Add exception for invalid encodings (#88)

Co-authored-by: Mathias Kuhring <mathias.kuhring@bih-charite.de> Co-authored-by: Thomas Sell <thomas.sell@bih-charite.de>
bihealth · Jan 15, 2024 · 2892ffb · 2892ffb
1 parent 8a9ecbd
commit 2892ffb
Show file tree

Hide file tree

Showing 2 changed files with 9 additions and 0 deletions.
diff --git a/altamisa/isatab/parse_assay_study.py b/altamisa/isatab/parse_assay_study.py
@@ -848,6 +848,9 @@ def _read_next_line(self):
                 self.unique_rows.add("\t".join(self._line))
         except StopIteration:
             self._line = None
+        except UnicodeDecodeError as e:  # pragma: no cover
+            msg = f"Invalid encoding of study file '{self._filename}' (use Unicode/UTF-8)."
+            raise ParseIsatabException(msg) from e
         return prev_line
 
     def read(self):
@@ -970,6 +973,9 @@ def _read_next_line(self):
                 self.unique_rows.add("\t".join(self._line))
         except StopIteration:
             self._line = None
+        except UnicodeDecodeError as e:  # pragma: no cover
+            msg = f"Invalid encoding of assay file '{self._filename}' (use Unicode/UTF-8)."
+            raise ParseIsatabException(msg) from e
         return prev_line
 
     def read(self):

diff --git a/altamisa/isatab/parse_investigation.py b/altamisa/isatab/parse_investigation.py
@@ -128,6 +128,9 @@ def _read_next_line(self) -> Optional[List[str]]:
                 self._line = list_strip(next(self._reader))
         except StopIteration:
             self._line = None
+        except UnicodeDecodeError as e:  # pragma: no cover
+            msg = f"Invalid encoding of investigation file '{self._filename}' (use Unicode/UTF-8)."
+            raise ParseIsatabException(msg) from e
         return prev_line
 
     def _next_line_startswith_comment(self):