Tilda special character in path parameters fix

Author: p1c2u

openapi_core/templating/paths/parsers.py

@@ -1,4 +1,9 @@
+from __future__ import annotations
+
+from dataclasses import dataclass
 from typing import Any
+from typing import Dict
+from typing import Optional
 
 from parse import Parser
 
@@ -18,16 +23,115 @@ class PathParser(Parser):  # type: ignore
     def __init__(
         self, pattern: str, pre_expression: str = "", post_expression: str = ""
     ) -> None:
+        self._orig_to_safe: Dict[str, str] = {}
+        self._safe_to_orig: Dict[str, str] = {}
+        self._safe_suffix_counters: Dict[str, int] = {}
         extra_types = {
             self.parse_path_parameter.name: self.parse_path_parameter
         }
-        super().__init__(pattern, extra_types)
+        sanitized_pattern = self._sanitize_pattern(pattern)
+        super().__init__(sanitized_pattern, extra_types)
         self._expression: str = (
             pre_expression + self._expression + post_expression
         )
 
+    def search(self, string: str, pos: int = 0, endpos: Optional[int] = None) -> Any:
+        result = super().search(string, pos=pos, endpos=endpos)
+        if not result:
+            return result
+        return _RemappedResult(result, self._safe_to_orig)
+
+    def parse(self, string: str, pos: int = 0, endpos: Optional[int] = None) -> Any:
+        result = super().parse(string, pos=pos, endpos=endpos)
+        if not result:
+            return result
+        return _RemappedResult(result, self._safe_to_orig)
+
+    def _get_safe_field_name(self, original: str) -> str:
+        existing = self._orig_to_safe.get(original)
+        if existing is not None:
+            return existing
+
+        safe_parts = []
+        for ch in original:
+            if ch == "_" or ch.isalnum():
+                safe_parts.append(ch)
+            else:
+                safe_parts.append(f"__{ord(ch):x}__")
+
+        safe = "".join(safe_parts) or "p"
+        # `parse` and Python `re` named groups are most reliable when the group name
+        # starts with a letter.
+        if not safe[0].isalpha():
+            safe = f"p_{safe}"
+
+        # Ensure uniqueness across fields within this parser
+        if safe in self._safe_to_orig and self._safe_to_orig[safe] != original:
+            base = safe
+            suffix = self._safe_suffix_counters.get(base, 1)
+            while True:
+                candidate = f"{base}__{suffix}"
+                if candidate not in self._safe_to_orig:
+                    safe = candidate
+                    self._safe_suffix_counters[base] = suffix + 1
+                    break
+                suffix += 1
+
+        self._orig_to_safe[original] = safe
+        self._safe_to_orig[safe] = original
+        return safe
+
+    def _sanitize_pattern(self, pattern: str) -> str:
+        # Pre-sanitize field names inside `{...}` before `parse` processes them.
+        # This ensures special characters (e.g. `~`) and digit-leading names are
+        # treated as named fields instead of literals or positional groups.
+        if "{" not in pattern:
+            return pattern
+
+        out: list[str] = []
+        i = 0
+        n = len(pattern)
+        while i < n:
+            ch = pattern[i]
+            if ch != "{":
+                out.append(ch)
+                i += 1
+                continue
+
+            end = pattern.find("}", i + 1)
+            if end == -1:
+                out.append(ch)
+                i += 1
+                continue
+
+            original = pattern[i + 1 : end]
+            safe = self._get_safe_field_name(original)
+            out.append("{")
+            out.append(safe)
+            out.append("}")
+            i = end + 1
+
+        return "".join(out)
+
     def _handle_field(self, field: str) -> Any:
         # handle as path parameter field
-        field = field[1:-1]
-        path_parameter_field = "{%s:PathParameter}" % field
+        safe_field = field[1:-1]
+        path_parameter_field = "{%s:PathParameter}" % safe_field
         return super()._handle_field(path_parameter_field)
+
+
+@dataclass(frozen=True)
+class _RemappedResult:
+    _result: Any
+    _safe_to_orig: Dict[str, str]
+
+    @property
+    def named(self) -> Dict[str, Any]:
+        named = getattr(self._result, "named", {})
+        return {self._safe_to_orig.get(k, k): v for k, v in named.items()}
+
+    def __bool__(self) -> bool:
+        return bool(self._result)
+
+    def __getattr__(self, item: str) -> Any:
+        return getattr(self._result, item)

tests/unit/templating/test_paths_parsers.py

@@ -32,6 +32,7 @@ def test_exact(self):
             ("/{test_id}/test", {"test_id": "test"}),
             ("/{test.id}/test", {"test.id": "test"}),
             ("/{test-id}/test", {"test-id": "test"}),
+            ("/{0test}/test", {"0test": "test"}),
         ],
     )
     def test_chars_valid(self, path_pattern, expected):
@@ -42,17 +43,11 @@ def test_chars_valid(self, path_pattern, expected):
 
         assert result.named == expected
 
-    @pytest.mark.xfail(
-        reason=(
-            "Special characters of regex not supported. "
-            "See https://github.com/python-openapi/openapi-core/issues/672"
-        ),
-        strict=True,
-    )
     @pytest.mark.parametrize(
         "path_pattern,expected",
         [
             ("/{test~id}/test", {"test~id": "test"}),
+            ("/{a-b~c.d}/test", {"a-b~c.d": "test"}),
         ],
     )
     def test_special_chars_valid(self, path_pattern, expected):