fix(cli): make geoparquet a feature

Author: gadomski

.github/workflows/ci.yml

@@ -29,6 +29,7 @@ jobs:
           - "-p stac -p stac-api -F geo"
           - "-p stac-async"
           - "-p stac-cli --no-default-features"
+          - "-p stac-cli --no-default-features -F geoparquet"
           - "-p stac-duckdb -F duckdb/bundled"
           - "-p stac-server"
           - "-p stac-server -F axum"

cli/Cargo.toml

@@ -11,21 +11,20 @@ keywords = ["geospatial", "stac", "metadata", "geo", "raster"]
 categories = ["science", "data-structures"]
 
 [features]
-default = ["gdal", "pgstac"]
+default = ["gdal", "geoparquet", "pgstac"]
 duckdb = ["dep:stac-duckdb"]
 gdal = ["stac/gdal"]
+geoparquet = ["dep:bytes", "stac/geoparquet-compression"]
 pgstac = ["stac-server/pgstac"]
 
 [dependencies]
 axum = "0.7"
-bytes = "1"
+bytes = { version = "1", optional = true }
 clap = { version = "4", features = ["derive"] }
 reqwest = "0.12"
 serde = "1"
 serde_json = "1"
-stac = { version = "0.8", path = "../core", features = [
-    "geoparquet-compression",
-] }
+stac = { version = "0.8", path = "../core" }
 stac-api = { version = "0.4", path = "../api" }
 stac-async = { version = "0.5", path = "../async" }
 stac-duckdb = { version = "0.1", path = "../duckdb", optional = true }

cli/src/format.rs

@@ -1,12 +1,11 @@
 use crate::{Error, Result};
-use bytes::Bytes;
 use clap::ValueEnum;
 use serde::de::DeserializeOwned;
-use std::{fs::File, io::Read};
 
 /// The STAC output format.
 #[derive(Clone, Copy, Debug, Default, ValueEnum)]
 pub enum Format {
+    #[cfg(feature = "geoparquet")]
     /// stac-geoparquet
     Parquet,
 
@@ -24,11 +23,12 @@ impl Format {
     pub(crate) async fn read_href<D: DeserializeOwned>(&self, href: Option<&str>) -> Result<D> {
         if let Some(href) = href.and_then(|href| if href == "-" { None } else { Some(href) }) {
             match *self {
+                #[cfg(feature = "geoparquet")]
                 Format::Parquet => {
                     let item_collection = if let Some(url) = stac::href_to_url(href) {
                         stac::geoparquet::from_reader(reqwest::blocking::get(url)?.bytes()?)?
                     } else {
-                        let file = File::open(href)?;
+                        let file = std::fs::File::open(href)?;
                         stac::geoparquet::from_reader(file)?
                     };
                     serde_json::from_value(serde_json::to_value(item_collection)?)
@@ -38,10 +38,12 @@ impl Format {
             }
         } else {
             match *self {
+                #[cfg(feature = "geoparquet")]
                 Format::Parquet => {
+                    use std::io::Read;
                     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::Bytes::from(buf))?;
                     serde_json::from_value(serde_json::to_value(item_collection)?)
                         .map_err(Error::from)
                 }

cli/src/runner.rs

@@ -46,6 +46,7 @@ where
                         writeln!(self.writer, "{}", value)?;
                     }
                 }
+                #[cfg(feature = "geoparquet")]
                 Format::Parquet => {
                     if let Some(value) = value.to_stac() {
                         stac::geoparquet::to_writer(&mut self.writer, value)?;

scripts/validate-stac-geoparquet

@@ -37,7 +37,7 @@ def clean_report(report: dict[str, Any]) -> dict[str, Any]:
 try:
     # Writing
     subprocess.check_call(
-        ["cargo", "run", "--no-default-features", "--", "translate", path, parquet_path]
+        ["cargo", "run", "--no-default-features", "-F", "geoparquet", "--", "translate", path, parquet_path]
     )
     table = pyarrow.parquet.read_table(parquet_path)
     after = next(stac_geoparquet.arrow.stac_table_to_items(table))
@@ -55,7 +55,7 @@ try:
     table = stac_geoparquet.arrow.parse_stac_items_to_arrow([before])
     stac_geoparquet.arrow.to_parquet(table, parquet_path)
     item_collection = json.loads(subprocess.check_output(
-        ["cargo", "run", "--no-default-features", "--", "translate", parquet_path]
+        ["cargo", "run", "--no-default-features", "-F", "geoparquet", "--", "translate", parquet_path]
     ))
     assert len(item_collection["features"]) == 1
     clean_item(item_collection["features"][0])  # stac-geoparquet writes as a multi-polygon