Added a warning in sort.py when a column is not in pairsam_format.DTYPES_PAIRSAM, informing users it’s treated as a string, which may affect sorting if numeric.

Modified test_sort.py to create a temporary .pairsam file with custom_col in the header for test_custom_column_warning, ensuring the data type warning is triggered and tested.

Author: ShigrafS

pairtools/cli/sort.py

@@ -68,9 +68,9 @@
     nargs=1,
     type=str,
     multiple=True,
-    help="Extra column (name or numerical index) that is also used for sorting."
-    "The option can be provided multiple times."
-    'Example: --extra-col "phase1" --extra-col "phase2". [output format option]',
+    help="Extra column (name or numerical index) to sort by. "
+         "If not defined in pairsam format, treated as a string and a warning is issued. "
+         "Can be provided multiple times, e.g., --extra-col phase1 --extra-col phase2.",
 )
 @click.option(
     "--nproc",
@@ -230,8 +230,14 @@ def sort_py(
     for col in sort_columns:
         if col is None:
             continue  # Skip optional columns
-        dtype = pairsam_format.DTYPES_PAIRSAM.get(column_names[col], str)
-        cols.append(f"-k {col+1},{col+1}{'n' if issubclass(dtype, int) else ''}")
+        col_name = column_names[col]
+        dtype = pairsam_format.DTYPES_PAIRSAM.get(col_name, str)
+        if col_name not in pairsam_format.DTYPES_PAIRSAM:
+            logger.warning(
+                f"Column '{col_name}' not found in pairsam format definitions. "
+                "Assuming string type for sorting, which may affect sort order if numeric."
+            )
+        cols.append(f"-k {col + 1},{col + 1}{'n' if issubclass(dtype, int) else ''}")
     cols = " ".join(cols)
 
     command = rf"""

tests/test_sort.py

@@ -2,7 +2,7 @@
 import os
 import subprocess
 import sys
-# import pytest
+import pytest
 
 testdir = os.path.dirname(os.path.realpath(__file__))
 
@@ -61,3 +61,60 @@ def test_mock_pairsam():
                         assert int(cur_pair[3]) >= int(prev_pair[3])
 
         prev_pair = cur_pair
+
+
+def test_custom_column_warning(tmpdir):
+    """Test that a warning is emitted when sorting with a custom column not in pairsam format."""
+    # Create a temporary .pairsam file with custom_col in the header
+    mock_pairsam_path = os.path.join(tmpdir, "test.pairsam")
+    output_path = os.path.join(tmpdir, "sorted_output.pairsam")
+
+    # Write a minimal .pairsam file
+    with open(mock_pairsam_path, "w") as f:
+        f.write("## pairs format v1.0\n")
+        f.write("#columns: readID chr1 pos1 chr2 pos2 strand1 strand2 pair_type custom_col\n")
+        f.write("read1\tchr1\t100\tchr2\t200\t+\t-\tUU\t42\n")
+        f.write("read2\tchr2\t150\tchr1\t250\t-\t+\tUU\t99\n")
+
+    # Run sort command with a custom column
+    cmd = [
+        "python",
+        "-m",
+        "pairtools",
+        "sort",
+        mock_pairsam_path,
+        "--output",
+        output_path,
+        "--extra-col",
+        "custom_col",
+    ]
+
+    # Capture stderr to check for warning
+    process = subprocess.Popen(
+        cmd,
+        stderr=subprocess.PIPE,
+        stdout=subprocess.PIPE,
+        universal_newlines=True,
+    )
+    stdout, stderr = process.communicate()
+
+    # Check that the command completed successfully
+    assert process.returncode == 0, f"Command failed: {stderr}"
+
+    # Verify warning was emitted
+    assert "Column 'custom_col' not found in pairsam format definitions" in stderr
+    assert "Assuming string type for sorting" in stderr
+
+    # Verify output file exists and has content
+    assert os.path.exists(output_path)
+    with open(output_path, "r") as f:
+        output_lines = f.readlines()
+    assert len(output_lines) > 0
+
+    # Check that the output is sorted (basic check on header and body)
+    output_header = [line.strip() for line in output_lines if line.startswith("#")]
+    output_body = [
+        line.strip() for line in output_lines if not line.startswith("#") and line.strip()
+    ]
+    assert any(line.startswith("#sorted") for line in output_header)
+    assert len(output_body) > 0