Skip to content

Commit e6172dd

Browse files
committed
refine docs for postgres
1 parent 7210f84 commit e6172dd

File tree

1 file changed

+18
-10
lines changed

1 file changed

+18
-10
lines changed

docs/step-by-step-guide-for-benchmark.md

Lines changed: 18 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -31,6 +31,11 @@
3131
- [Step 2. Modify the configuration](#step-2-modify-the-configuration-5)
3232
- [Step 3. Run the tests](#step-3-run-the-tests-5)
3333
- [Step 4. View the test results](#step-4-view-the-test-results-5)
34+
- [Postgres (Pgvector \& Pgvecto.rs)](#postgres-pgvector--pgvectors)
35+
- [Step 1. Create Server](#step-1-create-server)
36+
- [Step 2. Modify the configuration](#step-2-modify-the-configuration-6)
37+
- [Step 3. Run the tests](#step-3-run-the-tests-6)
38+
- [Step 4. View the test results](#step-4-view-the-test-results-6)
3439

3540
For any cloud vector database, the testing process follows the flowchart below:
3641

@@ -314,12 +319,11 @@ grep -E 'rps|mean_precision' $(ls -t)
314319

315320
![OpenSearchResults.jpg](../images/OpenSearchResults.png)
316321

317-
318-
## PGVector
322+
## Postgres (Pgvector & Pgvecto.rs)
319323

320324
### Step 1. Create Server
321325

322-
For deploying PGVector implemented in C, you can use Docker. Below is an example docker-compose.yaml configuration:
326+
For deploying Postgres with [pgvector](https://github.com/pgvector/pgvector) (a Postgres plugin written in C), you can use Docker. Below is an example `docker-compose.yaml` configuration:
323327

324328
```yaml
325329
version: '3'
@@ -335,7 +339,7 @@ services:
335339
- "5432:5432"
336340
```
337341
338-
Similarly, for PGVector written in Rust, Docker can also be utilized. Here's a corresponding docker-compose.yaml example:
342+
Similarly, for Postgres with [pgvecto.rs](https://github.com/tensorchord/pgvecto.rs) (a Postgres plugin written in Rust), Docker can also be utilized. Here's a corresponding `docker-compose.yaml` example:
339343

340344
```yaml
341345
version: '3'
@@ -353,16 +357,20 @@ services:
353357

354358
### Step 2. Modify the configuration
355359

356-
We have provided four configuration files for testing PGVector:
357-
for PGVector written in C:
360+
We have provided four configuration files for testing.
361+
362+
For pgvector:
363+
358364
- [pgvector_c_HNSW_single_node_laion-768-5m-ip.json](../experiments/needs_editing/pgvector_c_HNSW_single_node_laion-768-5m-ip.json)
359365
- [pgvector_c_HNSW_single_node_laion-768-5m-probability-ip.json](../experiments/needs_editing/pgvector_c_HNSW_single_node_laion-768-5m-probability-ip.json)
360-
for PGVector written in Rust:
366+
367+
For pgvecto.rs:
368+
361369
- [pgvector_rust_HNSW_single_node_laion-768-5m-ip.json](../experiments/needs_editing/pgvector_rust_HNSW_single_node_laion-768-5m-ip.json)
362370
- [pgvector_rust_HNSW_single_node_laion-768-5m-probability-ip.json](../experiments/needs_editing/pgvector_rust_HNSW_single_node_laion-768-5m-probability-ip.json)
363371

364-
After deploying your own PGVector service, you need to modify the `connection_params` fields in the configuration file.
365-
Additionally, you can append custom `search_params` for testing purposes.
372+
After deploying your own Postgres service, you need to modify the `connection_params` fields in the configuration file.
373+
Additionally, you can append custom `search_params` for testing purposes.
366374
You can also alter `upload_params` to modify the parameters for index creation.
367375

368376
```shell
@@ -387,4 +395,4 @@ cd results
387395
grep -E 'rps|mean_precision' $(ls -t)
388396
```
389397

390-
![PGVectorResults.png](../images/PGVectorResults.png)
398+
![PGVectorResults.png](../images/PGVectorResults.png)

0 commit comments

Comments
 (0)