Verify that patterns are anchored (#499)

As we transpile to some of the systems which always expect the regular
expressions to be anchored at the start and the end (*e.g.*, XSD
engines), we verify in the patch that all verification patterns are
appropriately anchored. This ensures that the patterns will also be
correctly transpiled down the line.

aas_core_codegen/intermediate/_translate.py

@@ -101,7 +101,10 @@
     TranspilableVerification,
     type_annotations_equal,
 )
-from aas_core_codegen.parse import tree as parse_tree
+from aas_core_codegen.parse import (
+    retree as parse_retree,
+    tree as parse_tree,
+)
 
 # pylint: disable=unused-argument
 
@@ -4403,6 +4406,85 @@ def _verify_only_simple_type_patterns(symbol_table: SymbolTable) -> List[Error]:
     return errors
 
 
+def _verify_patterns_anchored_at_start_and_end(
+    symbol_table: SymbolTable,
+) -> List[Error]:
+    """
+    Check that the patterns are anchored at the start (``^``) and at the end (``$``).
+
+    We need to make sure that all pattern verification functions can be transpiled into
+    many kinds of regular expressions. Some regular expression engines support
+    only patterns anchored at both ends (*e.g.*, XSD and the transpilation of regular
+    expressions to instructions of a virtual machine). To ensure inter-operability,
+    we decide to fail-fast here to prevent that non-anchored patterns propagate
+    downstream, which we might probably detect only much later.
+    """
+    errors = []  # type: List[Error]
+    for verification in symbol_table.verification_functions:
+        if isinstance(verification, PatternVerification):
+            regex, error = parse_retree.parse([verification.pattern])
+            if error is not None:
+                regex_line, pointer_line = parse_retree.render_pointer(error.cursor)
+
+                errors.append(
+                    Error(
+                        verification.parsed.node,
+                        f"Failed to parse the pattern of "
+                        f"the pattern verification function:\n"
+                        f"{error.message}\n"
+                        f"{regex_line}\n"
+                        f"{pointer_line}",
+                    )
+                )
+                continue
+
+            assert regex is not None
+
+            if (
+                len(regex.union.uniates) == 0
+                or len(regex.union.uniates[0].concatenants) == 0
+            ):
+                errors.append(
+                    Error(
+                        verification.parsed.node,
+                        f"The pattern is empty. We expect only non-empty patterns "
+                        f"all anchored at the start (``^``) and at the end (``$``) "
+                        f"for inter-operability with different regex engines, *e.g.*, "
+                        f"XSD engines. The pattern in question inferred for "
+                        f"the verification function {verification.name!r} "
+                        f"was: {verification.pattern}",
+                    )
+                )
+                continue
+
+            first_term = regex.union.uniates[0].concatenants[0]
+            last_term = regex.union.uniates[0].concatenants[-1]
+
+            if (
+                len(regex.union.uniates) != 1
+                or not isinstance(first_term.value, parse_retree.Symbol)
+                or not (first_term.value.kind is parse_retree.SymbolKind.START)
+                or not isinstance(last_term.value, parse_retree.Symbol)
+                or not (last_term.value.kind is parse_retree.SymbolKind.END)
+            ):
+                errors.append(
+                    Error(
+                        verification.parsed.node,
+                        f"(mristin, 2024-05-31): We expect all the patterns to be "
+                        f"anchored at the start (``^``) and at the end (``$``) for "
+                        f"inter-operability with different regex engines, *e.g.*, XSD "
+                        f"engines. Please consider re-writing your pattern with "
+                        f"a prefix ``^.*``, if you want to match an arbitrary prefix, "
+                        f"and a suffix ``.*$``, if you want to match an arbitrary "
+                        f"suffix. The pattern in question inferred for "
+                        f"the verification function {verification.name!r} "
+                        f"was: {verification.pattern}",
+                    )
+                )
+
+    return errors
+
+
 def _assert_interfaces_defined_correctly(
     symbol_table: SymbolTable, ontology: _hierarchy.Ontology
 ) -> None:
@@ -4497,6 +4579,8 @@ def _verify(symbol_table: SymbolTable, ontology: _hierarchy.Ontology) -> List[Er
 
     errors.extend(_verify_only_simple_type_patterns(symbol_table=symbol_table))
 
+    errors.extend(_verify_patterns_anchored_at_start_and_end(symbol_table=symbol_table))
+
     if len(errors) > 0:
         return errors
 

test_data/csharp/test_verification/pattern_verification/utf32_as_literal/in_union/expected_verification.cs

@@ -22,7 +22,7 @@ public static class Verification
         [CodeAnalysis.SuppressMessage("ReSharper", "StringLiteralTypo")]
         private static Regex _constructMatchSomething()
         {
-            var pattern = "^\\ud800\\udc00|something$";
+            var pattern = "^(\\ud800\\udc00|something)$";
 
             return new Regex(pattern);
         }

test_data/csharp/test_verification/pattern_verification/utf32_as_literal/in_union/meta_model.py

@@ -1,6 +1,6 @@
 @verification
 def match_something(text: str) -> bool:
-    pattern = f"^\\U00010000|something$"
+    pattern = "^(\\U00010000|something)$"
     return match(pattern, text) is not None
 
 

tests/csharp/test_description.py

@@ -471,7 +471,7 @@ def test_only_summary(self) -> None:
             @verification
             def verify_something(text: str) -> bool:
                 """Verify something."""
-                return match(r'.*', text) is not None
+                return match(r'^.*$', text) is not None
 
             __version__ = "dummy"
             __xml_namespace__ = "https://dummy.com"

tests/infer_for_schema/test_patterns_on_properties.py

@@ -58,7 +58,7 @@ def test_single_pattern(self) -> None:
             @verification
             def matches_something(text: str) -> bool:
                 prefix = "something"
-                return match(f"{prefix}-[a-zA-Z]+", text) is not None
+                return match(f"^{prefix}-[a-zA-Z]+$", text) is not None
 
 
             @invariant(
@@ -100,7 +100,7 @@ def __init__(self, some_property: str) -> None:
                   patterns_by_property={
                     'some_property': [
                       PatternConstraint(
-                        pattern='something-[a-zA-Z]+')]},
+                        pattern='^something-[a-zA-Z]+$')]},
                   set_of_primitives_by_property={},
                   set_of_enumeration_literals_by_property={})"""
             ),
@@ -112,11 +112,11 @@ def test_two_patterns(self) -> None:
             """\
             @verification
             def matches_something(text: str) -> bool:
-                return match("something-[a-zA-Z]+", text) is not None
+                return match("^something-[a-zA-Z]+$", text) is not None
 
             @verification
             def matches_acme(text: str) -> bool:
-                return match(".*acme.*", text) is not None
+                return match("^.*acme.*$", text) is not None
 
 
             @invariant(
@@ -162,9 +162,9 @@ def __init__(self, some_property: str) -> None:
                   patterns_by_property={
                     'some_property': [
                       PatternConstraint(
-                        pattern='something-[a-zA-Z]+'),
+                        pattern='^something-[a-zA-Z]+$'),
                       PatternConstraint(
-                        pattern='.*acme.*')]},
+                        pattern='^.*acme.*$')]},
                   set_of_primitives_by_property={},
                   set_of_enumeration_literals_by_property={})"""
             ),
@@ -176,7 +176,7 @@ def test_conditioned_on_property(self) -> None:
             """\
             @verification
             def matches_something(text: str) -> bool:
-                return match("something-[a-zA-Z]+", text) is not None
+                return match("^something-[a-zA-Z]+$", text) is not None
 
             @invariant(
                 lambda self:
@@ -224,7 +224,7 @@ def __init__(self, some_property: Optional[str] = None) -> None:
                   patterns_by_property={
                     'some_property': [
                       PatternConstraint(
-                        pattern='something-[a-zA-Z]+')]},
+                        pattern='^something-[a-zA-Z]+$')]},
                   set_of_primitives_by_property={},
                   set_of_enumeration_literals_by_property={})"""
             ),
@@ -236,11 +236,11 @@ def test_no_inheritance_by_default(self) -> None:
             """\
             @verification
             def matches_something(text: str) -> bool:
-                return match("something-[a-zA-Z]+", text) is not None
+                return match("^something-[a-zA-Z]+$", text) is not None
 
             @verification
             def matches_acme(text: str) -> bool:
-                return match(".*acme.*", text) is not None
+                return match("^.*acme.*$", text) is not None
 
             @invariant(
                 lambda self: matches_something(self.some_property),
@@ -298,7 +298,7 @@ def __init__(self, some_property: str) -> None:
                   patterns_by_property={
                     'some_property': [
                       PatternConstraint(
-                        pattern='.*acme.*')]},
+                        pattern='^.*acme.*$')]},
                   set_of_primitives_by_property={},
                   set_of_enumeration_literals_by_property={})"""
             ),
@@ -313,7 +313,7 @@ def test_no_inheritance_involved(self) -> None:
             @verification
             def matches_something(text: str) -> bool:
                 prefix = "something"
-                return match(f"{prefix}-[a-zA-Z]+", text) is not None
+                return match(f"^{prefix}-[a-zA-Z]+$", text) is not None
 
 
             @invariant(
@@ -370,7 +370,7 @@ def __init__(self, some_property: str) -> None:
                   patterns_by_property={
                     'some_property': [
                       PatternConstraint(
-                        pattern='something-[a-zA-Z]+')]},
+                        pattern='^something-[a-zA-Z]+$')]},
                   set_of_primitives_by_property={},
                   set_of_enumeration_literals_by_property={})"""
             ),
@@ -383,7 +383,7 @@ def test_inheritance_from_parent_with_no_patterns_of_own(self) -> None:
             @verification
             def matches_something(text: str) -> bool:
                 prefix = "something"
-                return match(f"{prefix}-[a-zA-Z]+", text) is not None
+                return match(f"^{prefix}-[a-zA-Z]+$", text) is not None
 
 
             @invariant(
@@ -440,7 +440,7 @@ def __init__(self, some_property: str) -> None:
                   patterns_by_property={
                     'some_property': [
                       PatternConstraint(
-                        pattern='something-[a-zA-Z]+')]},
+                        pattern='^something-[a-zA-Z]+$')]},
                   set_of_primitives_by_property={},
                   set_of_enumeration_literals_by_property={})"""
             ),
@@ -452,11 +452,11 @@ def test_merge_with_parent(self) -> None:
             """\
             @verification
             def matches_something(text: str) -> bool:
-                return match("something-[a-zA-Z]+", text) is not None
+                return match("^something-[a-zA-Z]+$", text) is not None
 
             @verification
             def matches_acme(text: str) -> bool:
-                return match(".*acme.*", text) is not None
+                return match("^.*acme.*$", text) is not None
 
             @invariant(
                 lambda self: matches_something(self.some_property),
@@ -521,9 +521,9 @@ def __init__(self, some_property: str) -> None:
                   patterns_by_property={
                     'some_property': [
                       PatternConstraint(
-                        pattern='.*acme.*'),
+                        pattern='^.*acme.*$'),
                       PatternConstraint(
-                        pattern='something-[a-zA-Z]+')]},
+                        pattern='^something-[a-zA-Z]+$')]},
                   set_of_primitives_by_property={},
                   set_of_enumeration_literals_by_property={})"""
             ),
@@ -535,7 +535,7 @@ def test_merge_with_parent_and_grand_parent(self) -> None:
             """\
             @verification
             def matches_something(text: str) -> bool:
-                return match("something-[a-zA-Z]+", text) is not None
+                return match("^something-[a-zA-Z]+$", text) is not None
 
             @invariant(
                 lambda self: matches_something(self.some_property),
@@ -599,7 +599,7 @@ def __init__(self, some_property: str) -> None:
                   patterns_by_property={
                     'some_property': [
                       PatternConstraint(
-                        pattern='something-[a-zA-Z]+')]},
+                        pattern='^something-[a-zA-Z]+$')]},
                   set_of_primitives_by_property={},
                   set_of_enumeration_literals_by_property={})"""
             ),
@@ -611,7 +611,7 @@ def test_merge_with_parent_over_constrained_primitive(self) -> None:
             """\
             @verification
             def matches_something(text: str) -> bool:
-                return match("something-[a-zA-Z]+", text) is not None
+                return match("^something-[a-zA-Z]+$", text) is not None
 
             @invariant(
                 lambda self: matches_something(self),
@@ -674,7 +674,7 @@ def __init__(self, some_property: SomeConstrainedPrimitive) -> None:
                   patterns_by_property={
                     'some_property': [
                       PatternConstraint(
-                        pattern='something-[a-zA-Z]+')]},
+                        pattern='^something-[a-zA-Z]+$')]},
                   set_of_primitives_by_property={},
                   set_of_enumeration_literals_by_property={})"""
             ),

tests/infer_for_schema/test_patterns_on_self.py

@@ -61,7 +61,7 @@ def test_single_pattern(self) -> None:
             @verification
             def matches_something(text: str) -> bool:
                 prefix = "something"
-                return match(f"{prefix}-[a-zA-Z]+", text) is not None
+                return match(f"^{prefix}-[a-zA-Z]+$", text) is not None
 
 
             @invariant(
@@ -106,7 +106,7 @@ def __init__(self, some_property: Some_constrained_primitive) -> None:
                   patterns_by_property={
                     'some_property': [
                       PatternConstraint(
-                        pattern='something-[a-zA-Z]+')]},
+                        pattern='^something-[a-zA-Z]+$')]},
                   set_of_primitives_by_property={},
                   set_of_enumeration_literals_by_property={})"""
             ),
@@ -118,11 +118,11 @@ def test_two_patterns(self) -> None:
             """\
             @verification
             def matches_something(text: str) -> bool:
-                return match("something-[a-zA-Z]+", text) is not None
+                return match("^something-[a-zA-Z]+$", text) is not None
 
             @verification
             def matches_acme(text: str) -> bool:
-                return match(".*acme.*", text) is not None
+                return match("^.*acme.*$", text) is not None
 
 
             @invariant(
@@ -171,9 +171,9 @@ def __init__(self, some_property: Some_constrained_primitive) -> None:
                   patterns_by_property={
                     'some_property': [
                       PatternConstraint(
-                        pattern='something-[a-zA-Z]+'),
+                        pattern='^something-[a-zA-Z]+$'),
                       PatternConstraint(
-                        pattern='.*acme.*')]},
+                        pattern='^.*acme.*$')]},
                   set_of_primitives_by_property={},
                   set_of_enumeration_literals_by_property={})"""
             ),
@@ -185,11 +185,11 @@ def test_inheritance_between_constrained_primitives_by_default(self) -> None:
             """\
             @verification
             def matches_something(text: str) -> bool:
-                return match("something-[a-zA-Z]+", text) is not None
+                return match("^something-[a-zA-Z]+$", text) is not None
 
             @verification
             def matches_acme(text: str) -> bool:
-                return match(".*acme.*", text) is not None
+                return match("^.*acme.*$", text) is not None
 
             @invariant(
                 lambda self: matches_something(self),
@@ -244,9 +244,9 @@ def __init__(self, some_property: Some_constrained_primitive) -> None:
                   patterns_by_property={
                     'some_property': [
                       PatternConstraint(
-                        pattern='something-[a-zA-Z]+'),
+                        pattern='^something-[a-zA-Z]+$'),
                       PatternConstraint(
-                        pattern='.*acme.*')]},
+                        pattern='^.*acme.*$')]},
                   set_of_primitives_by_property={},
                   set_of_enumeration_literals_by_property={})"""
             ),