = PG_GETARG_INTVEC_P(1);
|
||||
int8 *ax = a->x;
|
||||
int8 *bx = b->x;
|
||||
int distance = 0;
|
||||
|
||||
CheckDims(a, b);
|
||||
|
||||
/* Auto-vectorized */
|
||||
for (int i = 0; i < a->dim; i++)
|
||||
distance += abs(ax[i] - bx[i]);
|
||||
|
||||
PG_RETURN_FLOAT8((double) distance);
|
||||
}
|
||||
23
src/intvec.h
Normal file
23
src/intvec.h
Normal file
@@ -0,0 +1,23 @@
|
||||
#ifndef INTVEC_H
|
||||
#define INTVEC_H
|
||||
|
||||
#include "vector.h"
|
||||
|
||||
#define INTVEC_MAX_DIM VECTOR_MAX_DIM
|
||||
|
||||
#define INTVEC_SIZE(_dim) (offsetof(IntVector, x) + sizeof(int8)*(_dim))
|
||||
#define DatumGetIntVector(x) ((IntVector *) PG_DETOAST_DATUM(x))
|
||||
#define PG_GETARG_INTVEC_P(x) DatumGetIntVector(PG_GETARG_DATUM(x))
|
||||
#define PG_RETURN_INTVEC_P(x) PG_RETURN_POINTER(x)
|
||||
|
||||
typedef struct IntVector
|
||||
{
|
||||
int32 vl_len_; /* varlena header (do not touch directly!) */
|
||||
int16 dim; /* number of dimensions */
|
||||
int16 unused;
|
||||
int8 x[FLEXIBLE_ARRAY_MEMBER];
|
||||
} IntVector;
|
||||
|
||||
IntVector *InitIntVector(int dim);
|
||||
|
||||
#endif
|
||||
Reference in New Issue
Block a user