Merge pull request #88 from compomics/feature/tsv-raise-exception

Improve TSV reading error handling

psm_utils/io/tsv.py

@@ -57,10 +57,10 @@
 from pydantic import ValidationError
 
 from psm_utils.io._base_classes import ReaderBase, WriterBase
+from psm_utils.io._utils import set_csv_field_size_limit
 from psm_utils.io.exceptions import PSMUtilsIOException
 from psm_utils.psm import PSM
 from psm_utils.psm_list import PSMList
-from psm_utils.io._utils import set_csv_field_size_limit
 
 set_csv_field_size_limit()
 
@@ -74,12 +74,21 @@ def __iter__(self):
         """Iterate over file and return PSMs one-by-one."""
         with open(self.filename, "rt") as open_file:
             reader = csv.DictReader(open_file, delimiter="\t")
+            failed_rows = 0
             for row in reader:
                 try:
                     yield PSM(**self._parse_entry(row))
-                except ValidationError:
+                except ValidationError as e:
+                    failed_rows += 1
                     logger.warning(f"Could not parse PSM from row: `{row}`")
-                    continue
+                    if failed_rows >= 3:
+                        raise PSMUtilsIOException(
+                            "Could not parse PSM from three consecutive rows. Verify that the "
+                            "file is formatted correctly as a psm_utils TSV file or that the "
+                            "correct file type reader is used."
+                        ) from e
+                else:
+                    failed_rows = 0
 
     @staticmethod
     def _parse_entry(entry: dict) -> dict:

tests/test_data/test.tsv

@@ -0,0 +1,4 @@
+spectrum_id	peptidoform
+peptide1	ACDEK/2
+peptide2	AC[Carbamidomethyl]DEFGR/3
+peptide3	[Acetyl]-AC[Carbamidomethyl]DEFGHIK/2

tests/test_io/test_tsv.py

@@ -1,6 +1,10 @@
 """Tests for psm_utils.io.tsv."""
 
-from psm_utils.io.tsv import TSVReader, TSVWriter  # noqa: F401
+import pytest
+
+from psm_utils.io.exceptions import PSMUtilsIOException
+from psm_utils.io.tsv import TSVReader
+from psm_utils.peptidoform import Peptidoform
 
 test_cases = [
     (
@@ -30,3 +34,19 @@ class TestTSVReader:
     def test__parse_entry(self):
         for test_in, expected_out in test_cases:
             assert TSVReader._parse_entry(test_in) == expected_out
+
+    def test_iter(self):
+        reader = TSVReader("tests/test_data/test.tsv")
+        for psm in reader:
+            assert psm.peptidoform == Peptidoform("ACDEK/2")
+            assert psm.spectrum_id == "peptide1"
+            assert psm.provenance_data == {}
+            assert psm.metadata == {}
+            assert psm.rescoring_features == {}
+            break
+
+    def test_iter_raises(self):
+        with TSVReader("tests/test_data/peprec.tsv") as reader:
+            with pytest.raises(PSMUtilsIOException):
+                for psm in reader:
+                    pass