Merge pull request #83 from sudlab/ns-rse/81-io-bam-vcf

isoslam/io.py

@@ -1,11 +1,12 @@
 """Module for reading files."""
 
 import argparse
+import gzip
 from collections.abc import Callable
 from datetime import datetime
 from importlib import resources
 from pathlib import Path
-from typing import Any
+from typing import Any, TextIO
 
 import pysam
 
@@ -201,15 +202,13 @@ def _get_loader(file_ext: str = "bam") -> Callable:  # type: ignore[type-arg]
     """
     if file_ext == ".bam":
         return _load_bam
-    if file_ext == ".bed":
+    if file_ext == ".bed" or file_ext == ".bed.gz":
         return _load_bed
     if file_ext == ".gtf":
         return _load_gtf
     if file_ext == ".tbi":
         return _load_tbi
-    if file_ext == ".vcf":
-        return _load_vcf
-    if file_ext == ".vcf.gz":
+    if file_ext == ".vcf" or file_ext == ".vcf.gz":
         return _load_vcf
     raise ValueError(file_ext)
 
@@ -234,9 +233,26 @@ def _load_bam(bam_file: str | Path) -> pysam.libcalignmentfile.AlignmentFile:
         raise e
 
 
-def _load_bed() -> None:
-    """Load '.bed' file."""
-    return
+def _load_bed(bed_file: str | Path) -> TextIO:
+    """
+    Open '.bed' file for reading, supports gzip compressed formats.
+
+    Parameters
+    ----------
+    bed_file : str | Path
+        Path, as string or pathlib Path, to a '.bed' or '.bed.gz' file that is to be loaded.
+
+    Returns
+    -------
+    TextIO
+        Returns a connection to an open file object.
+    """
+    try:
+        if Path(bed_file).suffix == ".gz":
+            return gzip.open(bed_file, "rt", encoding="utf-8")
+        return Path(bed_file).open(mode="r", encoding="utf-8")
+    except OSError as e:
+        raise e
 
 
 def _load_gtf(gtf_file: str | Path) -> pysam.libctabix.tabix_generic_iterator:

tests/test_io.py

@@ -3,8 +3,9 @@
 import argparse
 from collections.abc import Callable
 from datetime import datetime
+from io import TextIOWrapper
 from pathlib import Path
-from typing import Any
+from typing import Any, TextIO
 
 import pysam
 import pytest
@@ -44,7 +45,7 @@ def test_str_to_path(tmp_path: Path) -> None:
 
 
 def test_path_to_str(tmp_path: Path) -> None:
-    """Test that Path objects are converted to strings."""
+    """Test that Path objects in dictionaries are converted to strings."""
     CONFIG_PATH = {
         "this": "is",
         "a": "test",
@@ -169,17 +170,16 @@ def test_load_bam(
     assert bam_file.compression == compression
 
 
-@pytest.mark.skip
 @pytest.mark.parametrize(
     ("file_path", "object_type"),
     [
-        pytest.param(RESOURCES / "bed" / "test_coding_introns.bed", str, id="bed file as Path"),
-        pytest.param("tests/resources/bed/test_coding_introns.bed", str, id="bed file as str"),
+        pytest.param(RESOURCES / "bed" / "test_coding_introns.bed", TextIOWrapper, id="bed file as Path"),
+        pytest.param("tests/resources/bed/test_coding_introns.bed", TextIOWrapper, id="bed file as str"),
     ],
 )
-def test_load_bed(file_path: str | Path, object_type: str) -> None:
+def test_load_bed(file_path: str | Path, object_type: TextIO) -> None:
     """Test loading of bed file."""
-    bed_file = io._load_bam(file_path)
+    bed_file = io._load_bed(file_path)
     assert isinstance(bed_file, object_type)