Use re2 instead of re for match expressions

CEL specifies that match should use re2 pattern matching and not re
semantics.

Author: ddn0

poetry.lock

@@ -44,6 +44,7 @@ python-dateutil = "^2.9.0.post0"
 pyyaml = "^6.0.1"
 types-pyyaml = "^6.0.12.20240311"
 types-python-dateutil = "^2.9.0.20240316"
+google-re2 = "^1.1.20240501"
 
 [tool.poetry.group.dev.dependencies]
 behave = "^1.2.6"

pyproject.toml

@@ -39,6 +39,7 @@
 import logging
 import operator
 import re
+import re2
 import sys
 from functools import reduce, wraps
 from typing import (Any, Callable, Dict, Iterable, Iterator, List, Mapping,
@@ -62,7 +63,6 @@
 
 logger = logging.getLogger("evaluation")
 
-
 class CELSyntaxError(Exception):
     """CEL Syntax error -- the AST did not have the expected structure."""
     def __init__(self, arg: Any, line: Optional[int] = None, column: Optional[int] = None) -> None:
@@ -293,6 +293,15 @@ def operator_in(item: Result, container: Result) -> Result:
     return result
 
 
+def function_matches(text: str, pattern: str):
+    try:
+        m = re2.search(pattern, text)
+    except re2.error as ex:
+        return CELEvalError("match error", ex.__class__, ex.args)
+
+    return celpy.celtypes.BoolType(m is not None)
+
+
 def function_size(container: Result) -> Result:
     """
     The size() function applied to a Value. Delegate to Python's :py:func:`len`.
@@ -340,7 +349,7 @@ def function_size(container: Result) -> Result:
     # StringType methods
     "endsWith": lambda s, text: celpy.celtypes.BoolType(s.endswith(text)),
     "startsWith": lambda s, text: celpy.celtypes.BoolType(s.startswith(text)),
-    "matches": lambda s, pattern: celpy.celtypes.BoolType(re.search(pattern, s) is not None),
+    "matches": function_matches,
     "contains": lambda s, text: celpy.celtypes.BoolType(text in s),
     # TimestampType methods. Type details are redundant, but required because of the lambdas
     "getDate": lambda ts, tz_name=None: celpy.celtypes.IntType(ts.getDate(tz_name)),

src/celpy/evaluation.py

@@ -158,6 +158,13 @@ def test_operator_in():
     assert isinstance(operator_in(celtypes.IntType(-1), container_2), CELEvalError)
 
 
+def test_function_matches():
+    empty_string = celtypes.StringType("")
+    # re2-specific patterns which behave differently than standard re
+    assert function_matches(empty_string, "^\\z")
+    assert isinstance(function_matches(empty_string, "^\\Z"), CELEvalError)
+
+
 def test_function_size():
     container_1 = celtypes.ListType([
         celtypes.IntType(42),