feat: validate in cli (#613)

gadomski · web-flow · commit ca0ea4eb5a2a · 2025-01-31T16:29:23.000Z
## Closes - Closes #612
diff --git a/README.md b/README.md
@@ -52,6 +52,9 @@ $ stacrs search items.parquet
 
 # Server
 $ stacrs serve items.parquet  # Opens a STAC API server on http://localhost:7822
+
+# Validate
+$ stacrs validate item.json
 ```
 
 ## Python
diff --git a/crates/cli/README.md b/crates/cli/README.md
@@ -43,6 +43,9 @@ $ stac search items.parquet
 
 # Server
 $ stacrs serve items.parquet  # Opens a STAC API server on http://localhost:7822
+
+# Validate
+$ stacrs validate item.json
 ```
 
 ## Usage
@@ -52,6 +55,7 @@ $ stacrs serve items.parquet  # Opens a STAC API server on http://localhost:7822
 - `stacrs search`: searches STAC APIs and geoparquet files
 - `stacrs serve`: serves a STAC API
 - `stacrs translate`: converts STAC from one format to another
+- `stacrs validate`: validates a STAC value
 
 Use the `--help` flag to see all available options for the CLI and the subcommands:
 
diff --git a/crates/cli/data/invalid-item.json b/crates/cli/data/invalid-item.json
@@ -0,0 +1,81 @@
+{
+  "stac_version": "1.1.0",
+  "stac_extensions": [],
+  "type": "Feature",
+  "id": "",
+  "bbox": [
+    172.91173669923782,
+    1.3438851951615003,
+    172.95469614953714,
+    1.3690476620161975
+  ],
+  "geometry": {
+    "type": "Polygon",
+    "coordinates": [
+      [
+        [
+          172.91173669923782,
+          1.3438851951615003
+        ],
+        [
+          172.95469614953714,
+          1.3438851951615003
+        ],
+        [
+          172.95469614953714,
+          1.3690476620161975
+        ],
+        [
+          172.91173669923782,
+          1.3690476620161975
+        ],
+        [
+          172.91173669923782,
+          1.3438851951615003
+        ]
+      ]
+    ]
+  },
+  "properties": {
+    "datetime": "2020-12-11T22:38:32.125000Z"
+  },
+  "collection": "simple-collection",
+  "links": [
+    {
+      "rel": "collection",
+      "href": "./collection.json",
+      "type": "application/json",
+      "title": "Simple Example Collection"
+    },
+    {
+      "rel": "root",
+      "href": "./collection.json",
+      "type": "application/json",
+      "title": "Simple Example Collection"
+    },
+    {
+      "rel": "parent",
+      "href": "./collection.json",
+      "type": "application/json",
+      "title": "Simple Example Collection"
+    }
+  ],
+  "assets": {
+    "visual": {
+      "href": "https://storage.googleapis.com/open-cogs/stac-examples/20201211_223832_CS2.tif",
+      "type": "image/tiff; application=geotiff; profile=cloud-optimized",
+      "title": "3-Band Visual",
+      "roles": [
+        "visual"
+      ]
+    },
+    "thumbnail": {
+      "href": "https://storage.googleapis.com/open-cogs/stac-examples/20201211_223832_CS2.jpg",
+      "title": "Thumbnail",
+      "type": "image/jpeg",
+      "roles": [
+        "thumbnail"
+      ]
+    }
+  }
+}
diff --git a/crates/cli/src/lib.rs b/crates/cli/src/lib.rs
@@ -1,10 +1,10 @@
 use anyhow::{anyhow, Error, Result};
 use clap::{Parser, Subcommand};
-use stac::{geoparquet::Compression, Collection, Format, Item, Links, Migrate};
+use stac::{geoparquet::Compression, Collection, Format, Item, Links, Migrate, Validate};
 use stac_api::{GetItems, GetSearch, Search};
 use stac_server::Backend;
 use std::{collections::HashMap, io::Write, str::FromStr};
-use tokio::{io::AsyncReadExt, net::TcpListener};
+use tokio::{io::AsyncReadExt, net::TcpListener, runtime::Handle};
 
 /// stacrs: A command-line interface for the SpatioTemporal Asset Catalog (STAC)
 #[derive(Debug, Parser)]
@@ -193,6 +193,17 @@ pub enum Command {
         #[arg(long, default_value_t = true)]
         create_collections: bool,
     },
+
+    /// Validates a STAC value.
+    ///
+    /// The default output format is plain text — use `--output-format=json` to
+    /// get structured output.
+    Validate {
+        /// The input file.
+        ///
+        /// To read from standard input, pass `-` or don't provide an argument at all.
+        infile: Option<String>,
+    },
 }
 
 #[derive(Debug)]
@@ -336,6 +347,40 @@ impl Stacrs {
                     load_and_serve(addr, backend, collections, items, create_collections).await
                 }
             }
+            Command::Validate { ref infile } => {
+                let value = self.get(infile.as_deref()).await?;
+                let result = Handle::current()
+                    .spawn_blocking(move || value.validate())
+                    .await?;
+                if let Err(error) = result {
+                    if let stac::Error::Validation(errors) = error {
+                        if let Some(format) = self.output_format {
+                            if let Format::Json(_) = format {
+                                let value = errors
+                                    .into_iter()
+                                    .map(|error| error.into_json())
+                                    .collect::<Vec<_>>();
+                                if self.compact_json.unwrap_or_default() {
+                                    serde_json::to_writer(std::io::stdout(), &value)?;
+                                } else {
+                                    serde_json::to_writer_pretty(std::io::stdout(), &value)?;
+                                }
+                                println!("");
+                            } else {
+                                return Err(anyhow!("invalid output format: {}", format));
+                            }
+                        } else {
+                            for error in errors {
+                                println!("{}", error);
+                            }
+                        }
+                    }
+                    std::io::stdout().flush()?;
+                    Err(anyhow!("one or more validation errors"))
+                } else {
+                    Ok(())
+                }
+            }
         }
     }
 
@@ -591,4 +636,18 @@ mod tests {
             Format::Geoparquet(Some(Compression::LZO))
         );
     }
+
+    #[rstest]
+    fn validate(mut command: Command) {
+        command
+            .arg("validate")
+            .arg("examples/simple-item.json")
+            .assert()
+            .success();
+        command
+            .arg("validate")
+            .arg("data/invalid-item.json")
+            .assert()
+            .failure();
+    }
 }
diff --git a/crates/core/CHANGELOG.md b/crates/core/CHANGELOG.md
@@ -6,6 +6,10 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 
 ## [Unreleased]
 
+### Added
+
+- `error::Validation::into_json` ([#613](https://github.com/stac-utils/stac-rs/pull/613))
+
 ### Removed
 
 - Async validation ([#611](https://github.com/stac-utils/stac-rs/pull/611))
diff --git a/crates/core/src/error.rs b/crates/core/src/error.rs
@@ -174,15 +174,6 @@ impl Validation {
         error: jsonschema::ValidationError<'_>,
         value: Option<&serde_json::Value>,
     ) -> Validation {
-        use std::borrow::Cow;
-
-        // Cribbed from https://docs.rs/jsonschema/latest/src/jsonschema/error.rs.html#21-30
-        let error = jsonschema::ValidationError {
-            instance_path: error.instance_path.clone(),
-            instance: Cow::Owned(error.instance.into_owned()),
-            kind: error.kind,
-            schema_path: error.schema_path,
-        };
         let mut id = None;
         let mut r#type = None;
         if let Some(value) = value.and_then(|v| v.as_object()) {
@@ -192,7 +183,21 @@ impl Validation {
                 .and_then(|v| v.as_str())
                 .and_then(|s| s.parse::<crate::Type>().ok());
         }
-        Validation { id, r#type, error }
+        Validation {
+            id,
+            r#type,
+            error: error.to_owned(),
+        }
+    }
+
+    /// Converts this validation error into a [serde_json::Value].
+    pub fn into_json(self) -> serde_json::Value {
+        let error_description = jsonschema::output::ErrorDescription::from(self.error);
+        serde_json::json!({
+            "id": self.id,
+            "type": self.r#type,
+            "error": error_description,
+        })
     }
 }
 
diff --git a/crates/core/src/lib.rs b/crates/core/src/lib.rs
@@ -221,7 +221,7 @@ pub const STAC_VERSION: Version = Version::v1_1_0;
 pub type Result<T> = std::result::Result<T, Error>;
 
 /// Enum for the four "types" of STAC values.
-#[derive(Clone, Copy, Debug, PartialEq, Eq, Hash)]
+#[derive(Clone, Copy, Debug, PartialEq, Eq, Hash, serde::Serialize)]
 pub enum Type {
     /// An item.
     Item,
