feat(stac): add geoarrow and geoparquet

The were their own crates, but that seemed to be overkill.

.github/dependabot.yml

@@ -8,6 +8,10 @@ updates:
     directory: "/stac"
     schedule:
       interval: "weekly"
+  - package-ecosystem: "cargo"
+    directory: "/python"
+    schedule:
+      interval: "weekly"
   - package-ecosystem: "cargo"
     directory: "/pgstac"
     schedule:
@@ -32,3 +36,7 @@ updates:
     directory: "/stac-validate"
     schedule:
       interval: "weekly"
+  - package-ecosystem: "pip"
+    directory: "/scripts"
+    schedule:
+      interval: "weekly"

.github/workflows/ci.yml

@@ -22,12 +22,13 @@ jobs:
         args:
           - "-p stac"
           - "-p stac -F reqwest"
+          - "-p stac -F geoarrow"
+          - "-p stac -F geoparquet"
+          - "-p stac -F geoparquet-compression"
           - "-p stac-api"
           - "-p stac -p stac-api -F geo"
-          - "-p stac-arrow"
           - "-p stac-async"
           - "-p stac-cli --no-default-features"
-          - "-p stac-geoparquet"
           - "-p stac-server --no-default-features"
           - "-p stac-server --no-default-features -F axum"
           - "-p stac-server --no-default-features -F memory-item-search"

Cargo.toml

@@ -5,21 +5,17 @@ members = [
     "pgstac",
     "python",
     "stac-api",
-    "stac-arrow",
     "stac-async",
     "stac-cli",
-    "stac-geoparquet",
     "stac-server",
     "stac-validate",
 ]
 default-members = [
     "python",
     "stac",
     "stac-api",
-    "stac-arrow",
     "stac-async",
     "stac-cli",
-    "stac-geoparquet",
     "stac-server",
     "stac-validate",
 ]

scripts/requirements-stac-server.in

@@ -1 +1,2 @@
+setuptools
 stac-api-validator

scripts/requirements-stac-server.txt

@@ -4,7 +4,7 @@ attrs==24.2.0
     # via
     #   jsonschema
     #   referencing
-certifi==2024.7.4
+certifi==2024.8.30
     # via
     #   requests
     #   stac-api-validator
@@ -17,7 +17,7 @@ click==8.1.7
     #   stac-validator
 deepdiff==6.7.1
     # via stac-api-validator
-idna==3.7
+idna==3.8
     # via requests
 jsonschema==4.23.0
     # via
@@ -29,11 +29,11 @@ jsonschema-specifications==2023.12.1
     # via jsonschema
 more-itertools==8.14.0
     # via stac-api-validator
-numpy==2.0.1
+numpy==2.1.0
     # via shapely
 ordered-set==4.1.0
     # via deepdiff
-orjson==3.10.6
+orjson==3.10.7
     # via pystac
 pystac==1.10.1
     # via
@@ -65,7 +65,9 @@ rpds-py==0.20.0
     # via
     #   jsonschema
     #   referencing
-shapely==2.0.5
+setuptools==74.0.0
+    # via -r scripts/requirements-stac-server.in
+shapely==2.0.6
     # via stac-api-validator
 six==1.16.0
     # via python-dateutil

scripts/validate-stac-server

@@ -1,4 +1,10 @@
 #!/usr/bin/env sh
+#
+# Validate stac server using stac-api-validator.
+#
+# To use this script on macos, you'll need `timeout`, which is provided by Homebrew's coreutils:
+#
+#   brew install coreutils
 
 set -e
 

stac-cli/Cargo.toml

@@ -22,10 +22,11 @@ clap = { version = "4", features = ["derive"] }
 reqwest = "0.12"
 serde = "1"
 serde_json = "1"
-stac = { version = "0.8", path = "../stac" }
+stac = { version = "0.8", path = "../stac", features = [
+    "geoparquet-compression",
+] }
 stac-api = { version = "0.4", path = "../stac-api" }
 stac-async = { version = "0.5", path = "../stac-async" }
-stac-geoparquet = { version = "0.0.1", path = "../stac-geoparquet" }
 stac-server = { version = "0.1", path = "../stac-server", features = [
     "memory-item-search",
     "tokio",

stac-cli/src/error.rs

@@ -33,10 +33,6 @@ pub enum Error {
     #[error(transparent)]
     StacAsync(#[from] stac_async::Error),
 
-    /// [stac_geoparquet::Error]
-    #[error(transparent)]
-    StacGeoparquet(#[from] stac_geoparquet::Error),
-
     /// [stac_server::Error]
     #[error(transparent)]
     StacServer(#[from] stac_server::Error),

stac-cli/src/format.rs

@@ -26,10 +26,10 @@ impl Format {
             match *self {
                 Format::Parquet => {
                     let item_collection = if let Some(url) = stac::href_to_url(href) {
-                        stac_geoparquet::from_reader(reqwest::blocking::get(url)?.bytes()?)?
+                        stac::geoparquet::from_reader(reqwest::blocking::get(url)?.bytes()?)?
                     } else {
                         let file = File::open(href)?;
-                        stac_geoparquet::from_reader(file)?
+                        stac::geoparquet::from_reader(file)?
                     };
                     serde_json::from_value(serde_json::to_value(item_collection)?)
                         .map_err(Error::from)
@@ -41,7 +41,7 @@ impl Format {
                 Format::Parquet => {
                     let mut buf = Vec::new();
                     let _ = std::io::stdin().read_to_end(&mut buf)?;
-                    let item_collection = stac_geoparquet::from_reader(Bytes::from(buf))?;
+                    let item_collection = stac::geoparquet::from_reader(Bytes::from(buf))?;
                     serde_json::from_value(serde_json::to_value(item_collection)?)
                         .map_err(Error::from)
                 }

stac-cli/src/runner.rs

@@ -48,7 +48,7 @@ where
                 }
                 Format::Parquet => {
                     if let Some(value) = value.to_stac() {
-                        stac_geoparquet::to_writer(&mut self.writer, value)?;
+                        stac::geoparquet::to_writer(&mut self.writer, value)?;
                     } else {
                         writeln!(self.writer, "{}", value)?;
                     }

stac/CHANGELOG.md

@@ -11,6 +11,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 - `Bbox` ([#303](https://github.com/stac-utils/stac-rs/pull/303))
 - Functions to create collections from items ([#304](https://github.com/stac-utils/stac-rs/pull/304))
 - Default implementation for `Version` ([#309](https://github.com/stac-utils/stac-rs/pull/309))
+- Experimental GeoParquet and GeoArrow support ([#316](https://github.com/stac-utils/stac-rs/pull/316))
 
 ### Changed
 

stac/Cargo.toml

@@ -13,15 +13,40 @@ categories = ["science", "data-structures"]
 [features]
 gdal = ["dep:gdal", "dep:gdal-sys"]
 geo = ["dep:geo"]
+geoarrow = [
+    "dep:geoarrow",
+    "dep:arrow-array",
+    "dep:arrow-cast",
+    "dep:arrow-json",
+    "dep:arrow-schema",
+    "dep:geo-types",
+]
+geoparquet = ["geoarrow/parquet", "dep:parquet"]
+geoparquet-compression = [
+    "geoparquet",
+    "geoarrow/parquet_compression",
+    "parquet/snap",
+    "parquet/brotli",
+    "parquet/flate2",
+    "parquet/lz4",
+    "parquet/zstd",
+]
 reqwest = ["dep:reqwest"]
 
 [dependencies]
+arrow-array = { version = "52", optional = true }
+arrow-cast = { version = "52", optional = true }
+arrow-json = { version = "52", optional = true }
+arrow-schema = { version = "52", optional = true }
 chrono = { version = "0.4", features = ["serde"] }
 gdal = { version = "0.17", optional = true }
 gdal-sys = { version = "0.10", optional = true }
 geo = { version = "0.28", optional = true }
+geo-types = { version = "0.7", optional = true }
+geoarrow = { git = "https://github.com/geoarrow/geoarrow-rs", rev = "6b877486bf98f280bc04b589eb7ce25b20e629f0", optional = true }
 geojson = { version = "0.24" }
 log = { version = "0.4" }
+parquet = { version = "52", default-features = false, optional = true }
 reqwest = { version = "0.12", optional = true, features = ["json", "blocking"] }
 serde = { version = "1", features = ["derive"] }
 serde_json = { version = "1", features = ["preserve_order"] }
@@ -30,6 +55,7 @@ url = "2"
 
 [dev-dependencies]
 assert-json-diff = "2"
+bytes = "1"
 rstest = "0.22"
 
 [package.metadata.docs.rs]