add geoarrow GeoTable conversion. Derive default in most cases

Author: JosiahParry

.github/workflows/cargo-doc.yml

@@ -33,7 +33,6 @@ jobs:
         uses: actions/checkout@v4
 
       - uses: dtolnay/rust-toolchain@nightly
-        if: matrix.config.rust == 'nightly'
 
       - name: Docs
         run: cargo doc --no-deps --all-features

.zed/settings.json

@@ -1,6 +1,6 @@
 [package]
 name = "serde_esri"
-version = "0.1.0"
+version = "0.2.0"
 edition = "2021"
 license = "MIT OR Apache-2.0"
 repository = "https://github.com/josiahparry/serde_esri"

Cargo.lock

@@ -7,48 +7,73 @@ This crate provides representations of Esri JSON objects with [`serde::Deseriali
 `serde_esri` has two additional features `geo` and `geoarrow`. 
 
 - `geo` implements `From` for the Esri JSON objects.
-- `geoarrow` provides compatibility with arrow and geoarrow by implementing geoarrow geometry traits as well as providing a utility function `featureset_to_arrow()` which converts a `FeatureSet` to an arrow `RecordBatch`.
+- `geoarrow` provides compatibility with arrow and geoarrow by implementing geoarrow geometry traits as well as providing a utility function `featureset_to_geoarrow()` which converts a `FeatureSet` to an arrow `GeoTable`.
+
 
 ## Example usage: 
 
-In this example, we query a feature service and convert the response to an Arrow `RecordBatch`. This requires the `geoarrow` feature to be enabled.
+This example reads a few features from a feature service and returns a `FeatureSet` struct. It illustrates the use of the geo and geoarrow features. 
+
+```toml
+[dependencies]
+geo = "0.28.0"
+geoarrow = "0.2.0"
+reqwest = { version = "0.12.3", features = ["blocking"] }
+serde_esri = { version = "0.2.0", features = ["geo", "geoarrow"] }
+serde_json = "1.0.115"
+```
 
 ```rust
+use geo::{GeodesicArea, Polygon};
+use serde_esri::arrow_compat::featureset_to_geoarrow;
 use serde_esri::features::FeatureSet;
-use serde_esri::arrow_compat::featureset_to_arrow;
-
-#[tokio::main]
-async fn main() {
-    
-    // query url 
-    let furl = "https://services.arcgis.com/P3ePLMYs2RVChkJx/ArcGIS/rest/services/USA_Counties_Generalized_Boundaries/FeatureServer/0/query?where=1=1&outFields=*&f=json&resultRecordCount=10";
-    
-    // make the request
-    let resp = reqwest::get(furl)
-        .await.unwrap()
-        .text()
-        .await.unwrap();
-    
-    // parse the response into a FeatureSet
-    let fset: FeatureSet<2> = serde_json::from_str(&resp).unwrap();
-
-    // convert the FeatureSet to an Arrow RecordBatch
-    let rb = featureset_to_arrow(fset).unwrap();
-
-    println!("{:#?}", rb.schema());
+use std::io::Read;
+
+fn main() {
+    let flayer_url = "https://services.arcgis.com/P3ePLMYs2RVChkJx/ArcGIS/rest/services/USA_Counties_Generalized_Boundaries/FeatureServer/0/query?where=1%3D1&outFields=*&returnGeometry=true&resultRecordCount=5&f=json";
+    let mut res = reqwest::blocking::get(flayer_url).unwrap();
+    let mut body = String::new();
+
+    // Read the request into a String
+    res.read_to_string(&mut body).unwrap();
+
+    // Parse into a 2D FeatureSet
+    let fset: FeatureSet<2> = serde_json::from_str(&body).unwrap();
+
+    // Utilize the `geo` feature by converting to Polygon
+    // and calculate geodesic area
+    // iterate through the features
+    let area = fset
+        .features
+        .clone()
+        .into_iter()
+        .map(|feat| {
+            // convert to a geo_types::Polygon
+            let poly = Polygon::from(feat.geometry.unwrap().as_polygon().unwrap());
+            // calculate geodesic area
+            poly.geodesic_area_unsigned()
+        })
+        .collect::<Vec<_>>();
+
+    // print areas
+    println!("{:?}", area);
+
+    // convert to a geoarrow GeoTable
+    println!("{:?}", featureset_to_geoarrow(fset).unwrap());
 }
 ```
 
 ## Supported Esri JSON objects:
 
 ### Geometry 
 
-[Esri Geometries Objects](https://developers.arcgis.com/documentation/common-data-types/geometry-objects.htm#CURVE) are encoded by the following structs: 
+[Esri Geometries Objects](https://developers.arcgis.com/documentation/common-data-types/geometry-objects.htm) are encoded by the following structs: 
 
 - `EsriPoint`
 - `EsriMultiPoint<N>`
 - `EsriPolyline<N>`
 - `EsriPolygon<N>`
+- `EsriEnvelope`
 
 They are encapsulated by the `EsriGeometry` enum:
 
@@ -58,6 +83,7 @@ enum EsriGeometry<const N: usize> {
     MultiPoint(EsriMultiPoint<N>),
     Polygon(EsriPolygon<N>),
     Polyline(EsriPolyline<N>),
+    Envelope(EsriEnvelope),
 }
 ```
 The parameter `N` is used to specify the dimension of the geometries. Use `<2>` for 2 dimensional data, `<3>` for Z values and `<4>` when `M` and `Z` are present. 
@@ -99,24 +125,3 @@ struct SpatialReference {
     wkt: Option<String>,
 }
 ```
-
-## Example usage: 
-
-This example reads a single feature from a feature service and returns a `FeatureSet` struct. 
-
-```rust
-use serde_esri::features::FeatureSet;
-use reqwest::Error;
-use std::io::Read;
-
-fn main() -> Result<(), Error> {
-    let flayer_url = "https://services.arcgis.com/P3ePLMYs2RVChkJx/ArcGIS/rest/services/USA_Counties_Generalized_Boundaries/FeatureServer/0/query?where=1%3D1&outFields=*&returnGeometry=true&resultRecordCount=1&f=json";
-    let mut res = reqwest::blocking::get(flayer_url)?;
-    let mut body = String::new();
-    res.read_to_string(&mut body).unwrap();
-
-    let fset: FeatureSet<2> = serde_json::from_str(&body).unwrap(); 
-    println!("{:#?}", fset);
-    Ok(())
-}
-```

Cargo.toml

@@ -18,7 +18,7 @@ use crate::{
     geometry::EsriGeometry,
 };
 
-use std::sync::Arc;
+use std::{result::Result, sync::Arc};
 
 use geoarrow::GeometryArrayTrait;
 use serde_json::Value;
@@ -71,6 +71,20 @@ pub fn featureset_to_arrow<const N: usize>(
     }
 }
 
+use geoarrow::table::GeoTable;
+
+// TODO create my own error types
+/// Given a `FeatureSet`, create a geoarrow `GeoTable`
+pub fn featureset_to_geoarrow<const N: usize>(
+    x: FeatureSet<N>,
+) -> Result<GeoTable, geoarrow::error::GeoArrowError> {
+    let arrow_res = featureset_to_arrow(x)?;
+    let schema_ref = arrow_res.schema_ref().clone();
+    let geometry_index = arrow_res.schema().fields.len();
+
+    GeoTable::try_new(schema_ref, vec![arrow_res], geometry_index)
+}
+
 // convert an esri field to a new arrow field
 impl From<Field> for AField {
     fn from(value: Field) -> Self {

README.md

@@ -56,5 +56,3 @@ pub mod arrays {
         deserializer.deserialize_tuple(N, ArrayVisitor::<T, N>(PhantomData))
     }
 }
-
-// pub struct ArrayVisitor<T, const N: usize>(PhantomData<T>);

src/arrow_compat.rs

@@ -20,15 +20,15 @@ use serde_with::{serde_as, DisplayFromStr};
 /// Note that both geometry and attributes are optional. This is because
 /// we can anticipate receiving _only_ geometries, or _only_ attributes
 /// or both together.
-#[derive(Debug, Deserialize, Serialize)]
+#[derive(Debug, Deserialize, Serialize, Clone, Default)]
 pub struct Feature<const N: usize> {
     pub geometry: Option<EsriGeometry<N>>,
     pub attributes: Option<Map<String, Value>>,
 }
 
 /// A set of geometries and their attributes
 #[allow(non_snake_case)]
-#[derive(Debug, Deserialize, Serialize)]
+#[derive(Debug, Deserialize, Serialize, Clone, Default)]
 pub struct FeatureSet<const N: usize> {
     pub objectIdFieldName: Option<String>,
     pub globalIdFieldName: Option<String>,
@@ -44,7 +44,7 @@ pub struct FeatureSet<const N: usize> {
 // TODO sqlType, field_type need to be Enums
 #[serde_as]
 #[allow(non_snake_case)]
-#[derive(Clone, Debug, Deserialize, Serialize)]
+#[derive(Clone, Debug, Deserialize, Serialize, Default)]
 pub struct Field {
     pub name: String,
     #[serde(rename = "type")]

src/de_array.rs

@@ -1,12 +1,13 @@
 //! Enumeration of valid esri field types
-use serde::{Deserialize, Serialize};
 
-#[derive(Clone, Debug, Deserialize, Serialize)]
+use serde::{Deserialize, Serialize};
+#[derive(Clone, Debug, Deserialize, Serialize, Default)]
 pub enum FieldType {
     EsriFieldTypeSmallInteger = 0,
     EsriFieldTypeInteger = 1,
     EsriFieldTypeSingle = 2,
     EsriFieldTypeDouble = 3,
+    #[default]
     EsriFieldTypeString = 4,
     EsriFieldTypeDate = 5,
     EsriFieldTypeOid = 6,

src/features.rs

@@ -22,7 +22,7 @@ pub struct EsriCoord<const N: usize>(#[serde(with = "arrays")] pub [f64; N]);
 /// An `esriGeometryPoint` with fields x, y, z, and m. x and y are both required.
 #[skip_serializing_none]
 #[allow(non_snake_case)]
-#[derive(Clone, Deserialize, Serialize, Debug)]
+#[derive(Clone, Deserialize, Serialize, Debug, Default)]
 pub struct EsriPoint {
     pub x: f64,
     pub y: f64,
@@ -38,7 +38,7 @@ pub struct EsriPoint {
 /// a `panic!`.
 #[skip_serializing_none]
 #[allow(non_snake_case)]
-#[derive(Clone, Deserialize, Serialize, Debug)]
+#[derive(Clone, Deserialize, Serialize, Debug, Default)]
 pub struct EsriMultiPoint<const N: usize> {
     pub hasZ: Option<bool>,
     pub hasM: Option<bool>,
@@ -91,7 +91,7 @@ impl<'a, const N: usize> ExactSizeIterator for EsriMultiPointIterator<'a, N> {
 /// This struct is used strictly for representing the internal LineStrings
 /// for the `EsriPolygon` and `EsriPolyline` structs. They do not represent
 /// any Esri JSON geometry objects.
-#[derive(Clone, Deserialize, Serialize, Debug)]
+#[derive(Clone, Deserialize, Serialize, Debug, Default)]
 pub struct EsriLineString<const N: usize>(pub Vec<EsriCoord<N>>);
 
 pub struct EsriLineStringIterator<'a, const N: usize> {
@@ -137,7 +137,7 @@ impl<const N: usize> EsriLineString<N> {
 /// a `panic!`.
 #[skip_serializing_none]
 #[allow(non_snake_case)]
-#[derive(Clone, Deserialize, Serialize, Debug)]
+#[derive(Clone, Deserialize, Serialize, Debug, Default)]
 pub struct EsriPolyline<const N: usize> {
     pub hasZ: Option<bool>,
     pub hasM: Option<bool>,
@@ -174,7 +174,7 @@ impl<'a, const N: usize> ExactSizeIterator for EsriPolylineIterator<'a, N> {
 /// a `panic!`.
 #[skip_serializing_none]
 #[allow(non_snake_case)]
-#[derive(Clone, Deserialize, Serialize, Debug)]
+#[derive(Clone, Deserialize, Serialize, Debug, Default)]
 pub struct EsriPolygon<const N: usize> {
     pub hasZ: Option<bool>,
     pub hasM: Option<bool>,
@@ -249,7 +249,7 @@ impl<const N: usize> EsriGeometry<N> {
 // TODO: esriGeometryEnvelope.
 
 #[allow(non_snake_case)]
-#[derive(Clone, Deserialize, Serialize, Debug)]
+#[derive(Clone, Deserialize, Serialize, Debug, Default)]
 #[skip_serializing_none]
 pub struct EsriEnvelope {
     xmin: f64,

src/field_type.rs

@@ -1,4 +1,4 @@
-//!
+#![doc = include_str!("../README.md")]
 //!
 //! Example usage:
 //!

src/geometry.rs

@@ -4,7 +4,7 @@ use serde_with::skip_serializing_none;
 
 /// Read more on [Esri docs site](https://developers.arcgis.com/documentation/common-data-types/geometry-objects.htm#GUID-DFF0E738-5A42-40BC-A811-ACCB5814BABC)
 #[skip_serializing_none]
-#[derive(Serialize, Deserialize, Debug, Clone)]
+#[derive(Serialize, Deserialize, Debug, Clone, Default)]
 pub struct SpatialReference {
     pub wkid: Option<u32>,
     pub latest_wkid: Option<u32>,