feat: update geoarrow and duckdb (#562)

- Closes #385
stac-utils · Dec 24, 2024 · 0a5a045 · 0a5a045
1 parent 55703ce
commit 0a5a045
Show file tree

Hide file tree

Showing 27 changed files with 517 additions and 661 deletions.
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
@@ -59,7 +59,6 @@ jobs:
         working-directory: crates/cli
     steps:
       - uses: actions/checkout@v4
-      - uses: dtolnay/rust-toolchain@1.81
       - uses: Swatinem/rust-cache@v2
       - name: Test
         run: cargo test -p stac-cli
@@ -71,7 +70,6 @@ jobs:
         working-directory: crates/cli
     steps:
       - uses: actions/checkout@v4
-      - uses: dtolnay/rust-toolchain@1.81
       - uses: Swatinem/rust-cache@v2
       - uses: actions/setup-python@v5
         with:
@@ -129,7 +127,7 @@ jobs:
         with:
           enable-cache: true
       - name: Test
-        run: cargo test -p pgstac
+        run: cargo test -p pgstac --all-features
       - name: Install Python dependencies
         run: uv sync --group stac-api-validator
       - name: Validate

diff --git a/Cargo.toml b/Cargo.toml
@@ -25,14 +25,14 @@ homepage = "https://stac-utils.github.io/stac-rs"
 repository = "https://github.com/stac-utils/stac-rs"
 license = "MIT OR Apache-2.0"
 categories = ["science", "data-structures"]
-rust-version = "1.80"
+rust-version = "1.82"
 
 [workspace.dependencies]
-arrow = "52.2"
-arrow-array = "52.2"
-arrow-cast = "52.2"
-arrow-json = "52.2"
-arrow-schema = "52.2"
+arrow = "53.1.0"
+arrow-array = "53.1.0"
+arrow-cast = "53.1.0"
+arrow-json = "53.1.0"
+arrow-schema = "53.1.0"
 assert-json-diff = "2.0"
 assert_cmd = "2.0"
 async-stream = "0.3.6"
@@ -43,23 +43,23 @@ bytes = "1.7"
 chrono = "0.4.38"
 clap = "4.5"
 cql2 = "0.3.0"
-duckdb = "=1.0.0"
+duckdb = "1.1.1"
 fluent-uri = "0.3.1"
 futures = "0.3.31"
 geo = "0.29.3"
 geo-types = "0.7.13"
-geoarrow = "0.3.0"
+geoarrow = "0.4.0-beta.3"
 geojson = "0.24.1"
 http = "1.1"
 jsonschema = { version = "0.22.3", default-features = false }
-libduckdb-sys = "=1.0.0"
+libduckdb-sys = "1.1.1"
 log = "0.4.22"
 mime = "0.3.17"
 mockito = "1.5"
 object_store = "0.11.0"
 openssl = { version = "0.10.68", features = ["vendored"] }
 openssl-src = "=300.4.1" # joinked from https://github.com/iopsystems/rpc-perf/commit/705b290d2105af6f33150da04b217422c6d68701#diff-2e9d962a08321605940b5a657135052fbcef87b5e360662bb527c96d9a615542R41 to cross-compile Python
-parquet = { version = "52.2", default-features = false }
+parquet = { version = "53.1.0", default-features = false }
 pgstac = { version = "0.2.2", path = "crates/pgstac" }
 pyo3 = "0.23.3"
 pythonize = "0.23.0"

diff --git a/README.md b/README.md
@@ -8,6 +8,14 @@
 
 Command Line Interface (CLI) and Rust libraries for the [SpatioTemporal Asset Catalog (STAC)](https://stacspec.org/) specification.
 
+There's a couple Python projects based on **stac-rs** that might be of interest to you, as well:
+
+- [stacrs](https://github.com/gadomski/stacrs) provides a Python API to **stac-rs**, including
+  - Reading and writing [stac-geoparquet](https://github.com/stac-utils/stac-geoparquet)
+  - Migrating to [STAC v1.1](https://github.com/radiantearth/stac-spec/releases/tag/v1.1.0)
+  - [More...](https://www.gadom.ski/posts/stacrs-python-v0-1/)
+- [pgstacrs](https://github.com/stac-utils/pgstacrs) is a Python library for working with [pgstac](https://github.com/stac-utils/pgstac)
+
 ## Command line interface
 
 Our command line interface (CLI) can query STAC APIs, validate STAC, and more.

diff --git a/crates/api/CHANGELOG.md b/crates/api/CHANGELOG.md
@@ -6,6 +6,14 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 
 ## [Unreleased]
 
+### Added
+
+- Convenience methods on `Search` ([#562](https://github.com/stac-utils/stac-rs/pull/562))
+
+### Changed
+
+- Many vectors on `Search` and `Items` are no longer `Options` ([#562](https://github.com/stac-utils/stac-rs/pull/562))
+
 ## [0.6.2] - 2024-11-12
 
 ### Added

diff --git a/crates/api/src/client.rs b/crates/api/src/client.rs
@@ -147,7 +147,7 @@ impl Client {
     /// use futures::StreamExt;
     ///
     /// let client = Client::new("https://planetarycomputer.microsoft.com/api/stac/v1").unwrap();
-    /// let mut search = Search { collections: Some(vec!["sentinel-2-l2a".to_string()]), ..Default::default() };
+    /// let mut search = Search { collections: vec!["sentinel-2-l2a".to_string()], ..Default::default() };
     /// # tokio_test::block_on(async {
     /// let items: Vec<_> = client
     ///     .search(search)
@@ -272,7 +272,7 @@ impl BlockingClient {
     /// use stac_api::{Search, BlockingClient};
     ///
     /// let client = BlockingClient::new("https://planetarycomputer.microsoft.com/api/stac/v1").unwrap();
-    /// let mut search = Search { collections: Some(vec!["sentinel-2-l2a".to_string()]), ..Default::default() };
+    /// let mut search = Search { collections: vec!["sentinel-2-l2a".to_string()], ..Default::default() };
     /// let items: Vec<_> = client
     ///     .search(search)
     ///     .unwrap()
@@ -430,7 +430,7 @@ mod tests {
 
         let client = Client::new(&server.url()).unwrap();
         let mut search = Search {
-            collections: Some(vec!["sentinel-2-l2a".to_string()]),
+            collections: vec!["sentinel-2-l2a".to_string()],
             ..Default::default()
         };
         search.items.limit = Some(1);

diff --git a/crates/api/src/error.rs b/crates/api/src/error.rs
@@ -23,7 +23,7 @@ pub enum Error {
 
     /// [cql2::Error]
     #[error(transparent)]
-    Cql2(#[from] cql2::Error),
+    Cql2(#[from] Box<cql2::Error>),
 
     /// [geojson::Error]
     #[error(transparent)]

diff --git a/crates/api/src/filter.rs b/crates/api/src/filter.rs
@@ -23,7 +23,7 @@ impl Filter {
         match self {
             Filter::Cql2Json(_) => Ok(self),
             Filter::Cql2Text(text) => {
-                let expr = cql2::parse_text(&text)?;
+                let expr = cql2::parse_text(&text).map_err(Box::new)?;
                 Ok(Filter::Cql2Json(serde_json::from_value(
                     serde_json::to_value(expr)?,
                 )?))
@@ -37,7 +37,7 @@ impl Filter {
             Filter::Cql2Text(_) => Ok(self),
             Filter::Cql2Json(json) => {
                 let expr: Expr = serde_json::from_value(Value::Object(json))?;
-                Ok(Filter::Cql2Text(expr.to_text()?))
+                Ok(Filter::Cql2Text(expr.to_text().map_err(Box::new)?))
             }
         }
     }

diff --git a/crates/api/src/items.rs b/crates/api/src/items.rs
@@ -28,8 +28,8 @@ pub struct Items {
     pub fields: Option<Fields>,
 
     /// Fields by which to sort results.
-    #[serde(skip_serializing_if = "Option::is_none")]
-    pub sortby: Option<Vec<Sortby>>,
+    #[serde(skip_serializing_if = "Vec::is_empty", default)]
+    pub sortby: Vec<Sortby>,
 
     /// Recommended to not be passed, but server must only accept
     /// <http://www.opengis.net/def/crs/OGC/1.3/CRS84> as a valid value, may
@@ -281,14 +281,14 @@ impl Items {
     ///     ..Default::default()
     /// };
     /// let search = items.search_collection("collection-id");
-    /// assert_eq!(search.collections.unwrap(), vec!["collection-id"]);
+    /// assert_eq!(search.collections, vec!["collection-id"]);
     /// ```
     pub fn search_collection(self, collection_id: impl ToString) -> Search {
         Search {
             items: self,
             intersects: None,
-            ids: None,
-            collections: Some(vec![collection_id.to_string()]),
+            ids: Vec::new(),
+            collections: vec![collection_id.to_string()],
         }
     }
 
@@ -327,13 +327,18 @@ impl TryFrom<Items> for GetItems {
             }),
             datetime: items.datetime,
             fields: items.fields.map(|fields| fields.to_string()),
-            sortby: items.sortby.map(|sortby| {
-                sortby
-                    .into_iter()
-                    .map(|s| s.to_string())
-                    .collect::<Vec<_>>()
-                    .join(",")
-            }),
+            sortby: if items.sortby.is_empty() {
+                None
+            } else {
+                Some(
+                    items
+                        .sortby
+                        .into_iter()
+                        .map(|s| s.to_string())
+                        .collect::<Vec<_>>()
+                        .join(","),
+                )
+            },
             filter_crs: items.filter_crs,
             filter_lang: if filter.is_some() {
                 Some("cql2-text".to_string())
@@ -364,15 +369,16 @@ impl TryFrom<GetItems> for Items {
             None
         };
 
-        let sortby = if let Some(value) = get_items.sortby {
-            let mut sortby = Vec::new();
-            for s in value.split(',') {
-                sortby.push(s.parse().expect("infallible"));
-            }
-            Some(sortby)
-        } else {
-            None
-        };
+        let sortby = get_items
+            .sortby
+            .map(|s| {
+                let mut sortby = Vec::new();
+                for s in s.split(',') {
+                    sortby.push(s.parse().expect("infallible"));
+                }
+                sortby
+            })
+            .unwrap_or_default();
 
         Ok(Items {
             limit: get_items.limit.map(|limit| limit.parse()).transpose()?,
@@ -452,7 +458,7 @@ mod tests {
             }
         );
         assert_eq!(
-            items.sortby.unwrap(),
+            items.sortby,
             vec![Sortby {
                 field: "foo".to_string(),
                 direction: Direction::Descending,
@@ -478,10 +484,10 @@ mod tests {
                 include: vec!["foo".to_string()],
                 exclude: vec!["bar".to_string()],
             }),
-            sortby: Some(vec![Sortby {
+            sortby: vec![Sortby {
                 field: "foo".to_string(),
                 direction: Direction::Descending,
-            }]),
+            }],
             filter_crs: None,
             filter: Some(Filter::Cql2Text("dummy text".to_string())),
             query: None,

diff --git a/crates/api/src/python.rs b/crates/api/src/python.rs
@@ -40,12 +40,14 @@ pub fn search<'py>(
         .map(|query| pythonize::depythonize(&query))
         .transpose()?;
     let bbox = bbox.map(Bbox::try_from).transpose().map_err(Error::from)?;
-    let sortby = sortby.map(|sortby| {
-        Vec::<String>::from(sortby)
-            .into_iter()
-            .map(|s| s.parse::<Sortby>().unwrap()) // the parse is infallible
-            .collect::<Vec<_>>()
-    });
+    let sortby = sortby
+        .map(|sortby| {
+            Vec::<String>::from(sortby)
+                .into_iter()
+                .map(|s| s.parse::<Sortby>().unwrap()) // the parse is infallible
+                .collect::<Vec<_>>()
+        })
+        .unwrap_or_default();
     let filter = filter
         .map(|filter| match filter {
             StringOrDict::Dict(cql_json) => pythonize::depythonize(&cql_json).map(Filter::Cql2Json),
@@ -80,8 +82,8 @@ pub fn search<'py>(
                 .map_err(|err| PyValueError::new_err(err.to_string())),
         })
         .transpose()?;
-    let ids = ids.map(|ids| ids.into());
-    let collections = collections.map(|ids| ids.into());
+    let ids = ids.map(|ids| ids.into()).unwrap_or_default();
+    let collections = collections.map(|ids| ids.into()).unwrap_or_default();
     Ok(Search {
         items,
         intersects,