This is an automated email from the ASF dual-hosted git repository.
jiayu pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/sedona-spatialbench.git
The following commit(s) were added to refs/heads/main by this push:
new c85f331 docs: Fixes typos on SpatialBench homepage, fixes mobile nav
black text, removes PostGIS line (#38)
c85f331 is described below
commit c85f331308d8bfebf1345365f6a0ede72fec362a
Author: Kelly-Ann Dolor <[email protected]>
AuthorDate: Tue Sep 23 21:38:44 2025 -0700
docs: Fixes typos on SpatialBench homepage, fixes mobile nav black text,
removes PostGIS line (#38)
* Fixes typos on SpatialBench homepage, fixes mobile nav black text,
removes PostGIS line
* resolving comment
---
docs/index.md | 48 +++++++++++++++++++++++-------------------
docs/single-node-benchmarks.md | 13 ++++++------
docs/stylesheets/extra.css | 48 +++++++++++++++++++++++++++++++-----------
3 files changed, 68 insertions(+), 41 deletions(-)
diff --git a/docs/index.md b/docs/index.md
index dbcfeb1..55046a6 100644
--- a/docs/index.md
+++ b/docs/index.md
@@ -19,44 +19,48 @@ title: SpatialBench
under the License.
-->
-SpatialBench is a benchmark for assessing geospatial SQL analytics query
performance across database systems.
+SpatialBench is a benchmark for assessing geospatial SQL analytics query
performance across database systems, making it easy to run tests on a realistic
dataset with any query engine.
-SpatialBench makes it easy to run spatial benchmarks on a realistic dataset
with any query engine.
-
-The methodology is unbiased and the benchmarks in any environment to compare
relative performance between runtimes.
+The methodology is unbiased, allowing you to run the benchmarks in any
environment to compare the relative performance between runtimes.
## Why SpatialBench
-SpatialBench is a geospatial benchmark for testing and optimizing spatial
analytical query performance in database systems. Inspired by the SSB and NYC
taxi data, it combines realistic urban mobility scenarios with a star schema
extended with spatial attributes like pickup/dropoff points, zones, and
building footprints.
+SpatialBench was created because standard database benchmarks don't adequately
test the unique demands of geospatial queries. SpatialBench provides an
open-source, standardized, and scalable framework designed specifically for
geospatial analytics.
+
+Inspired by the Star Schema Benchmark (SSB) and NYC taxi data, SpatialBench
combines realistic urban mobility scenarios
+with a star schema extended with spatial attributes like pickup/dropoff
points, zones, and building footprints.
This design enables evaluation of the following geospatial operations:
-* spatial joins
-* distance queries
-* aggregations
-* point-in-polygon analysis
+* Spatial joins
+* Distance queries
+* Aggregations
+* Point-in-polygon analysis
+
+Let's dive into the advantages of SpatialBench.
-Let’s dive into the advantages of SpatialBench.
+## Key Features
-## Key advantages
+To ensure fair and comprehensive testing, SpatialBench provides the following
advantages:
-* Uses spatial datasets with geometry columns.
-* Includes queries with different spatial predicates.
-* Easily reproducible results.
-* Includes a dataset generator to so results are reproducible.
-* The scale factors of the datasets can be changed so that you can run the
queries locally, in a data warehouse, or on a large cluster in the cloud.
-* All the specifications used to run the benchmarks are documented, and the
methodology is unbiased.
-* The code is open source, allowing the community to provide feedback and keep
the benchmarks up-to-date and reliable over time.
+* Features realistic spatial datasets with native geometry columns.
+* Includes a suite of queries that test various operations such as spatial
predicates and joins.
+* Provides a built-in synthetic data generator for creating consistent test
data.
+* Offers a configurable scale factor to benchmark performance across various
+ environments, from a single local machine to a large-scale cloud cluster.
+* Ensures consistent and reproducible benchmark results across all
environments.
+* Utilizes a fully documented and unbiased methodology to facilitate fair
comparisons.
+* Open-source and community-driven to foster transparency and continuous
improvement.
## Generate synthetic data
-Here’s how you can install the synthetic data generator:
+Here's how you can install the synthetic data generator:
```
cargo install --path ./spatialbench-cli
```
-Here’s how you can generate the synthetic dataset:
+Here's how you can generate the synthetic dataset:
```
spatialbench-cli -s 1 --format=parquet
@@ -66,7 +70,7 @@ See the project repository
[README](https://github.com/apache/sedona-spatialbenc
## Example query
-Here’s an example query that counts the number of trips that start within 500
meters of each building:
+Here's an example query that counts the number of trips that start within 500
meters of each building:
```sql
SELECT
@@ -80,7 +84,7 @@ GROUP BY b.b_buildingkey, b.b_name
ORDER BY nearby_pickup_count DESC;
```
-This query performs a distance join, followed by an aggregation. It’s a great
example of a query that’s useful for performance benchmarking a spatial engine
that can process vector geometries.
+This query performs a distance join, followed by an aggregation. It's a great
example of a query that's useful for performance benchmarking a spatial engine
that can process vector geometries.
## Join the community
diff --git a/docs/single-node-benchmarks.md b/docs/single-node-benchmarks.md
index 5a5d552..64e84c7 100644
--- a/docs/single-node-benchmarks.md
+++ b/docs/single-node-benchmarks.md
@@ -43,7 +43,7 @@ Here are the software versions used in this benchmark:
* DuckDB: 1.4.0
* SedonaDB: 0.1
-This benchmark report lists software versions, so it’s easy to track how
engine performance improves over time. We use the default settings of all
software unless otherwise noted. For DuckDB, we explicitly set
enable_external_file_cache to false to focus on the cold start queries runtime,
consistent with the other engines.
+This benchmark report lists software versions, so it's easy to track how
engine performance improves over time. We use the default settings of all
software unless otherwise noted. For DuckDB, we explicitly set
enable_external_file_cache to false to focus on the cold start queries runtime,
consistent with the other engines.
The code execution runtime includes the entire query runtime for all engines.
The query timeout is set to 1200 seconds.
@@ -51,7 +51,7 @@ The code execution runtime includes the entire query runtime
for all engines. Th
The GeoPandas queries are written in Python, since GeoPandas does not support
SQL. GeoPandas executes queries by loading data fully into memory and then
processing it directly.
-Since GeoPandas runs in a single thread and lacks a query optimizer, any
parallelization or optimization must be implemented manually. This benchmark
implemented a straightforward implementation that mirrors the SQL queries used
for other engines. If you’re a GeoPandas expert, we’d be glad to collaborate on
a more optimized and/or parallelized version.
+Since GeoPandas runs in a single thread and lacks a query optimizer, any
parallelization or optimization must be implemented manually. This benchmark
implemented a straightforward implementation that mirrors the SQL queries used
for other engines. If you're a GeoPandas expert, we'd be glad to collaborate on
a more optimized and/or parallelized version.
## Result analysis
@@ -65,7 +65,7 @@ SedonaDB consistently delivers strong results on heavier
joins, particularly Q10
### Geometric computations (Q5, Q7, Q9)
-SedonaDB is especially effective on intersection/IoU (Q9), showing substantial
efficiency improvements, while Q5 (convex hull aggregation) highlights areas
where DuckDB currently performs faster. SedonaDB’s overhead in geometry copying
in spatial aggregation is a known bottleneck and is planned for improvement.
+SedonaDB is especially effective on intersection/IoU (Q9), showing substantial
efficiency improvements, while Q5 (convex hull aggregation) highlights areas
where DuckDB currently performs faster. SedonaDB's overhead in geometry copying
in spatial aggregation is a known bottleneck and is planned for improvement.
### Nearest-neighbor joins (Q12)
@@ -79,7 +79,7 @@ SedonaDB demonstrates balanced strengths across all
categories and successfully
You can access and run the benchmark code in the [sedona-spatialbench
GitHub](https://github.com/apache/sedona-spatialbench) repository.
-It’s easy to generate the datasets locally or in the cloud. You can also run
the benchmarks locally or in the cloud.
+It's easy to generate the datasets locally or in the cloud. You can also run
the benchmarks locally or in the cloud.
The repository has an issue tracker where you can file bug reports or suggest
code improvements.
@@ -87,11 +87,10 @@ The repository has an issue tracker where you can file bug
reports or suggest co
It would be great to include other engines and databases in the future:
-* dask-geopandas for single-node parallelism across cores
-* PostGIS (Postgres SQL extension)
+* `dask-geopandas` for single-node parallelism across cores
* An R geospatial engine
-If you’re an expert in any of these technologies, we welcome you to take on
this project or reach out to us about collaborating.
+If you're an expert in any of these technologies, we welcome you to take on
this project or reach out to us about collaborating.
Note that compute engines designed for multi-node environments are
intentionally excluded from these single-node results for clarity and
simplicity.
diff --git a/docs/stylesheets/extra.css b/docs/stylesheets/extra.css
index 1b5a89f..be4ae4f 100644
--- a/docs/stylesheets/extra.css
+++ b/docs/stylesheets/extra.css
@@ -57,18 +57,6 @@
font-size: 0.65rem; /* NEW: Adjust font size */
}
-/* ==========================================================================
- Mobile Navigation Styles
- ==========================================================================
*/
-
-/* This targets the main container of the slide-out navigation on mobile */
-.md-nav--primary .md-nav__title,
-.md-nav__source {
- background-color: var(--color-red); /* Use your red color */
- box-shadow: none; /* Optional: removes the shadow */
-}
-
-
/* ==========================================================================
Logo Size Adjustment
==========================================================================
*/
@@ -82,3 +70,39 @@
height: 42px; /* This should match the Apache Sedona logo size */
width: auto; /* Ensures the width scales proportionally */
}
+
+/* ==========================================================================
+ Mobile Navigation Styles
+ ==========================================================================
*/
+
+/*
+ Force the entire mobile navigation header to be black, overriding theme
defaults.
+*/
+.md-nav--primary .md-nav__title,
+.md-nav--primary .md-nav__source {
+ background-color: var(--color-dark) !important;
+}
+
+/* ==========================================================================
+ Swap Logo ONLY in Mobile Navigation
+ ==========================================================================
*/
+
+/*
+ Target the logo link (<a> tag) directly and apply the new logo
+ as a background image, overriding the theme's default icon.
+*/
+.md-sidebar--primary .md-nav__title a.md-nav__button.md-logo {
+ background-image: url('/docs-overrides/.icons/sedona_logo_symbol_white.svg');
+ background-size: contain;
+ background-repeat: no-repeat;
+ background-position: center;
+
+ /* Hide the theme's default icon which is applied via a mask */
+ -webkit-mask: none;
+ mask: none;
+}
+
+/* Change the mobile nav header label text to white */
+label.md-nav__title {
+ color: #FFFFFF !important;
+}
\ No newline at end of file
