Strip stray unquoted backslashes in action normalization

bionetgen/modelapi/bngparser.py

@@ -27,9 +27,43 @@ def _normalize_action_text(action: str) -> str:
     """
     text = _strip_comment_outside_quotes(action)
     text = _collapse_unquoted_whitespace(text)
+    text = _strip_unquoted_backslashes(text)
     return text.strip()
 
 
+def _strip_unquoted_backslashes(text: str) -> str:
+    """Drop ``\\`` characters that appear outside string literals."""
+    return _filter_outside_quotes(text, lambda ch: ch != "\\")
+
+
+def _filter_outside_quotes(text: str, keep) -> str:
+    out = []
+    in_single = False
+    in_double = False
+    escaped = False
+    for ch in text:
+        if escaped:
+            out.append(ch)
+            escaped = False
+            continue
+        if ch == "\\" and (in_single or in_double):
+            out.append(ch)
+            escaped = True
+            continue
+        if ch == '"' and not in_single:
+            in_double = not in_double
+            out.append(ch)
+            continue
+        if ch == "'" and not in_double:
+            in_single = not in_single
+            out.append(ch)
+            continue
+        if not in_single and not in_double and not keep(ch):
+            continue
+        out.append(ch)
+    return "".join(out)
+
+
 def _strip_comment_outside_quotes(text: str) -> str:
     out = []
     in_single = False

tests/test_bng_parsing.py

@@ -73,3 +73,20 @@ def test_pattern_canonicalization():
             break
     # assert that everything matched up
     assert res is True
+
+
+def test_action_normalization_drops_stray_backslashes_outside_quotes():
+    from bionetgen.modelapi.bngparser import _normalize_action_text
+
+    out = _normalize_action_text(
+        'parameter_scan({n_scan_pts=>101,\\log_scale=>1,method=>"ode"})'
+    )
+    assert "\\" not in out
+    assert "log_scale=>1" in out
+
+
+def test_action_normalization_preserves_backslashes_inside_quotes():
+    from bionetgen.modelapi.bngparser import _normalize_action_text
+
+    out = _normalize_action_text('action({arg=>"a\\b"})')
+    assert '"a\\b"' in out