Fix tiff reader

ci/environment.yml

@@ -27,7 +27,8 @@ dependencies:
   - fsspec
   - s3fs
   - fastparquet
-  # for opening tiff files
+  # for opening and creating test tiff files
   - tifffile
+  - pillow
   # for opening FITS files
   - astropy

conftest.py

@@ -120,3 +120,16 @@ def simple_netcdf4(tmpdir):
     ds.to_netcdf(filepath)
 
     return filepath
+
+
+@pytest.fixture
+def random_tiff(tmpdir):
+    from PIL import Image
+
+    array = np.random.randint(0, 255, (128, 128), dtype=np.uint8)
+    img = Image.fromarray(array)
+
+    filepath = tmpdir / "rand.tiff"
+    img.save(filepath)
+
+    return str(filepath)

pyproject.toml

@@ -48,6 +48,8 @@ test = [
     "ruff",
     "s3fs",
     "scipy",
+    "tifffile",
+    "pillow",
 ]
 
 

virtualizarr/readers/tiff.py

@@ -9,7 +9,6 @@
     open_loadable_vars_and_indexes,
 )
 from virtualizarr.translators.kerchunk import (
-    extract_group,
     virtual_vars_and_metadata_from_kerchunk_refs,
 )
 from virtualizarr.types.kerchunk import KerchunkStoreRefs
@@ -45,7 +44,9 @@ def open_virtual_dataset(
         # handle inconsistency in kerchunk, see GH issue https://github.com/zarr-developers/VirtualiZarr/issues/160
         refs = KerchunkStoreRefs({"refs": tiff_to_zarr(filepath, **reader_options)})
 
-        refs = extract_group(refs, group)
+        print(refs)
+
+        # refs = extract_group(refs, group)
 
         virtual_vars, attrs, coord_names = virtual_vars_and_metadata_from_kerchunk_refs(
             refs,

virtualizarr/tests/__init__.py

@@ -37,6 +37,7 @@ def _importorskip(
 has_s3fs, requires_s3fs = _importorskip("s3fs")
 has_scipy, requires_scipy = _importorskip("scipy")
 has_tifffile, requires_tifffile = _importorskip("tifffile")
+has_pillow, requires_pillow = _importorskip("PIL")
 
 
 def create_manifestarray(

virtualizarr/tests/test_readers/test_tiff.py

@@ -0,0 +1,26 @@
+import numpy as np
+from xarray import Dataset
+
+from virtualizarr import open_virtual_dataset
+from virtualizarr.manifests import ManifestArray
+from virtualizarr.tests import requires_pillow
+
+
+@requires_pillow
+def test_random_tiff(random_tiff):
+    vds = open_virtual_dataset(random_tiff, indexes={})
+
+    assert isinstance(vds, Dataset)
+
+    # TODO what is the name of this array expected to be??
+    assert list(vds.variables) == ["foo"]
+    vda = vds["foo"]
+
+    assert vda.sizes == {"X": 128, "Y": 128}
+    assert vda.dtype == np.uint8
+
+    assert isinstance(vda.data, ManifestArray)
+    manifest = vda.data.manifest
+    assert manifest.dict() == {
+        "0.0": {"path": random_tiff, "offset": 122, "length": 16384}
+    }