Compare commits

..

1 Commits

Author SHA1 Message Date
Andrew Kane
ec68b75892 Added bulk hashing [skip ci] 2024-10-29 11:41:57 -07:00
8 changed files with 48 additions and 21 deletions

View File

@@ -1,4 +1,4 @@
## 0.8.0 (2024-10-30) ## 0.8.0 (unreleased)
- Added support for iterative index scans - Added support for iterative index scans
- Added casts for arrays to `sparsevec` - Added casts for arrays to `sparsevec`

View File

@@ -2,7 +2,7 @@
"name": "vector", "name": "vector",
"abstract": "Open-source vector similarity search for Postgres", "abstract": "Open-source vector similarity search for Postgres",
"description": "Supports L2 distance, inner product, and cosine distance", "description": "Supports L2 distance, inner product, and cosine distance",
"version": "0.8.0", "version": "0.7.4",
"maintainer": [ "maintainer": [
"Andrew Kane <andrew@ankane.org>" "Andrew Kane <andrew@ankane.org>"
], ],
@@ -12,7 +12,7 @@
"prereqs": { "prereqs": {
"runtime": { "runtime": {
"requires": { "requires": {
"PostgreSQL": "13.0.0" "PostgreSQL": "12.0.0"
} }
} }
}, },
@@ -20,7 +20,7 @@
"vector": { "vector": {
"file": "sql/vector.sql", "file": "sql/vector.sql",
"docfile": "README.md", "docfile": "README.md",
"version": "0.8.0", "version": "0.7.4",
"abstract": "Open-source vector similarity search for Postgres" "abstract": "Open-source vector similarity search for Postgres"
} }
}, },

View File

@@ -1,5 +1,5 @@
EXTENSION = vector EXTENSION = vector
EXTVERSION = 0.8.0 EXTVERSION = 0.7.4
MODULE_big = vector MODULE_big = vector
DATA = $(wildcard sql/*--*--*.sql) DATA = $(wildcard sql/*--*--*.sql)

View File

@@ -1,5 +1,5 @@
EXTENSION = vector EXTENSION = vector
EXTVERSION = 0.8.0 EXTVERSION = 0.7.4
DATA_built = sql\$(EXTENSION)--$(EXTVERSION).sql DATA_built = sql\$(EXTENSION)--$(EXTVERSION).sql
OBJS = src\bitutils.obj src\bitvec.obj src\halfutils.obj src\halfvec.obj src\hnsw.obj src\hnswbuild.obj src\hnswinsert.obj src\hnswscan.obj src\hnswutils.obj src\hnswvacuum.obj src\ivfbuild.obj src\ivfflat.obj src\ivfinsert.obj src\ivfkmeans.obj src\ivfscan.obj src\ivfutils.obj src\ivfvacuum.obj src\sparsevec.obj src\vector.obj OBJS = src\bitutils.obj src\bitvec.obj src\halfutils.obj src\halfvec.obj src\hnsw.obj src\hnswbuild.obj src\hnswinsert.obj src\hnswscan.obj src\hnswutils.obj src\hnswvacuum.obj src\ivfbuild.obj src\ivfflat.obj src\ivfinsert.obj src\ivfkmeans.obj src\ivfscan.obj src\ivfutils.obj src\ivfvacuum.obj src\sparsevec.obj src\vector.obj

View File

@@ -17,11 +17,11 @@ Plus [ACID](https://en.wikipedia.org/wiki/ACID) compliance, point-in-time recove
### Linux and Mac ### Linux and Mac
Compile and install the extension (supports Postgres 13+) Compile and install the extension (supports Postgres 12+)
```sh ```sh
cd /tmp cd /tmp
git clone --branch v0.8.0 https://github.com/pgvector/pgvector.git git clone --branch v0.7.4 https://github.com/pgvector/pgvector.git
cd pgvector cd pgvector
make make
make install # may need sudo make install # may need sudo
@@ -46,7 +46,7 @@ Then use `nmake` to build:
```cmd ```cmd
set "PGROOT=C:\Program Files\PostgreSQL\16" set "PGROOT=C:\Program Files\PostgreSQL\16"
cd %TEMP% cd %TEMP%
git clone --branch v0.8.0 https://github.com/pgvector/pgvector.git git clone --branch v0.7.4 https://github.com/pgvector/pgvector.git
cd pgvector cd pgvector
nmake /F Makefile.win nmake /F Makefile.win
nmake /F Makefile.win install nmake /F Makefile.win install
@@ -324,7 +324,7 @@ For a large number of workers, you may also need to increase `max_parallel_worke
### Indexing Progress ### Indexing Progress
Check [indexing progress](https://www.postgresql.org/docs/current/progress-reporting.html#CREATE-INDEX-PROGRESS-REPORTING) Check [indexing progress](https://www.postgresql.org/docs/current/progress-reporting.html#CREATE-INDEX-PROGRESS-REPORTING) with Postgres 12+
```sql ```sql
SELECT phase, round(100.0 * blocks_done / nullif(blocks_total, 0), 1) AS "%" FROM pg_stat_progress_create_index; SELECT phase, round(100.0 * blocks_done / nullif(blocks_total, 0), 1) AS "%" FROM pg_stat_progress_create_index;
@@ -410,7 +410,7 @@ For a large number of workers, you may also need to increase `max_parallel_worke
### Indexing Progress ### Indexing Progress
Check [indexing progress](https://www.postgresql.org/docs/current/progress-reporting.html#CREATE-INDEX-PROGRESS-REPORTING) Check [indexing progress](https://www.postgresql.org/docs/current/progress-reporting.html#CREATE-INDEX-PROGRESS-REPORTING) with Postgres 12+
```sql ```sql
SELECT phase, round(100.0 * tuples_done / nullif(tuples_total, 0), 1) AS "%" FROM pg_stat_progress_create_index; SELECT phase, round(100.0 * tuples_done / nullif(tuples_total, 0), 1) AS "%" FROM pg_stat_progress_create_index;
@@ -477,7 +477,7 @@ CREATE TABLE items (embedding vector(3), category_id int) PARTITION BY LIST(cate
## Iterative Index Scans ## Iterative Index Scans
*Added in 0.8.0* *Unreleased*
With approximate indexes, queries with filtering can return less results since filtering is applied *after* the index is scanned. Starting with 0.8.0, you can enable iterative index scans, which will automatically scan more of the index until enough results are found (or it reaches `hnsw.max_scan_tuples` or `ivfflat.max_probes`). With approximate indexes, queries with filtering can return less results since filtering is applied *after* the index is scanned. Starting with 0.8.0, you can enable iterative index scans, which will automatically scan more of the index until enough results are found (or it reaches `hnsw.max_scan_tuples` or `ivfflat.max_probes`).
@@ -1153,7 +1153,7 @@ This adds pgvector to the [Postgres image](https://hub.docker.com/_/postgres) (r
You can also build the image manually: You can also build the image manually:
```sh ```sh
git clone --branch v0.8.0 https://github.com/pgvector/pgvector.git git clone --branch v0.7.4 https://github.com/pgvector/pgvector.git
cd pgvector cd pgvector
docker build --pull --build-arg PG_MAJOR=17 -t myuser/pgvector . docker build --pull --build-arg PG_MAJOR=17 -t myuser/pgvector .
``` ```

View File

@@ -722,6 +722,8 @@ CountElement(HnswElement skipElement, HnswElement e)
static void static void
HnswLoadUnvisitedFromMemory(char *base, HnswElement element, HnswUnvisited * unvisited, int *unvisitedLength, visited_hash * v, int lc, HnswNeighborArray * localNeighborhood, Size neighborhoodSize) HnswLoadUnvisitedFromMemory(char *base, HnswElement element, HnswUnvisited * unvisited, int *unvisitedLength, visited_hash * v, int lc, HnswNeighborArray * localNeighborhood, Size neighborhoodSize)
{ {
uint32 hashes[HNSW_MAX_M * 2];
/* Get the neighborhood at layer lc */ /* Get the neighborhood at layer lc */
HnswNeighborArray *neighborhood = HnswGetNeighbors(base, element, lc); HnswNeighborArray *neighborhood = HnswGetNeighbors(base, element, lc);
@@ -733,14 +735,33 @@ HnswLoadUnvisitedFromMemory(char *base, HnswElement element, HnswUnvisited * unv
*unvisitedLength = 0; *unvisitedLength = 0;
for (int i = 0; i < localNeighborhood->length; i++) for (int i = 0; i < localNeighborhood->length; i++)
hashes[i] = HnswPtrAccess(base, localNeighborhood->items[i].element)->hash;
if (base != NULL)
{ {
HnswCandidate *hc = &localNeighborhood->items[i]; for (int i = 0; i < localNeighborhood->length; i++)
bool found; {
HnswCandidate *hc = &localNeighborhood->items[i];
bool found;
AddToVisited(base, v, hc->element, true, &found); offsethash_insert_hash(v->offsets, HnswPtrOffset(hc->element), hashes[i], &found);
if (!found) if (!found)
unvisited[(*unvisitedLength)++].element = HnswPtrAccess(base, hc->element); unvisited[(*unvisitedLength)++].element = HnswPtrAccess(base, hc->element);
}
}
else
{
for (int i = 0; i < localNeighborhood->length; i++)
{
HnswCandidate *hc = &localNeighborhood->items[i];
bool found;
pointerhash_insert_hash(v->pointers, (uintptr_t) HnswPtrPointer(hc->element), hashes[i], &found);
if (!found)
unvisited[(*unvisitedLength)++].element = HnswPtrAccess(base, hc->element);
}
} }
} }

View File

@@ -4,7 +4,6 @@
#include <math.h> #include <math.h>
#include "catalog/pg_type.h" #include "catalog/pg_type.h"
#include "common/shortest_dec.h"
#include "common/string.h" #include "common/string.h"
#include "fmgr.h" #include "fmgr.h"
#include "halfutils.h" #include "halfutils.h"
@@ -13,10 +12,17 @@
#include "sparsevec.h" #include "sparsevec.h"
#include "utils/array.h" #include "utils/array.h"
#include "utils/builtins.h" #include "utils/builtins.h"
#include "utils/float.h"
#include "utils/lsyscache.h" #include "utils/lsyscache.h"
#include "vector.h" #include "vector.h"
#if PG_VERSION_NUM >= 120000
#include "common/shortest_dec.h"
#include "utils/float.h"
#else
#include <float.h>
#include "utils/builtins.h"
#endif
typedef struct SparseInputElement typedef struct SparseInputElement
{ {
int32 index; int32 index;

View File

@@ -1,4 +1,4 @@
comment = 'vector data type and ivfflat and hnsw access methods' comment = 'vector data type and ivfflat and hnsw access methods'
default_version = '0.8.0' default_version = '0.7.4'
module_pathname = '$libdir/vector' module_pathname = '$libdir/vector'
relocatable = true relocatable = true