liuyu/pgvector
1
0
Fork 0
mirror of https://github.com/pgvector/pgvector.git synced 2026-07-14 16:46:54 +08:00
Code Issues Packages Projects Releases Wiki Activity

Updated readme for 0.7.0 [skip ci]

Browse Source
This commit is contained in:
Andrew Kane
2024-04-29 09:15:24 -07:00
parent a72511db7f
commit df178472d1
1 changed files with 65 additions and 64 deletions
Download Patch File Download Diff File Expand all files Collapse all files

129
README.md
View File

@@ -5,7 +5,8 @@ Open-source vector similarity search for Postgres
Store your vectors with the rest of your data. Supports: Store your vectors with the rest of your data. Supports:
- exact and approximate nearest neighbor search - exact and approximate nearest neighbor search
- L2 distance, inner product, and cosine distance - single-precision, half-precision, binary, and sparse vectors
- L2 distance, inner product, cosine distance, L1 distance, Hamming distance, and Jaccard distance
- any [language](#languages) with a Postgres client - any [language](#languages) with a Postgres client
Plus [ACID](https://en.wikipedia.org/wiki/ACID) compliance, point-in-time recovery, JOINs, and all of the other [great features](https://www.postgresql.org/about/) of Postgres Plus [ACID](https://en.wikipedia.org/wiki/ACID) compliance, point-in-time recovery, JOINs, and all of the other [great features](https://www.postgresql.org/about/) of Postgres
@@ -81,7 +82,7 @@ Get the nearest neighbors by L2 distance
SELECT * FROM items ORDER BY embedding <-> '[3,1,2]' LIMIT 5; SELECT * FROM items ORDER BY embedding <-> '[3,1,2]' LIMIT 5;
``` ```
Also supports inner product (`<#>`), cosine distance (`<=>`), and L1 distance (`<+>`, unreleased) Also supports inner product (`<#>`), cosine distance (`<=>`), and L1 distance (`<+>`, added in 0.7.0)
Note: `<#>` returns the negative inner product since Postgres only supports `ASC` order index scans on operators Note: `<#>` returns the negative inner product since Postgres only supports `ASC` order index scans on operators
@@ -143,7 +144,7 @@ Supported distance functions are:
- `<->` - L2 distance - `<->` - L2 distance
- `<#>` - (negative) inner product - `<#>` - (negative) inner product
- `<=>` - cosine distance - `<=>` - cosine distance
- `<+>` - L1 distance (unreleased) - `<+>` - L1 distance (added in 0.7.0)
Get the nearest neighbors to a row Get the nearest neighbors to a row
@@ -228,19 +229,19 @@ Cosine distance
CREATE INDEX ON items USING hnsw (embedding vector_cosine_ops); CREATE INDEX ON items USING hnsw (embedding vector_cosine_ops);
``` ```
L1 distance - unreleased L1 distance - added in 0.7.0
```sql ```sql
CREATE INDEX ON items USING hnsw (embedding vector_l1_ops); CREATE INDEX ON items USING hnsw (embedding vector_l1_ops);
``` ```
Hamming distance - unreleased Hamming distance - added in 0.7.0
```sql ```sql
CREATE INDEX ON items USING hnsw (embedding bit_hamming_ops); CREATE INDEX ON items USING hnsw (embedding bit_hamming_ops);
``` ```
Jaccard distance - unreleased Jaccard distance - added in 0.7.0
```sql ```sql
CREATE INDEX ON items USING hnsw (embedding bit_jaccard_ops); CREATE INDEX ON items USING hnsw (embedding bit_jaccard_ops);
@@ -249,9 +250,9 @@ CREATE INDEX ON items USING hnsw (embedding bit_jaccard_ops);
Supported types are: Supported types are:
- `vector` - up to 2,000 dimensions - `vector` - up to 2,000 dimensions
- `halfvec` - up to 4,000 dimensions (unreleased) - `halfvec` - up to 4,000 dimensions (added in 0.7.0)
- `bit` - up to 64,000 dimensions (unreleased) - `bit` - up to 64,000 dimensions (added in 0.7.0)
- `sparsevec` - up to 1,000 non-zero elements (unreleased) - `sparsevec` - up to 1,000 non-zero elements (added in 0.7.0)
### Index Options ### Index Options
@@ -356,7 +357,7 @@ Cosine distance
CREATE INDEX ON items USING ivfflat (embedding vector_cosine_ops) WITH (lists = 100); CREATE INDEX ON items USING ivfflat (embedding vector_cosine_ops) WITH (lists = 100);
``` ```
Hamming distance - unreleased Hamming distance - added in 0.7.0
```sql ```sql
CREATE INDEX ON items USING ivfflat (embedding bit_hamming_ops) WITH (lists = 100); CREATE INDEX ON items USING ivfflat (embedding bit_hamming_ops) WITH (lists = 100);
@@ -365,8 +366,8 @@ CREATE INDEX ON items USING ivfflat (embedding bit_hamming_ops) WITH (lists = 10
Supported types are: Supported types are:
- `vector` - up to 2,000 dimensions - `vector` - up to 2,000 dimensions
- `halfvec` - up to 4,000 dimensions (unreleased) - `halfvec` - up to 4,000 dimensions (added in 0.7.0)
- `bit` - up to 64,000 dimensions (unreleased) - `bit` - up to 64,000 dimensions (added in 0.7.0)
### Query Options ### Query Options
@@ -440,9 +441,9 @@ Use [partitioning](https://www.postgresql.org/docs/current/ddl-partitioning.html
CREATE TABLE items (embedding vector(3), category_id int) PARTITION BY LIST(category_id); CREATE TABLE items (embedding vector(3), category_id int) PARTITION BY LIST(category_id);
``` ```
## Half Vectors ## Half-Precision Vectors
*Unreleased* *Added in 0.7.0*
Use the `halfvec` type to store half-precision vectors Use the `halfvec` type to store half-precision vectors
@@ -450,9 +451,9 @@ Use the `halfvec` type to store half-precision vectors
CREATE TABLE items (id bigserial PRIMARY KEY, embedding halfvec(3)); CREATE TABLE items (id bigserial PRIMARY KEY, embedding halfvec(3));
``` ```
## Half Indexing ## Half-Precision Indexing
*Unreleased* *Added in 0.7.0*
Index vectors at half precision for smaller indexes and faster build times Index vectors at half precision for smaller indexes and faster build times
@@ -475,23 +476,23 @@ CREATE TABLE items (id bigserial PRIMARY KEY, embedding bit(3));
INSERT INTO items (embedding) VALUES ('000'), ('111'); INSERT INTO items (embedding) VALUES ('000'), ('111');
``` ```
Get the nearest neighbors by Hamming distance Get the nearest neighbors by Hamming distance (added in 0.7.0)
```sql
SELECT * FROM items ORDER BY bit_count(embedding # '101') LIMIT 5;
```
Or (unreleased)
```sql ```sql
SELECT * FROM items ORDER BY embedding <~> '101' LIMIT 5; SELECT * FROM items ORDER BY embedding <~> '101' LIMIT 5;
``` ```
Or (before 0.7.0)
```sql
SELECT * FROM items ORDER BY bit_count(embedding # '101') LIMIT 5;
```
Also supports Jaccard distance (`<%>`) Also supports Jaccard distance (`<%>`)
## Binary Quantization ## Binary Quantization
*Unreleased* *Added in 0.7.0*
Use expression indexing for binary quantization Use expression indexing for binary quantization
@@ -515,7 +516,7 @@ SELECT * FROM (
## Sparse Vectors ## Sparse Vectors
*Unreleased* *Added in 0.7.0*
Use the `sparsevec` type to store sparse vectors Use the `sparsevec` type to store sparse vectors
@@ -550,7 +551,7 @@ You can use [Reciprocal Rank Fusion](https://github.com/pgvector/pgvector-python
## Indexing Subvectors ## Indexing Subvectors
*Unreleased* *Added in 0.7.0*
Use expression indexing to index subvectors Use expression indexing to index subvectors
@@ -864,23 +865,23 @@ Operator | Description | Added
\+ | element-wise addition | \+ | element-wise addition |
\- | element-wise subtraction | \- | element-wise subtraction |
\* | element-wise multiplication | 0.5.0 \* | element-wise multiplication | 0.5.0
\|\| | concatenate | unreleased \|\| | concatenate | 0.7.0
<-> | Euclidean distance | <-> | Euclidean distance |
<#> | negative inner product | <#> | negative inner product |
<=> | cosine distance | <=> | cosine distance |
<+> | taxicab distance | unreleased <+> | taxicab distance | 0.7.0
### Vector Functions ### Vector Functions
Function | Description | Added Function | Description | Added
--- | --- | --- --- | --- | ---
binary_quantize(vector) → bit | binary quantize | unreleased binary_quantize(vector) → bit | binary quantize | 0.7.0
cosine_distance(vector, vector) → double precision | cosine distance | cosine_distance(vector, vector) → double precision | cosine distance |
inner_product(vector, vector) → double precision | inner product | inner_product(vector, vector) → double precision | inner product |
l1_distance(vector, vector) → double precision | taxicab distance | 0.5.0 l1_distance(vector, vector) → double precision | taxicab distance | 0.5.0
l2_distance(vector, vector) → double precision | Euclidean distance | l2_distance(vector, vector) → double precision | Euclidean distance |
l2_normalize(vector) → vector | Normalize with Euclidean norm | unreleased l2_normalize(vector) → vector | Normalize with Euclidean norm | 0.7.0
subvector(vector, integer, integer) → vector | subvector | unreleased subvector(vector, integer, integer) → vector | subvector | 0.7.0
vector_dims(vector) → integer | number of dimensions | vector_dims(vector) → integer | number of dimensions |
vector_norm(vector) → double precision | Euclidean norm | vector_norm(vector) → double precision | Euclidean norm |
@@ -899,35 +900,35 @@ Each half vector takes `2 * dimensions + 8` bytes of storage. Each element is a
Operator | Description | Added Operator | Description | Added
--- | --- | --- --- | --- | ---
\+ | element-wise addition | unreleased \+ | element-wise addition | 0.7.0
\- | element-wise subtraction | unreleased \- | element-wise subtraction | 0.7.0
\* | element-wise multiplication | unreleased \* | element-wise multiplication | 0.7.0
\|\| | concatenate | unreleased \|\| | concatenate | 0.7.0
<-> | Euclidean distance | unreleased <-> | Euclidean distance | 0.7.0
<#> | negative inner product | unreleased <#> | negative inner product | 0.7.0
<=> | cosine distance | unreleased <=> | cosine distance | 0.7.0
<+> | taxicab distance | unreleased <+> | taxicab distance | 0.7.0
### Halfvec Functions ### Halfvec Functions
Function | Description | Added Function | Description | Added
--- | --- | --- --- | --- | ---
binary_quantize(halfvec) → bit | binary quantize | unreleased binary_quantize(halfvec) → bit | binary quantize | 0.7.0
cosine_distance(halfvec, halfvec) → double precision | cosine distance | unreleased cosine_distance(halfvec, halfvec) → double precision | cosine distance | 0.7.0
inner_product(halfvec, halfvec) → double precision | inner product | unreleased inner_product(halfvec, halfvec) → double precision | inner product | 0.7.0
l1_distance(halfvec, halfvec) → double precision | taxicab distance | unreleased l1_distance(halfvec, halfvec) → double precision | taxicab distance | 0.7.0
l2_distance(halfvec, halfvec) → double precision | Euclidean distance | unreleased l2_distance(halfvec, halfvec) → double precision | Euclidean distance | 0.7.0
l2_norm(halfvec) → double precision | Euclidean norm | unreleased l2_norm(halfvec) → double precision | Euclidean norm | 0.7.0
l2_normalize(halfvec) → halfvec | Normalize with Euclidean norm | unreleased l2_normalize(halfvec) → halfvec | Normalize with Euclidean norm | 0.7.0
subvector(halfvec, integer, integer) → halfvec | subvector | unreleased subvector(halfvec, integer, integer) → halfvec | subvector | 0.7.0
vector_dims(halfvec) → integer | number of dimensions | unreleased vector_dims(halfvec) → integer | number of dimensions | 0.7.0
### Halfvec Aggregate Functions ### Halfvec Aggregate Functions
Function | Description | Added Function | Description | Added
--- | --- | --- --- | --- | ---
avg(halfvec) → halfvec | average | unreleased avg(halfvec) → halfvec | average | 0.7.0
sum(halfvec) → halfvec | sum | unreleased sum(halfvec) → halfvec | sum | 0.7.0
### Bit Type ### Bit Type
@@ -937,15 +938,15 @@ Each bit vector takes `dimensions / 8 + 8` bytes of storage. See the [Postgres d
Operator | Description | Added Operator | Description | Added
--- | --- | --- --- | --- | ---
<~> | Hamming distance | unreleased <~> | Hamming distance | 0.7.0
<%> | Jaccard distance | unreleased <%> | Jaccard distance | 0.7.0
### Bit Functions ### Bit Functions
Function | Description | Added Function | Description | Added
--- | --- | --- --- | --- | ---
hamming_distance(bit, bit) → double precision | Hamming distance | unreleased hamming_distance(bit, bit) → double precision | Hamming distance | 0.7.0
jaccard_distance(bit, bit) → double precision | Jaccard distance | unreleased jaccard_distance(bit, bit) → double precision | Jaccard distance | 0.7.0
### Sparsevec Type ### Sparsevec Type
@@ -955,21 +956,21 @@ Each sparse vector takes `8 * non-zero elements + 16` bytes of storage. Each ele
Operator | Description | Added Operator | Description | Added
--- | --- | --- --- | --- | ---
<-> | Euclidean distance | unreleased <-> | Euclidean distance | 0.7.0
<#> | negative inner product | unreleased <#> | negative inner product | 0.7.0
<=> | cosine distance | unreleased <=> | cosine distance | 0.7.0
<+> | taxicab distance | unreleased <+> | taxicab distance | 0.7.0
### Sparsevec Functions ### Sparsevec Functions
Function | Description | Added Function | Description | Added
--- | --- | --- --- | --- | ---
cosine_distance(sparsevec, sparsevec) → double precision | cosine distance | unreleased cosine_distance(sparsevec, sparsevec) → double precision | cosine distance | 0.7.0
inner_product(sparsevec, sparsevec) → double precision | inner product | unreleased inner_product(sparsevec, sparsevec) → double precision | inner product | 0.7.0
l1_distance(sparsevec, sparsevec) → double precision | taxicab distance | unreleased l1_distance(sparsevec, sparsevec) → double precision | taxicab distance | 0.7.0
l2_distance(sparsevec, sparsevec) → double precision | Euclidean distance | unreleased l2_distance(sparsevec, sparsevec) → double precision | Euclidean distance | 0.7.0
l2_norm(sparsevec) → double precision | Euclidean norm | unreleased l2_norm(sparsevec) → double precision | Euclidean norm | 0.7.0
l2_normalize(sparsevec) → sparsevec | Normalize with Euclidean norm | unreleased l2_normalize(sparsevec) → sparsevec | Normalize with Euclidean norm | 0.7.0
## Installation Notes - Linux and Mac ## Installation Notes - Linux and Mac