fix: Skip unknown GeoTag keys (#134)

* Add failing test for unknown tag

* Lint

* Add support for tag

* Add bindings

* Fix typo

* Add docstring

* Add tiny test

* Ignore unknown GeoKeyTag ids

* infer type in tests

* fix python compile

* fmt

* restore whitespaces

* Ensure geo and ome tiff tests are run

* fmt

---------

Co-authored-by: Max Jones <[email protected]>

Author: kylebarron

README.md

@@ -24,4 +24,5 @@ Download the following file for use in the tests.
 
 ```shell
 aws s3 cp s3://naip-visualization/ny/2022/60cm/rgb/40073/m_4007307_sw_18_060_20220803.tif ./ --request-payer
+aws s3 cp s3://prd-tnm/StagedProducts/Elevation/13/TIFF/current/s14w171/USGS_13_s14w171.tif ./ --no-sign-request --region us-west-2
 ```

python/DEVELOP.md

@@ -2,4 +2,5 @@
 uv sync --no-install-package async-tiff
 uv run --no-project maturin develop
 uv run --no-project mkdocs serve
+uv run --no-project pytest --verbose
 ```

src/ifd.rs

@@ -295,8 +295,15 @@ impl ImageFileDirectory {
                     .expect("There should be a chunk for each key.");
 
                 let key_id = chunk[0];
-                let tag_name =
-                    GeoKeyTag::try_from_primitive(key_id).expect("Unknown GeoKeyTag id: {key_id}");
+                let tag_name = if let Ok(tag_name) = GeoKeyTag::try_from_primitive(key_id) {
+                    tag_name
+                } else {
+                    // Skip unknown GeoKeyTag ids. Some GeoTIFFs include keys that were proposed
+                    // but not included in the GeoTIFF spec. See
+                    // https://github.com/developmentseed/async-tiff/pull/131 and
+                    // https://github.com/virtual-zarr/virtual-tiff/issues/52
+                    continue;
+                };
 
                 let tag_location = chunk[1];
                 let count = chunk[2];

tests/geo.rs

@@ -0,0 +1,25 @@
+use std::env;
+use std::sync::Arc;
+
+use async_tiff::metadata::{PrefetchBuffer, TiffMetadataReader};
+use async_tiff::reader::{AsyncFileReader, ObjectReader};
+use object_store::local::LocalFileSystem;
+
+#[tokio::test]
+async fn test_parse_file_with_unknown_geokey() {
+    let folder = env::current_dir().unwrap();
+    let path = object_store::path::Path::parse("tests/images/geogtowgs_subset_USGS_13_s14w171.tif")
+        .unwrap();
+    let store = Arc::new(LocalFileSystem::new_with_prefix(folder).unwrap());
+    let reader = Arc::new(ObjectReader::new(store, path)) as Arc<dyn AsyncFileReader>;
+    let prefetch_reader = PrefetchBuffer::new(reader.clone(), 32 * 1024)
+        .await
+        .unwrap();
+    let mut metadata_reader = TiffMetadataReader::try_open(&prefetch_reader)
+        .await
+        .unwrap();
+    let _ = metadata_reader
+        .read_all_ifds(&prefetch_reader)
+        .await
+        .unwrap();
+}

tests/images/geogtowgs_subset_USGS_13_s14w171.tif

@@ -0,0 +1,35 @@
+`geogtowgs_subset_USGS_13_s14w171.tif` was created from "s3://prd-tnm/StagedProducts/Elevation/13/TIFF/current/s14w171/USGS_13_s14w171.tif" using these commands:
+
+```bash
+gdal_translate USGS_13_s14w171.tif tiny.tif -srcwin 0 0 1 1 -co COMPRESS=DEFLATE
+listgeo USGS_13_s14w171.tif > metadata.txt  # Then modify to remove information related to spatial extent
+cp tiny.tif geogtowgs_subset_USGS_13_s14w171.tif
+geotifcp -g metadata.txt tiny.tif geogtowgs_subset_USGS_13_s14w171.tif
+listgeo geogtowgs_subset_USGS_13_s14w171.tif
+```
+
+and this workspace definition:
+
+```toml
+[project]
+name = "gdal-workspace"
+version = "0.1.0"
+description = "workspace for using gdal via pixi"
+readme = "README.md"
+requires-python = ">=3.12"
+dependencies = []
+
+[tool.pixi.workspace]
+channels = ["conda-forge"]
+platforms = ["osx-arm64"]
+
+[tool.pixi.pypi-dependencies]
+gdal-workspace = { path = ".", editable = true }
+
+[tool.pixi.tasks]
+
+[tool.pixi.dependencies]
+gdal = ">=3.11.5,<4"
+libgdal = ">=3.11.5,<4"
+geotiff = ">=1.7.4,<2"
+```

tests/images/readme.md

@@ -1 +1,3 @@
+mod geo;
 mod image_tiff;
+mod ome_tiff;

tests/mod.rs

@@ -1,12 +1,35 @@
-/// Integration tests on OME-TIFF files.
+//! Integration tests on OME-TIFF files.
+
+use std::sync::Arc;
+
+use async_tiff::metadata::{PrefetchBuffer, TiffMetadataReader};
+use async_tiff::reader::{AsyncFileReader, ObjectReader};
 use async_tiff::tiff::tags::PhotometricInterpretation;
+use async_tiff::TIFF;
+use reqwest::Url;
 
-mod util;
+async fn open_remote_tiff(url: &str) -> TIFF {
+    let parsed_url = Url::parse(url).expect("failed parsing url");
+    let (store, path) = object_store::parse_url(&parsed_url).unwrap();
+
+    let reader = Arc::new(ObjectReader::new(Arc::new(store), path)) as Arc<dyn AsyncFileReader>;
+    let prefetch_reader = PrefetchBuffer::new(reader.clone(), 32 * 1024)
+        .await
+        .unwrap();
+    let mut metadata_reader = TiffMetadataReader::try_open(&prefetch_reader)
+        .await
+        .unwrap();
+    let ifds = metadata_reader
+        .read_all_ifds(&prefetch_reader)
+        .await
+        .unwrap();
+    TIFF::new(ifds)
+}
 
 #[tokio::test]
 async fn test_ome_tiff_single_channel() {
     let tiff =
-        util::open_remote_tiff("https://downloads.openmicroscopy.org/images/OME-TIFF/2016-06/bioformats-artificial/single-channel.ome.tif").await;
+        open_remote_tiff("https://downloads.openmicroscopy.org/images/OME-TIFF/2016-06/bioformats-artificial/single-channel.ome.tif").await;
 
     assert_eq!(tiff.ifds().len(), 1);
     let ifd = &tiff.ifds()[0];