Override matches with re2 implementation

Author: Steve Ayers

.github/CONTRIBUTING.md

@@ -19,8 +19,12 @@ git clone git@github.com:bufbuild/protovalidate-python.git
 cd protovalidate-python
 ```
 
-Then, make any changes you'd like. We use a Makefile to test and lint our code,
-so you'll need a few non-Python tools:
+Next, install dependencies. You will need:
+
+* Python >=v3.9
+* [Pipenv](https://pipenv.pypa.io/en/latest/index.html)
+
+We use a Makefile to test and lint our code, so you'll also need a few non-Python tools:
 
 * GNU Make (to use the Makefile): part of the `build-essential` package on
   Debian-derived Linux distributions (including Ubuntu), and part of

Makefile

@@ -62,7 +62,6 @@ lint: install ## Lint code
 
 .PHONY: install
 install: ## Install dependencies
-	$(PYTHON) -m pip install --upgrade pip pipenv
 	pipenv --python $(PYTHON) sync --dev
 
 .PHONY: checkgenerate

Pipfile

@@ -6,6 +6,7 @@ name = "pypi"
 [packages]
 cel-python = "==0.2.*"
 protobuf = "==6.*"
+google-re2 = "*"
 
 [dev-packages]
 pytest = "*"

Pipfile.lock

@@ -29,6 +29,19 @@
 )
 
 
+# Currently cel-python does not support re2. So we are overriding their `matches`
+# function with our own that leverages re2. Note there is a PR in cel-python to
+# add re2 support. See https://github.com/cloud-custodian/cel-python/pull/67.
+# Once that lands, this `matches` override can be removed.
+def cel_matches(text: str, pattern: str) -> celpy.Result:
+    try:
+        m = re2.search(pattern, text)
+    except re2.error as ex:
+        msg = "match error"
+        raise celpy.CELEvalError(msg, ex.__class__, ex.args) from ex
+    return celtypes.BoolType(m is not None)
+
+
 def cel_get_field(message: celtypes.Value, field_name: celtypes.Value) -> celpy.Result:
     if not isinstance(message, MessageType):
         msg = "invalid argument, expected message"
@@ -1553,22 +1566,14 @@ def __peek(self, char: str) -> bool:
         return self._index < len(self._string) and self._string[self._index] == char
 
 
-def matches(text: str, pattern: str) -> celpy.Result:
-    try:
-        m = re2.search(pattern, text)
-    except re2.error as ex:
-        return celpy.CELEvalError("match error", ex.__class__, ex.args)
-    return celtypes.BoolType(m is not None)
-
-
 def make_extra_funcs(locale: str) -> dict[str, celpy.CELFunction]:
     # TODO(#257): Fix types and add tests for StringFormat.
     # For now, ignoring the type.
     string_fmt = string_format.StringFormat(locale)  # type: ignore
     return {
         # Missing standard functions
         "format": string_fmt.format,
-        "matches": matches,
+        "matches": cel_matches,
         # protovalidate specific functions
         "getField": cel_get_field,
         "isNan": cel_is_nan,

protovalidate/internal/extra_func.py

@@ -6,7 +6,8 @@ build-backend = "hatchling.build"
 name = "protovalidate"
 description = "Protocol Buffer Validation for Python"
 readme = "README.md"
-license = { file = "LICENSE" }
+license-expression = "Apache 2.0"
+license-file = "LICENSE"
 keywords = ["validate", "protobuf", "protocol buffer"]
 requires-python = ">=3.9"
 classifiers = [
@@ -15,12 +16,11 @@ classifiers = [
   "Programming Language :: Python :: 3.11",
   "Programming Language :: Python :: 3.12",
   "Programming Language :: Python :: 3.13",
-  "License :: OSI Approved :: Apache Software License",
   "Operating System :: OS Independent",
   "Typing :: Typed",
 ]
 dynamic = ["version"]
-dependencies = ["protobuf", "cel-python"]
+dependencies = ["protobuf", "cel-python", "google-re2"]
 
 [project.urls]
 Homepage = "https://github.com/bufbuild/protovalidate-python"

pyproject.toml

@@ -11,3 +11,21 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+
+import unittest
+
+import pytest
+from celpy import CELEvalError, celtypes
+
+from protovalidate.internal.extra_func import cel_matches
+
+
+class TestMatches(unittest.TestCase):
+    def test_re2_end_of_string(self):
+        empty_string = celtypes.StringType("")
+        self.assertTrue(cel_matches(empty_string, "^\\z"))
+
+    def test_re2_invalid_end_of_string(self):
+        empty_string = celtypes.StringType("")
+        with pytest.raises(CELEvalError):
+            self.assertTrue(cel_matches(empty_string, "^\\Z"))