Improve inspect error handling

Closes #303

Author: jeromekelleher

bio2zarr/cli.py

@@ -319,7 +319,7 @@ def dexplode_finalise(icf_path, verbose):
 
 
 @click.command
-@click.argument("path", type=click.Path())
+@click.argument("path", type=click.Path(exists=True))
 @verbose
 def inspect(path, verbose):
     """

bio2zarr/vcf2zarr/vcz.py

@@ -23,13 +23,15 @@
 
 def inspect(path):
     path = pathlib.Path(path)
-    # TODO add support for the Zarr format also
+    if not path.exists():
+        raise ValueError(f"Path not found: {path}")
     if (path / "metadata.json").exists():
         obj = icf.IntermediateColumnarFormat(path)
+    # NOTE: this is too strict, we should support more general Zarrs, see #276
     elif (path / ".zmetadata").exists():
         obj = VcfZarr(path)
     else:
-        raise ValueError("Format not recognised")  # NEEDS TEST
+        raise ValueError(f"{path} not in ICF or VCF Zarr format")
     return obj.summary_table()
 
 

tests/test_cli.py

@@ -807,6 +807,19 @@ def test_convert(self, tmp_path):
         assert "variant_position" in result.stdout
 
 
+class TestBadPaths:
+    @pytest.mark.parametrize("bad_path", ["PPP", "/dev/no_such_thing"])
+    def test_inspect(self, bad_path):
+        runner = ct.CliRunner(mix_stderr=False)
+        result = runner.invoke(
+            cli.vcf2zarr_main,
+            f"inspect {bad_path}",
+            catch_exceptions=False,
+        )
+        assert result.exit_code == 2
+        assert "Invalid value for" in result.stderr
+
+
 class TestVcfPartition:
     path = "tests/data/vcf/NA12878.prod.chr20snippet.g.vcf.gz"
     paths = (path, "tests/data/vcf/1kg_2020_chrM.vcf.gz")

tests/test_vcz.py

@@ -1,5 +1,7 @@
 import json
 
+import numpy.testing as nt
+import pandas as pd
 import pysam
 import pytest
 import sgkit as sg
@@ -662,3 +664,100 @@ def test_removed_samples(self, tmp_path, schema, icf_path, samples):
             json.dump(d, f)
         with pytest.raises(ValueError, match="Subsetting or reordering samples"):
             vcf2zarr.encode(icf_path, tmp_path / "z", schema_path=schema_path)
+
+
+class TestInspect:
+    def test_icf(self, icf_path):
+        df = pd.DataFrame(vcz_mod.inspect(icf_path))
+        assert list(df) == [
+            "name",
+            "type",
+            "chunks",
+            "size",
+            "compressed",
+            "max_n",
+            "min_val",
+            "max_val",
+        ]
+        nt.assert_array_equal(
+            df["name"].values,
+            [
+                "CHROM",
+                "POS",
+                "QUAL",
+                "ID",
+                "FILTERS",
+                "REF",
+                "ALT",
+                "rlen",
+                "INFO/NS",
+                "INFO/AN",
+                "INFO/AC",
+                "INFO/DP",
+                "INFO/AF",
+                "INFO/AA",
+                "INFO/DB",
+                "INFO/H2",
+                "FORMAT/GT",
+                "FORMAT/GQ",
+                "FORMAT/DP",
+                "FORMAT/HQ",
+            ],
+        )
+
+    def test_vcz(self, zarr_path):
+        df = pd.DataFrame(vcz_mod.inspect(zarr_path))
+        cols = [
+            "name",
+            "dtype",
+            "stored",
+            "size",
+            "ratio",
+            "nchunks",
+            "chunk_size",
+            "avg_chunk_stored",
+            "shape",
+            "chunk_shape",
+            "compressor",
+            "filters",
+        ]
+        assert list(df) == cols
+        fields = [
+            "/call_genotype",
+            "/call_HQ",
+            "/call_genotype_mask",
+            "/call_GQ",
+            "/call_DP",
+            "/call_genotype_phased",
+            "/variant_allele",
+            "/variant_AC",
+            "/variant_AF",
+            "/region_index",
+            "/variant_filter",
+            "/variant_id",
+            "/variant_contig",
+            "/variant_AA",
+            "/variant_quality",
+            "/variant_position",
+            "/variant_AN",
+            "/variant_length",
+            "/variant_NS",
+            "/variant_DB",
+            "/variant_DP",
+            "/variant_H2",
+            "/sample_id",
+            "/variant_id_mask",
+            "/filter_id",
+            "/contig_id",
+        ]
+        nt.assert_array_equal(df["name"], fields)
+
+    @pytest.mark.parametrize("bad_path", ["/NO_WAY", "TTTTTT"])
+    def test_no_such_path(self, bad_path):
+        with pytest.raises(ValueError, match=f"Path not found: {bad_path}"):
+            vcz_mod.inspect(bad_path)
+
+    @pytest.mark.parametrize("path", ["./", "tests/data/vcf/sample.vcf.gz"])
+    def test_unknown_format(self, path):
+        with pytest.raises(ValueError, match="not in ICF or VCF Zarr format"):
+            vcz_mod.inspect(path)