liuyu/pgvector
1
0
Fork 0
mirror of https://github.com/pgvector/pgvector.git synced 2026-07-14 00:26:55 +08:00
Code Issues Packages Projects Releases Wiki Activity

Improved CheckCenters code [skip ci]

Browse Source
This commit is contained in:
Andrew Kane
2024-04-25 17:41:53 -07:00
parent dc88135515
commit cd95d6dfa4
1 changed files with 32 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

50
src/ivfkmeans.c
View File

@@ -498,18 +498,13 @@ ElkanKmeans(Relation index, VectorArray samples, VectorArray centers, const Ivff
} }
/* /*
* Detect issues with centers * Ensure no NaN or infinite values
*/ */
static void static void
CheckCenters(Relation index, VectorArray centers, const IvfflatTypeInfo * typeInfo) CheckElements(VectorArray centers, const IvfflatTypeInfo * typeInfo)
{ {
FmgrInfo *normprocinfo;
float *scratch = palloc(sizeof(float) * centers->dim); float *scratch = palloc(sizeof(float) * centers->dim);
if (centers->length != centers->maxlen)
elog(ERROR, "Not enough centers. Please report a bug.");
/* Ensure no NaN or infinite values */
for (int i = 0; i < centers->length; i++) for (int i = 0; i < centers->length; i++)
{ {
for (int j = 0; j < centers->dim; j++) for (int j = 0; j < centers->dim; j++)
@@ -527,24 +522,43 @@ CheckCenters(Relation index, VectorArray centers, const IvfflatTypeInfo * typeIn
elog(ERROR, "Infinite value detected. Please report a bug."); elog(ERROR, "Infinite value detected. Please report a bug.");
} }
} }
}
/* Ensure no zero vectors for cosine distance */ /*
* Ensure no zero vectors for cosine distance
*/
static void
CheckNorms(VectorArray centers, Relation index)
{
/* Check NORM_PROC instead of KMEANS_NORM_PROC */ /* Check NORM_PROC instead of KMEANS_NORM_PROC */
normprocinfo = IvfflatOptionalProcInfo(index, IVFFLAT_NORM_PROC); FmgrInfo *normprocinfo = IvfflatOptionalProcInfo(index, IVFFLAT_NORM_PROC);
if (normprocinfo != NULL) Oid collation = index->rd_indcollation[0];
if (normprocinfo == NULL)
return;
for (int i = 0; i < centers->length; i++)
{ {
Oid collation = index->rd_indcollation[0]; double norm = DatumGetFloat8(FunctionCall1Coll(normprocinfo, collation, PointerGetDatum(VectorArrayGet(centers, i))));
for (int i = 0; i < centers->length; i++) if (norm == 0)
{ elog(ERROR, "Zero norm detected. Please report a bug.");
double norm = DatumGetFloat8(FunctionCall1Coll(normprocinfo, collation, PointerGetDatum(VectorArrayGet(centers, i))));
if (norm == 0)
elog(ERROR, "Zero norm detected. Please report a bug.");
}
} }
} }
/*
* Detect issues with centers
*/
static void
CheckCenters(Relation index, VectorArray centers, const IvfflatTypeInfo * typeInfo)
{
if (centers->length != centers->maxlen)
elog(ERROR, "Not enough centers. Please report a bug.");
CheckElements(centers, typeInfo);
CheckNorms(centers, index);
}
/* /*
* Perform naive k-means centering * Perform naive k-means centering
* We use spherical k-means for inner product and cosine * We use spherical k-means for inner product and cosine