Compare commits

..

24 Commits

Author SHA1 Message Date
Andrew Kane
007defb6dd Updated [skip ci] 2025-12-12 14:34:20 -08:00
Andrew Kane
627d3fcc27 Merge branch 'master' into includes 2025-12-12 14:33:42 -08:00
Andrew Kane
d557f1c9b1 Added link to pgvector-tcl [skip ci] 2025-12-11 23:22:56 -08:00
Andrew Kane
3d6dd20469 Removed unneeded check 2025-12-11 16:28:45 -08:00
Andrew Kane
c711da411c Improved includes for indexes 2025-12-11 15:35:37 -08:00
Andrew Kane
3975e212cc Improved includes for ivfkmeans [skip ci] 2025-12-11 13:26:36 -08:00
Andrew Kane
8983103d58 Merge branch 'master' into includes 2025-12-11 00:37:44 -08:00
Andrew Kane
2b741ef068 Improved includes for types 2025-12-11 00:35:18 -08:00
Andrew Kane
6aba9f4422 Added more headers 2025-12-10 23:49:14 -08:00
Andrew Kane
364e4149a6 Added more headers [skip ci] 2025-12-10 23:45:25 -08:00
Andrew Kane
08412c5704 Added more headers 2025-12-10 23:38:46 -08:00
Andrew Kane
88f9c495ba Added more headers [skip ci] 2025-12-10 23:37:12 -08:00
Andrew Kane
fc8c736858 Added more headers [skip ci] 2025-12-10 23:32:21 -08:00
Andrew Kane
a233d68365 Added more headers [skip ci] 2025-12-10 23:19:52 -08:00
Andrew Kane
831ba3f3de Added more headers [skip ci] 2025-12-10 23:13:40 -08:00
Andrew Kane
0c5e1e89d3 Added more headers [skip ci] 2025-12-10 23:05:28 -08:00
Andrew Kane
eda94a46d8 Added more headers 2025-12-10 22:55:54 -08:00
Andrew Kane
8ca5e37c83 Fixed CI 2025-12-10 22:30:32 -08:00
Andrew Kane
18f8562ce9 Improved includes 2025-12-10 22:25:38 -08:00
Andrew Kane
f85a7d3fd0 Keep clang-tidy happy [skip ci] 2025-12-10 20:03:26 -08:00
Andrew Kane
5c25f81789 Fixed inconsistent parameter names [skip ci] 2025-12-10 18:49:00 -08:00
Andrew Kane
d72ee71f23 Fixed compilation errors with Postgres 19 2025-12-10 14:59:29 -08:00
Andrew Kane
b7c0085afe Added instructions for APK [skip ci] 2025-12-01 22:52:23 -08:00
Andrew Kane
2b742576b4 Added links [skip ci] 2025-11-24 12:41:30 -08:00
22 changed files with 305 additions and 199 deletions

View File

@@ -37,10 +37,10 @@ jobs:
- run: make installcheck
- if: ${{ failure() }}
run: cat regression.diffs
- run: |
sudo apt-get update
sudo apt-get install libipc-run-perl
- run: make prove_installcheck
# - run: |
# sudo apt-get update
# sudo apt-get install libipc-run-perl
# - run: make prove_installcheck
mac:
runs-on: ${{ matrix.os }}
if: ${{ !startsWith(github.ref_name, 'windows') }}
@@ -65,21 +65,21 @@ jobs:
- if: ${{ failure() }}
run: cat regression.diffs
# Homebrew Postgres does not enable TAP tests, so need to download
- run: |
brew install cpanm
cpanm --notest IPC::Run
wget -q https://github.com/postgres/postgres/archive/refs/tags/$TAG.tar.gz
tar xf $TAG.tar.gz
mv postgres-$TAG postgres
env:
TAG: ${{ matrix.postgres == 18 && 'REL_18_0' || 'REL_14_19' }}
- run: make prove_installcheck PROVE_FLAGS="-I ./postgres/src/test/perl -I ./test/perl"
env:
PERL5LIB: /Users/runner/perl5/lib/perl5
- run: make clean && $(brew --prefix llvm@$LLVM_VERSION)/bin/scan-build --status-bugs make
env:
LLVM_VERSION: ${{ matrix.os == 'macos-26' && 20 || 18 }}
PG_CFLAGS: -DUSE_ASSERT_CHECKING
# - run: |
# brew install cpanm
# cpanm --notest IPC::Run
# wget -q https://github.com/postgres/postgres/archive/refs/tags/$TAG.tar.gz
# tar xf $TAG.tar.gz
# mv postgres-$TAG postgres
# env:
# TAG: ${{ matrix.postgres == 18 && 'REL_18_0' || 'REL_14_19' }}
# - run: make prove_installcheck PROVE_FLAGS="-I ./postgres/src/test/perl -I ./test/perl"
# env:
# PERL5LIB: /Users/runner/perl5/lib/perl5
# - run: make clean && $(brew --prefix llvm@$LLVM_VERSION)/bin/scan-build --status-bugs make
# env:
# LLVM_VERSION: ${{ matrix.os == 'macos-26' && 20 || 18 }}
# PG_CFLAGS: -DUSE_ASSERT_CHECKING
windows:
runs-on: ${{ matrix.os }}
if: ${{ !startsWith(github.ref_name, 'mac') }}
@@ -106,38 +106,4 @@ jobs:
shell: cmd
- if: ${{ failure() }}
run: cat regression.diffs
i386:
if: ${{ !startsWith(github.ref_name, 'mac') && !startsWith(github.ref_name, 'windows') }}
runs-on: ubuntu-latest
container:
image: debian:12
options: --platform linux/386
steps:
- run: apt-get update && apt-get install -y build-essential git libipc-run-perl postgresql-15 postgresql-server-dev-15 sudo
- run: service postgresql start
- run: |
git clone https://github.com/${{ github.repository }}.git pgvector
cd pgvector
git fetch origin ${{ github.ref }}
git reset --hard FETCH_HEAD
make
make install
chown -R postgres .
sudo -u postgres make installcheck
sudo -u postgres make prove_installcheck
env:
PG_CFLAGS: -DUSE_ASSERT_CHECKING -Wall -Wextra -Werror -Wno-unused-parameter -Wno-sign-compare
- if: ${{ failure() }}
run: cat pgvector/regression.diffs
valgrind:
if: ${{ !startsWith(github.ref_name, 'mac') && !startsWith(github.ref_name, 'windows') }}
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v5
- uses: ankane/setup-postgres-valgrind@v1
with:
postgres-version: 18
check-ub: yes
- run: make OPTFLAGS=""
- run: sudo --preserve-env=PG_CONFIG make install
- run: make installcheck

View File

@@ -29,7 +29,7 @@ make install # may need sudo
See the [installation notes](#installation-notes---linux-and-mac) if you run into issues
You can also install it with [Docker](#docker), [Homebrew](#homebrew), [PGXN](#pgxn), [APT](#apt), [Yum](#yum), [pkg](#pkg), or [conda-forge](#conda-forge), and it comes preinstalled with [Postgres.app](#postgresapp) and many [hosted providers](#hosted-postgres). There are also instructions for [GitHub Actions](https://github.com/pgvector/setup-pgvector).
You can also install it with [Docker](#docker), [Homebrew](#homebrew), [PGXN](#pgxn), [APT](#apt), [Yum](#yum), [pkg](#pkg), [APK](#apk), or [conda-forge](#conda-forge), and it comes preinstalled with [Postgres.app](#postgresapp) and many [hosted providers](#hosted-postgres). There are also instructions for [GitHub Actions](https://github.com/pgvector/setup-pgvector).
### Windows
@@ -772,6 +772,7 @@ Use pgvector from any language with a Postgres client. You can even generate and
Language | Libraries / Examples
--- | ---
Ada | [pgvector-ada](https://github.com/pgvector/pgvector-ada)
Algol | [pgvector-algol](https://github.com/pgvector/pgvector-algol)
C | [pgvector-c](https://github.com/pgvector/pgvector-c)
C++ | [pgvector-cpp](https://github.com/pgvector/pgvector-cpp)
C#, F#, Visual Basic | [pgvector-dotnet](https://github.com/pgvector/pgvector-dotnet)
@@ -794,12 +795,15 @@ OCaml | [pgvector-ocaml](https://github.com/pgvector/pgvector-ocaml)
Pascal | [pgvector-pascal](https://github.com/pgvector/pgvector-pascal)
Perl | [pgvector-perl](https://github.com/pgvector/pgvector-perl)
PHP | [pgvector-php](https://github.com/pgvector/pgvector-php)
Prolog | [pgvector-prolog](https://github.com/pgvector/pgvector-prolog)
Python | [pgvector-python](https://github.com/pgvector/pgvector-python)
R | [pgvector-r](https://github.com/pgvector/pgvector-r)
Racket | [pgvector-racket](https://github.com/pgvector/pgvector-racket)
Raku | [pgvector-raku](https://github.com/pgvector/pgvector-raku)
Ruby | [pgvector-ruby](https://github.com/pgvector/pgvector-ruby), [Neighbor](https://github.com/ankane/neighbor)
Rust | [pgvector-rust](https://github.com/pgvector/pgvector-rust)
Swift | [pgvector-swift](https://github.com/pgvector/pgvector-swift)
Tcl | [pgvector-tcl](https://github.com/pgvector/pgvector-tcl)
Zig | [pgvector-zig](https://github.com/pgvector/pgvector-zig)
## Frequently Asked Questions
@@ -1229,6 +1233,14 @@ cd /usr/ports/databases/pgvector
make install
```
### APK
Install the Alpine package with:
```sh
apk add postgresql-pgvector
```
### conda-forge
With Conda Postgres, install from [conda-forge](https://anaconda.org/conda-forge/pgvector) with:

View File

@@ -2,6 +2,7 @@
#include "bitutils.h"
#include "bitvec.h"
#include "fmgr.h"
#include "utils/varbit.h"
#include "vector.h"

View File

@@ -1,5 +1,7 @@
#include "postgres.h"
#include <math.h>
#include "halfutils.h"
#include "halfvec.h"

View File

@@ -13,12 +13,16 @@
#include "port.h" /* for strtof() */
#include "sparsevec.h"
#include "utils/array.h"
#include "utils/builtins.h"
#include "utils/float.h"
#include "utils/fmgrprotos.h"
#include "utils/lsyscache.h"
#include "utils/numeric.h"
#include "utils/varbit.h"
#include "vector.h"
#if PG_VERSION_NUM >= 160000
#include "varatt.h"
#endif
#define STATE_DIMS(x) (ARR_DIMS(x)[0] - 1)
#define CreateStateDatums(dim) palloc(sizeof(Datum) * (dim + 1))

View File

@@ -1,18 +1,28 @@
#include "postgres.h"
#include <float.h>
#include <limits.h>
#include <math.h>
#include "access/amapi.h"
#include "access/genam.h"
#include "access/reloptions.h"
#include "commands/progress.h"
#include "commands/vacuum.h"
#include "fmgr.h"
#include "hnsw.h"
#include "miscadmin.h"
#include "nodes/nodes.h"
#include "nodes/pathnodes.h"
#include "nodes/pg_list.h"
#include "storage/lockdefs.h"
#include "storage/shmem.h"
#include "utils/float.h"
#include "utils/guc.h"
#include "utils/relcache.h"
#include "utils/selfuncs.h"
#include "utils/spccache.h"
#include "vector.h"
#if PG_VERSION_NUM < 150000
#define MarkGUCPrefixReserved(x) EmitWarningsOnPlaceholders(x)

View File

@@ -3,6 +3,8 @@
#include "postgres.h"
#include <math.h>
#include "access/genam.h"
#include "access/parallel.h"
#include "lib/pairingheap.h"
@@ -12,6 +14,10 @@
#include "utils/sampling.h"
#include "vector.h"
#if PG_VERSION_NUM >= 190000
typedef Pointer Item;
#endif
#define HNSW_MAX_DIM 2000
#define HNSW_MAX_NNZ 1000
@@ -362,13 +368,6 @@ typedef union
ItemPointerData indextid;
} HnswUnvisited;
typedef struct HnswReadStreamData
{
HnswUnvisited *unvisited;
int unvisitedLength;
int visited;
} HnswReadStreamData;
typedef struct HnswScanOpaqueData
{
const HnswTypeInfo *typeInfo;
@@ -424,14 +423,14 @@ bool HnswCheckNorm(HnswSupport * support, Datum value);
Buffer HnswNewBuffer(Relation index, ForkNumber forkNum);
void HnswInitPage(Buffer buf, Page page);
void HnswInit(void);
List *HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation index, HnswSupport * support, int m, bool inserting, HnswElement skipElement, visited_hash * v, pairingheap **discarded, bool initVisited, int64 *tuples, bool maintenance);
List *HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation index, HnswSupport * support, int m, bool inserting, HnswElement skipElement, visited_hash * v, pairingheap **discarded, bool initVisited, int64 *tuples);
HnswElement HnswGetEntryPoint(Relation index);
void HnswGetMetaPageInfo(Relation index, int *m, HnswElement * entryPoint);
void *HnswAlloc(HnswAllocator * allocator, Size size);
HnswElement HnswInitElement(char *base, ItemPointer tid, int m, double ml, int maxLevel, HnswAllocator * alloc);
HnswElement HnswInitElementFromBlock(BlockNumber blkno, OffsetNumber offno);
void HnswFindElementNeighbors(char *base, HnswElement element, HnswElement entryPoint, Relation index, HnswSupport * support, int m, int efConstruction, bool existing, bool maintenance);
HnswSearchCandidate *HnswEntryCandidate(char *base, HnswElement em, HnswQuery * q, Relation rel, HnswSupport * support, bool loadVec);
void HnswFindElementNeighbors(char *base, HnswElement element, HnswElement entryPoint, Relation index, HnswSupport * support, int m, int efConstruction, bool existing);
HnswSearchCandidate *HnswEntryCandidate(char *base, HnswElement entryPoint, HnswQuery * q, Relation index, HnswSupport * support, bool loadVec);
void HnswUpdateMetaPage(Relation index, int updateEntry, HnswElement entryPoint, BlockNumber insertPage, ForkNumber forkNum, bool building);
void HnswSetNeighborTuple(char *base, HnswNeighborTuple ntup, HnswElement e, int m);
void HnswAddHeapTid(HnswElement element, ItemPointer heaptid);

View File

@@ -36,23 +36,47 @@
*/
#include "postgres.h"
#include <math.h>
#include "access/genam.h"
#include "access/parallel.h"
#include "access/relscan.h"
#include "access/table.h"
#include "access/tableam.h"
#include "access/tupdesc.h"
#include "access/xact.h"
#include "access/xloginsert.h"
#include "catalog/index.h"
#include "catalog/pg_type_d.h"
#include "commands/progress.h"
#include "common/relpath.h"
#include "hnsw.h"
#include "miscadmin.h"
#include "nodes/execnodes.h"
#include "optimizer/optimizer.h"
#include "storage/block.h"
#include "storage/buf.h"
#include "storage/bufmgr.h"
#include "storage/bufpage.h"
#include "storage/condition_variable.h"
#include "storage/dsm.h"
#include "storage/itemid.h"
#include "storage/itemptr.h"
#include "storage/lockdefs.h"
#include "storage/lwlock.h"
#include "storage/off.h"
#include "storage/shm_toc.h"
#include "storage/shmem.h"
#include "storage/spin.h"
#include "tcop/tcopprot.h"
#include "utils/datum.h"
#include "utils/memutils.h"
#include "utils/rel.h"
#include "utils/relcache.h"
#include "utils/snapmgr.h"
#include "utils/snapshot.h"
#if PG_VERSION_NUM < 190000
#include "storage/item.h"
#endif
#if PG_VERSION_NUM >= 160000
#include "varatt.h"
@@ -66,6 +90,11 @@
#if PG_VERSION_NUM >= 140000
#include "utils/backend_status.h"
#endif
#if PG_VERSION_NUM >= 170000
#include "utils/wait_event_types.h"
#elif PG_VERSION_NUM >= 140000
#include "utils/wait_event.h"
#endif
@@ -461,7 +490,7 @@ InsertTupleInMemory(HnswBuildState * buildstate, HnswElement element)
}
/* Find neighbors for element */
HnswFindElementNeighbors(base, element, entryPoint, NULL, support, m, efConstruction, false, true);
HnswFindElementNeighbors(base, element, entryPoint, NULL, support, m, efConstruction, false);
/* Update graph in memory */
UpdateGraphInMemory(support, element, m, entryPoint, buildstate);
@@ -549,7 +578,7 @@ InsertTuple(Relation index, Datum *values, bool *isnull, ItemPointer heaptid, Hn
/* Copy the datum */
memcpy(valuePtr, DatumGetPointer(value), valueSize);
HnswPtrStore(base, element->value, valuePtr);
HnswPtrStore(base, element->value, (char *) valuePtr);
/* Create a lock for the element */
LWLockInitialize(&element->lock, hnsw_lock_tranche_id);

View File

@@ -1,13 +1,27 @@
#include "postgres.h"
#include <math.h>
#include "access/genam.h"
#include "access/generic_xlog.h"
#include "common/relpath.h"
#include "hnsw.h"
#include "nodes/execnodes.h"
#include "storage/block.h"
#include "storage/buf.h"
#include "storage/bufmgr.h"
#include "storage/bufpage.h"
#include "storage/itemid.h"
#include "storage/itemptr.h"
#include "storage/lmgr.h"
#include "storage/lockdefs.h"
#include "storage/off.h"
#include "utils/datum.h"
#include "utils/memutils.h"
#include "utils/rel.h"
#include "utils/relcache.h"
#if PG_VERSION_NUM < 190000
#include "storage/item.h"
#endif
#if PG_VERSION_NUM >= 160000
#include "varatt.h"
@@ -712,7 +726,7 @@ HnswInsertTupleOnDisk(Relation index, HnswSupport * support, Datum value, ItemPo
/* Create an element */
element = HnswInitElement(base, heaptid, m, HnswGetMl(m), HnswGetMaxLevel(m), NULL);
HnswPtrStore(base, element->value, DatumGetPointer(value));
HnswPtrStore(base, element->value, (char *) DatumGetPointer(value));
/* Prevent concurrent inserts when likely updating entry point */
if (entryPoint == NULL || element->level > entryPoint->level)
@@ -729,7 +743,7 @@ HnswInsertTupleOnDisk(Relation index, HnswSupport * support, Datum value, ItemPo
}
/* Find neighbors for element */
HnswFindElementNeighbors(base, element, entryPoint, index, support, m, efConstruction, false, building);
HnswFindElementNeighbors(base, element, entryPoint, index, support, m, efConstruction, false);
/* Update graph on disk */
UpdateGraphOnDisk(index, support, element, m, entryPoint, building);

View File

@@ -1,12 +1,25 @@
#include "postgres.h"
#include "access/genam.h"
#include "access/relscan.h"
#include "access/sdir.h"
#include "access/skey.h"
#include "hnsw.h"
#include "lib/pairingheap.h"
#include "miscadmin.h"
#include "nodes/pg_list.h"
#include "pgstat.h"
#include "storage/bufmgr.h"
#include "storage/itemptr.h"
#include "storage/lmgr.h"
#include "storage/lockdefs.h"
#include "utils/float.h"
#include "utils/memutils.h"
#include "utils/relcache.h"
#include "utils/snapmgr.h"
#if PG_VERSION_NUM >= 160000
#include "varatt.h"
#endif
/*
* Algorithm 5 from paper
@@ -37,11 +50,11 @@ GetScanItems(IndexScanDesc scan, Datum value)
for (int lc = entryPoint->level; lc >= 1; lc--)
{
w = HnswSearchLayer(base, q, ep, 1, lc, index, support, m, false, NULL, NULL, NULL, true, NULL, false);
w = HnswSearchLayer(base, q, ep, 1, lc, index, support, m, false, NULL, NULL, NULL, true, NULL);
ep = w;
}
return HnswSearchLayer(base, q, ep, hnsw_ef_search, 0, index, support, m, false, NULL, &so->v, hnsw_iterative_scan != HNSW_ITERATIVE_SCAN_OFF ? &so->discarded : NULL, true, &so->tuples, false);
return HnswSearchLayer(base, q, ep, hnsw_ef_search, 0, index, support, m, false, NULL, &so->v, hnsw_iterative_scan != HNSW_ITERATIVE_SCAN_OFF ? &so->discarded : NULL, true, &so->tuples);
}
/*
@@ -72,7 +85,7 @@ ResumeScanItems(IndexScanDesc scan)
ep = lappend(ep, sc);
}
return HnswSearchLayer(base, &so->q, ep, batch_size, 0, index, &so->support, so->m, false, NULL, &so->v, &so->discarded, false, &so->tuples, false);
return HnswSearchLayer(base, &so->q, ep, batch_size, 0, index, &so->support, so->m, false, NULL, &so->v, &so->discarded, false, &so->tuples);
}
/*

View File

@@ -2,27 +2,33 @@
#include <math.h>
#include "access/genam.h"
#include "access/generic_xlog.h"
#include "catalog/pg_type.h"
#include "catalog/pg_type_d.h"
#include "common/hashfn.h"
#include "common/relpath.h"
#include "fmgr.h"
#include "hnsw.h"
#include "lib/pairingheap.h"
#include "nodes/pg_list.h"
#include "port/atomics.h"
#include "sparsevec.h"
#include "storage/block.h"
#include "storage/buf.h"
#include "storage/bufmgr.h"
#include "storage/bufpage.h"
#include "storage/itemptr.h"
#include "storage/lwlock.h"
#include "storage/off.h"
#include "utils/datum.h"
#include "utils/memdebug.h"
#include "utils/rel.h"
#include "utils/relcache.h"
#include "vector.h"
#if PG_VERSION_NUM >= 160000
#include "varatt.h"
#endif
#if PG_VERSION_NUM >= 170000
#include "storage/read_stream.h"
#endif
#if PG_VERSION_NUM < 170000
static inline uint64
murmurhash64(uint64 data)
@@ -264,7 +270,7 @@ HnswInitElement(char *base, ItemPointer heaptid, int m, double ml, int maxLevel,
HnswInitNeighbors(base, element, m, allocator);
HnswPtrStore(base, element->value, (Pointer) NULL);
HnswPtrStore(base, element->value, (char *) NULL);
return element;
}
@@ -290,7 +296,7 @@ HnswInitElementFromBlock(BlockNumber blkno, OffsetNumber offno)
element->blkno = blkno;
element->offno = offno;
HnswPtrStore(base, element->neighbors, (HnswNeighborArrayPtr *) NULL);
HnswPtrStore(base, element->value, (Pointer) NULL);
HnswPtrStore(base, element->value, (char *) NULL);
return element;
}
@@ -516,7 +522,7 @@ HnswLoadElementFromTuple(HnswElement element, HnswElementTuple etup, bool loadHe
char *base = NULL;
Datum value = datumCopy(PointerGetDatum(&etup->data), false, -1);
HnswPtrStore(base, element->value, DatumGetPointer(value));
HnswPtrStore(base, element->value, (char *) DatumGetPointer(value));
}
}
@@ -533,12 +539,14 @@ HnswGetDistance(Datum a, Datum b, HnswSupport * support)
* Load an element and optionally get its distance from q
*/
static void
HnswLoadElementImpl(Buffer buf, OffsetNumber offno, double *distance, HnswQuery * q, Relation index, HnswSupport * support, bool loadVec, double *maxDistance, HnswElement * element)
HnswLoadElementImpl(BlockNumber blkno, OffsetNumber offno, double *distance, HnswQuery * q, Relation index, HnswSupport * support, bool loadVec, double *maxDistance, HnswElement * element)
{
Buffer buf;
Page page;
HnswElementTuple etup;
/* Read vector */
buf = ReadBuffer(index, blkno);
LockBuffer(buf, BUFFER_LOCK_SHARE);
page = BufferGetPage(buf);
@@ -559,7 +567,7 @@ HnswLoadElementImpl(Buffer buf, OffsetNumber offno, double *distance, HnswQuery
if (distance == NULL || maxDistance == NULL || *distance < *maxDistance)
{
if (*element == NULL)
*element = HnswInitElementFromBlock(BufferGetBlockNumber(buf), offno);
*element = HnswInitElementFromBlock(blkno, offno);
HnswLoadElementFromTuple(*element, etup, true, loadVec);
}
@@ -573,9 +581,7 @@ HnswLoadElementImpl(Buffer buf, OffsetNumber offno, double *distance, HnswQuery
void
HnswLoadElement(HnswElement element, double *distance, HnswQuery * q, Relation index, HnswSupport * support, bool loadVec, double *maxDistance)
{
Buffer buf = ReadBuffer(index, element->blkno);
HnswLoadElementImpl(buf, element->offno, distance, q, index, support, loadVec, maxDistance, &element);
HnswLoadElementImpl(element->blkno, element->offno, distance, q, index, support, loadVec, maxDistance, &element);
}
/*
@@ -815,31 +821,11 @@ HnswLoadUnvisitedFromDisk(HnswElement element, HnswUnvisited * unvisited, int *u
}
}
#if PG_VERSION_NUM >= 170000
/*
* Get next block number for read stream
*/
static BlockNumber
HnswReadStreamNextBlock(ReadStream *stream, void *callback_private_data, void *per_buffer_data)
{
HnswReadStreamData *streamData = callback_private_data;
OffsetNumber *offno = per_buffer_data;
HnswUnvisited *uv;
if (streamData->visited == streamData->unvisitedLength)
return InvalidBlockNumber;
uv = &streamData->unvisited[streamData->visited++];
*offno = ItemPointerGetOffsetNumber(&uv->indextid);
return ItemPointerGetBlockNumber(&uv->indextid);
}
#endif
/*
* Algorithm 2 from paper
*/
List *
HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation index, HnswSupport * support, int m, bool inserting, HnswElement skipElement, visited_hash * v, pairingheap **discarded, bool initVisited, int64 *tuples, bool maintenance)
HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation index, HnswSupport * support, int m, bool inserting, HnswElement skipElement, visited_hash * v, pairingheap **discarded, bool initVisited, int64 *tuples)
{
List *w = NIL;
pairingheap *C = pairingheap_allocate(CompareNearestCandidates, NULL);
@@ -854,27 +840,6 @@ HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation in
int unvisitedLength;
bool inMemory = index == NULL;
#if PG_VERSION_NUM >= 170000
HnswReadStreamData streamData;
ReadStream *stream = NULL;
if (!inMemory)
{
int flags = READ_STREAM_DEFAULT;
if (maintenance)
{
flags |= READ_STREAM_MAINTENANCE;
}
#if PG_VERSION_NUM >= 180000
flags |= READ_STREAM_USE_BATCHING;
#endif
stream = read_stream_begin_relation(flags, NULL, index, MAIN_FORKNUM, HnswReadStreamNextBlock, &streamData, sizeof(OffsetNumber));
}
#endif
if (v == NULL)
{
v = &vh;
@@ -937,23 +902,13 @@ HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation in
if (inMemory)
HnswLoadUnvisitedFromMemory(base, cElement, unvisited, &unvisitedLength, v, lc, localNeighborhood, neighborhoodSize);
else
{
HnswLoadUnvisitedFromDisk(cElement, unvisited, &unvisitedLength, v, index, m, lm, lc);
#if PG_VERSION_NUM >= 170000
read_stream_resume(stream);
streamData.unvisited = unvisited;
streamData.unvisitedLength = unvisitedLength;
streamData.visited = 0;
#endif
}
/* OK to count elements instead of tuples */
if (tuples != NULL)
(*tuples) += unvisitedLength;
for (int i = 0;; i++)
for (int i = 0; i < unvisitedLength; i++)
{
HnswElement eElement;
HnswSearchCandidate *e;
@@ -964,46 +919,24 @@ HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation in
if (inMemory)
{
if (i == unvisitedLength)
break;
eElement = unvisited[i].element;
eDistance = GetElementDistance(base, eElement, q, support);
}
else
{
Buffer buf;
OffsetNumber offno;
#if PG_VERSION_NUM >= 170000
void *offnoPtr;
buf = read_stream_next_buffer(stream, &offnoPtr);
if (!BufferIsValid(buf))
break;
offno = *((OffsetNumber *) offnoPtr);
#else
ItemPointer indextid;
if (i == unvisitedLength)
break;
indextid = &unvisited[i].indextid;
buf = ReadBuffer(index, ItemPointerGetBlockNumber(indextid));
offno = ItemPointerGetOffsetNumber(indextid);
#endif
ItemPointer indextid = &unvisited[i].indextid;
BlockNumber blkno = ItemPointerGetBlockNumber(indextid);
OffsetNumber offno = ItemPointerGetOffsetNumber(indextid);
/* Avoid any allocations if not adding */
eElement = NULL;
HnswLoadElementImpl(buf, offno, &eDistance, q, index, support, inserting, alwaysAdd || discarded != NULL ? NULL : &f->distance, &eElement);
HnswLoadElementImpl(blkno, offno, &eDistance, q, index, support, inserting, alwaysAdd || discarded != NULL ? NULL : &f->distance, &eElement);
if (eElement == NULL)
continue;
}
if (eElement == NULL || !(eDistance < f->distance || alwaysAdd))
if (!(eDistance < f->distance || alwaysAdd))
{
if (discarded != NULL)
{
@@ -1053,11 +986,6 @@ HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation in
w = lappend(w, sc);
}
#if PG_VERSION_NUM >= 170000
if (!inMemory)
read_stream_end(stream);
#endif
return w;
}
@@ -1353,7 +1281,7 @@ PrecomputeHash(char *base, HnswElement element)
* Algorithm 1 from paper
*/
void
HnswFindElementNeighbors(char *base, HnswElement element, HnswElement entryPoint, Relation index, HnswSupport * support, int m, int efConstruction, bool existing, bool maintenance)
HnswFindElementNeighbors(char *base, HnswElement element, HnswElement entryPoint, Relation index, HnswSupport * support, int m, int efConstruction, bool existing)
{
List *ep;
List *w;
@@ -1380,7 +1308,7 @@ HnswFindElementNeighbors(char *base, HnswElement element, HnswElement entryPoint
/* 1st phase: greedy search to insert level */
for (int lc = entryLevel; lc >= level + 1; lc--)
{
w = HnswSearchLayer(base, &q, ep, 1, lc, index, support, m, true, skipElement, NULL, NULL, true, NULL, maintenance);
w = HnswSearchLayer(base, &q, ep, 1, lc, index, support, m, true, skipElement, NULL, NULL, true, NULL);
ep = w;
}
@@ -1399,7 +1327,7 @@ HnswFindElementNeighbors(char *base, HnswElement element, HnswElement entryPoint
List *lw = NIL;
ListCell *lc2;
w = HnswSearchLayer(base, &q, ep, efConstruction, lc, index, support, m, true, skipElement, NULL, NULL, true, NULL, maintenance);
w = HnswSearchLayer(base, &q, ep, efConstruction, lc, index, support, m, true, skipElement, NULL, NULL, true, NULL);
/* Convert search candidates to candidates */
foreach(lc2, w)

View File

@@ -1,13 +1,26 @@
#include "postgres.h"
#include <math.h>
#include "access/genam.h"
#include "access/generic_xlog.h"
#include "commands/vacuum.h"
#include "common/relpath.h"
#include "hnsw.h"
#include "nodes/pg_list.h"
#include "storage/block.h"
#include "storage/buf.h"
#include "storage/bufmgr.h"
#include "storage/bufpage.h"
#include "storage/itemptr.h"
#include "storage/lmgr.h"
#include "storage/lockdefs.h"
#include "storage/off.h"
#include "utils/memutils.h"
#include "utils/rel.h"
#include "utils/relcache.h"
#if PG_VERSION_NUM < 190000
#include "storage/item.h"
#endif
#if PG_VERSION_NUM >= 160000
#include "varatt.h"
@@ -178,7 +191,12 @@ NeedsUpdated(HnswVacuumState * vacuumstate, HnswElement element)
/* Also update if layer 0 is not full */
/* This could indicate too many candidates being deleted during insert */
if (!needsUpdated)
{
/* Keep clang-tidy happy */
Assert(ntup->count > 0);
needsUpdated = !ItemPointerIsValid(&ntup->indextids[ntup->count - 1]);
}
UnlockReleaseBuffer(buf);
@@ -212,7 +230,7 @@ RepairGraphElement(HnswVacuumState * vacuumstate, HnswElement element, HnswEleme
element->heaptidsLength = 0;
/* Find neighbors for element, skipping itself */
HnswFindElementNeighbors(base, element, entryPoint, index, support, m, efConstruction, true, true);
HnswFindElementNeighbors(base, element, entryPoint, index, support, m, efConstruction, true);
/* Zero memory for each element */
MemSet(ntup, 0, HNSW_TUPLE_ALLOC_SIZE);
@@ -528,8 +546,9 @@ MarkDeleted(HnswVacuumState * vacuumstate)
ntup = (HnswNeighborTuple) PageGetItem(npage, PageGetItemId(npage, neighborOffno));
/* Overwrite element */
/* Use memset instead of MemSet to keep clang-tidy happy */
etup->deleted = 1;
MemSet(&etup->data, 0, VARSIZE_ANY(&etup->data));
memset(&etup->data, 0, VARSIZE_ANY(&etup->data));
/* Overwrite neighbors */
for (int i = 0; i < ntup->count; i++)

View File

@@ -2,23 +2,52 @@
#include <float.h>
#include "access/attnum.h"
#include "access/genam.h"
#include "access/generic_xlog.h"
#include "access/itup.h"
#include "access/relscan.h"
#include "access/table.h"
#include "access/tableam.h"
#include "access/tupdesc.h"
#include "access/parallel.h"
#include "access/xact.h"
#include "bitvec.h"
#include "access/xloginsert.h"
#include "catalog/index.h"
#include "catalog/pg_operator_d.h"
#include "catalog/pg_type_d.h"
#include "commands/progress.h"
#include "halfvec.h"
#include "common/relpath.h"
#include "executor/tuptable.h"
#include "fmgr.h"
#include "ivfflat.h"
#include "miscadmin.h"
#include "nodes/execnodes.h"
#include "optimizer/optimizer.h"
#include "storage/block.h"
#include "storage/buf.h"
#include "storage/bufmgr.h"
#include "storage/bufpage.h"
#include "storage/condition_variable.h"
#include "storage/dsm.h"
#include "storage/itemptr.h"
#include "storage/lockdefs.h"
#include "storage/off.h"
#include "storage/shm_toc.h"
#include "storage/shmem.h"
#include "storage/spin.h"
#include "tcop/tcopprot.h"
#include "utils/memutils.h"
#include "vector.h"
#include "utils/rel.h"
#include "utils/relcache.h"
#include "utils/sampling.h"
#include "utils/snapmgr.h"
#include "utils/snapshot.h"
#include "utils/tuplesort.h"
#if PG_VERSION_NUM < 190000
#include "storage/item.h"
#endif
#if PG_VERSION_NUM >= 160000
#include "varatt.h"
@@ -32,6 +61,11 @@
#if PG_VERSION_NUM >= 140000
#include "utils/backend_status.h"
#endif
#if PG_VERSION_NUM >= 170000
#include "utils/wait_event_types.h"
#elif PG_VERSION_NUM >= 140000
#include "utils/wait_event.h"
#endif

View File

@@ -3,14 +3,22 @@
#include <float.h>
#include "access/amapi.h"
#include "access/genam.h"
#include "access/reloptions.h"
#include "commands/progress.h"
#include "commands/vacuum.h"
#include "fmgr.h"
#include "ivfflat.h"
#include "nodes/nodes.h"
#include "nodes/pathnodes.h"
#include "nodes/pg_list.h"
#include "storage/lockdefs.h"
#include "utils/float.h"
#include "utils/guc.h"
#include "utils/relcache.h"
#include "utils/selfuncs.h"
#include "utils/spccache.h"
#include "vector.h"
#if PG_VERSION_NUM < 150000
#define MarkGUCPrefixReserved(x) EmitWarningsOnPlaceholders(x)

View File

@@ -25,6 +25,10 @@
#include "portability/instr_time.h"
#endif
#if PG_VERSION_NUM >= 190000
typedef Pointer Item;
#endif
#define IVFFLAT_MAX_DIM 2000
/* Support functions */

View File

@@ -2,11 +2,29 @@
#include <float.h>
#include "access/genam.h"
#include "access/generic_xlog.h"
#include "access/itup.h"
#include "common/relpath.h"
#include "fmgr.h"
#include "ivfflat.h"
#include "nodes/execnodes.h"
#include "storage/block.h"
#include "storage/buf.h"
#include "storage/bufmgr.h"
#include "storage/bufpage.h"
#include "storage/itemid.h"
#include "storage/itemptr.h"
#include "storage/lmgr.h"
#include "storage/lockdefs.h"
#include "storage/off.h"
#include "utils/memutils.h"
#include "utils/rel.h"
#include "utils/relcache.h"
#if PG_VERSION_NUM < 190000
#include "storage/item.h"
#endif
/*
* Find the list that minimizes the distance function

View File

@@ -1,17 +1,15 @@
#include "postgres.h"
#include <float.h>
#include <limits.h>
#include <math.h>
#include "bitvec.h"
#include "halfutils.h"
#include "halfvec.h"
#include "access/genam.h"
#include "fmgr.h"
#include "ivfflat.h"
#include "miscadmin.h"
#include "utils/builtins.h"
#include "utils/datum.h"
#include "utils/memutils.h"
#include "vector.h"
#include "utils/relcache.h"
#if PG_VERSION_NUM >= 160000
#include "varatt.h"

View File

@@ -2,15 +2,38 @@
#include <float.h>
#include "access/attnum.h"
#include "access/genam.h"
#include "access/itup.h"
#include "access/relscan.h"
#include "access/sdir.h"
#include "access/skey.h"
#include "access/tupdesc.h"
#include "catalog/pg_operator_d.h"
#include "catalog/pg_type_d.h"
#include "common/relpath.h"
#include "executor/tuptable.h"
#include "fmgr.h"
#include "lib/pairingheap.h"
#include "ivfflat.h"
#include "miscadmin.h"
#include "pgstat.h"
#include "storage/block.h"
#include "storage/buf.h"
#include "storage/bufmgr.h"
#include "storage/bufpage.h"
#include "storage/itemid.h"
#include "storage/itemptr.h"
#include "storage/off.h"
#include "utils/memutils.h"
#include "utils/rel.h"
#include "utils/relcache.h"
#include "utils/snapmgr.h"
#include "utils/tuplesort.h"
#if PG_VERSION_NUM >= 160000
#include "varatt.h"
#endif
#define GetScanList(ptr) pairingheap_container(IvfflatScanList, ph_node, ptr)
#define GetScanListConst(ptr) pairingheap_const_container(IvfflatScanList, ph_node, ptr)

View File

@@ -1,13 +1,23 @@
#include "postgres.h"
#include "access/genam.h"
#include "access/generic_xlog.h"
#include "bitvec.h"
#include "catalog/pg_type.h"
#include "common/relpath.h"
#include "fmgr.h"
#include "halfutils.h"
#include "halfvec.h"
#include "ivfflat.h"
#include "storage/block.h"
#include "storage/buf.h"
#include "storage/bufmgr.h"
#include "storage/bufpage.h"
#include "utils/relcache.h"
#include "utils/varbit.h"
#include "vector.h"
#if PG_VERSION_NUM >= 160000
#include "varatt.h"
#endif
/*
* Allocate a vector array

View File

@@ -1,9 +1,18 @@
#include "postgres.h"
#include "access/genam.h"
#include "access/generic_xlog.h"
#include "access/itup.h"
#include "commands/vacuum.h"
#include "common/relpath.h"
#include "ivfflat.h"
#include "storage/block.h"
#include "storage/buf.h"
#include "storage/bufmgr.h"
#include "storage/bufpage.h"
#include "storage/itemptr.h"
#include "storage/off.h"
#include "utils/relcache.h"
#if PG_VERSION_NUM >= 180000
#define vacuum_delay_point() vacuum_delay_point(false)

View File

@@ -5,18 +5,23 @@
#include "catalog/pg_type.h"
#include "common/shortest_dec.h"
#include "common/string.h"
#include "fmgr.h"
#include "halfutils.h"
#include "halfvec.h"
#include "lib/stringinfo.h"
#include "libpq/pqformat.h"
#include "sparsevec.h"
#include "utils/array.h"
#include "utils/builtins.h"
#include "utils/float.h"
#include "utils/fmgrprotos.h"
#include "utils/lsyscache.h"
#include "vector.h"
#if PG_VERSION_NUM >= 160000
#include "varatt.h"
#endif
typedef struct SparseInputElement
{
int32 index;

View File

@@ -16,10 +16,10 @@
#include "port.h" /* for strtof() */
#include "sparsevec.h"
#include "utils/array.h"
#include "utils/builtins.h"
#include "utils/float.h"
#include "utils/fmgrprotos.h"
#include "utils/lsyscache.h"
#include "utils/numeric.h"
#include "utils/varbit.h"
#include "vector.h"
#if PG_VERSION_NUM >= 160000