use serde::{Deserialize, Serialize};
use std::collections::HashMap;
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct GeoParquetMetadata {
pub version: String,
pub primary_column: String,
pub columns: HashMap<String, GeoParquetColumnMeta>,
}
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct GeoParquetColumnMeta {
pub encoding: String,
pub geometry_types: Vec<String>,
pub crs: serde_json::Value,
#[serde(skip_serializing_if = "Option::is_none")]
pub bbox: Option<[f64; 4]>,
}
impl GeoParquetMetadata {
pub fn single_column(
column_name: &str,
geometry_types: Vec<String>,
bbox: Option<[f64; 4]>,
) -> Self {
let mut columns = HashMap::new();
columns.insert(
column_name.to_string(),
GeoParquetColumnMeta {
encoding: "WKB".to_string(),
geometry_types,
crs: serde_json::json!({
"type": "GeographicCRS",
"name": "WGS 84",
"id": { "authority": "EPSG", "code": 4326 }
}),
bbox,
},
);
Self {
version: "1.1.0".to_string(),
primary_column: column_name.to_string(),
columns,
}
}
pub fn to_json(&self) -> Result<String, sonic_rs::Error> {
sonic_rs::to_string(self)
}
pub const PARQUET_KEY: &'static str = "geo";
}
pub const GEOARROW_EXTENSION_NAME: &str = "geoarrow.wkb";
pub fn geoarrow_extension_metadata(crs_epsg: u32) -> String {
serde_json::json!({
"crs": {
"type": "GeographicCRS",
"name": "WGS 84",
"id": { "authority": "EPSG", "code": crs_epsg }
}
})
.to_string()
}
pub const ARROW_EXTENSION_NAME_KEY: &str = "ARROW:extension:name";
pub const ARROW_EXTENSION_METADATA_KEY: &str = "ARROW:extension:metadata";
pub fn geoarrow_field_metadata() -> HashMap<String, String> {
let mut meta = HashMap::new();
meta.insert(
ARROW_EXTENSION_NAME_KEY.to_string(),
GEOARROW_EXTENSION_NAME.to_string(),
);
meta.insert(
ARROW_EXTENSION_METADATA_KEY.to_string(),
geoarrow_extension_metadata(4326),
);
meta
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn geoparquet_metadata_json() {
let meta = GeoParquetMetadata::single_column(
"geom",
vec!["Point".to_string(), "Polygon".to_string()],
Some([-180.0, -90.0, 180.0, 90.0]),
);
let json = meta.to_json().unwrap();
assert!(json.contains("\"version\":\"1.1.0\""));
assert!(json.contains("\"primary_column\":\"geom\""));
assert!(json.contains("\"encoding\":\"WKB\""));
assert!(json.contains("EPSG"));
}
#[test]
fn geoparquet_key() {
assert_eq!(GeoParquetMetadata::PARQUET_KEY, "geo");
}
#[test]
fn geoarrow_field_meta() {
let meta = geoarrow_field_metadata();
assert_eq!(meta[ARROW_EXTENSION_NAME_KEY], "geoarrow.wkb");
assert!(meta[ARROW_EXTENSION_METADATA_KEY].contains("EPSG"));
}
#[test]
fn geoarrow_extension_name() {
assert_eq!(GEOARROW_EXTENSION_NAME, "geoarrow.wkb");
}
#[test]
fn roundtrip_parquet_metadata() {
let meta = GeoParquetMetadata::single_column("location", vec!["Point".into()], None);
let json = meta.to_json().unwrap();
let parsed: GeoParquetMetadata = sonic_rs::from_str(&json).unwrap();
assert_eq!(parsed.primary_column, "location");
assert_eq!(parsed.columns["location"].encoding, "WKB");
}
}