Change zip to fail if iterables are different lengths

Author: tomwhite

bio2zarr/plink.py

@@ -185,11 +185,11 @@ def validate(bed_path, zarr_path):
     assert call_genotype.shape[2] == 2
 
     row_id = 0
-    for bed_row, zarr_row in zip(bed_genotypes, call_genotype, strict=False):
+    for bed_row, zarr_row in zip(bed_genotypes, call_genotype, strict=True):
         # print("ROW", row_id)
         # print(bed_row, zarr_row)
         row_id += 1
-        for bed_call, zarr_call in zip(bed_row, zarr_row, strict=False):
+        for bed_call, zarr_call in zip(bed_row, zarr_row, strict=True):
             if bed_call == -127:
                 assert list(zarr_call) == [-1, -1]
             elif bed_call == 0:

bio2zarr/vcf2zarr/icf.py

@@ -289,7 +289,7 @@ def scan_vcf(path, target_num_partitions, *, local_alleles):
             samples=[Sample(sample_id) for sample_id in vcf.samples],
             contigs=[
                 Contig(contig_id, length)
-                for contig_id, length in zip(vcf.seqnames, contig_lengths, strict=False)
+                for contig_id, length in zip(vcf.seqnames, contig_lengths, strict=True)
             ],
             filters=filters,
             fields=fields,
@@ -766,7 +766,7 @@ def chunks(self, partition_id, start_chunk=0):
         for count, cumulative in zip(
             chunk_num_records[start_chunk:],
             chunk_cumulative_records[start_chunk + 1 :],
-            strict=False,
+            strict=True,
         ):
             path = partition_path / f"{cumulative}"
             chunk = self.read_chunk(path)

bio2zarr/vcf2zarr/vcz.py

@@ -760,7 +760,7 @@ def encode_alleles_partition(self, partition_index):
         for ref, alt in zip(
             ref_field.iter_values(partition.start, partition.stop),
             alt_field.iter_values(partition.start, partition.stop),
-            strict=False,
+            strict=True,
         ):
             j = alleles.next_buffer_row()
             alleles.buff[j, :] = constants.STR_FILL

bio2zarr/vcf2zarr/verification.py

@@ -114,7 +114,7 @@ def assert_format_val_equal(vcf_val, zarr_val, vcf_type, vcf_number):
     assert isinstance(vcf_val, np.ndarray)
     if vcf_type in ("String", "Character"):
         assert len(vcf_val) == len(zarr_val)
-        for v, z in zip(vcf_val, zarr_val, strict=False):
+        for v, z in zip(vcf_val, zarr_val, strict=True):
             if vcf_number == "1":
                 assert v == z
             else:

bio2zarr/vcf_utils.py

@@ -450,7 +450,7 @@ def __exit__(self, exc_type, exc_val, exc_tb):
         return False
 
     def contig_record_counts(self):
-        d = dict(zip(self.sequence_names, self.index.record_counts, strict=False))
+        d = dict(zip(self.sequence_names, self.index.record_counts, strict=True))
         if self.file_type == VcfFileType.BCF:
             d = {k: v for k, v in d.items() if v > 0}
         return d