Add correctness tests for lookahead expressions.

Signed-off-by: Igor Peshansky <ipeshansky@nvidia.com>

Author: igorpeshansky

integration_tests/src/main/python/regexp_test.py

@@ -1040,6 +1040,16 @@ def test_rlike_fallback_possessive_quantifier():
             'RLike',
             conf=_regexp_conf)
 
+@allow_non_gpu('ProjectExec', 'RLike')
+def test_rlike_fallback_lookaheads():
+    gen = mk_str_gen('(\u20ac|\\w){0,3}a[|b*.$\r\n]{0,2}c\\w{0,3}')
+    for pattern in ['a(?=a*)', 'a(?!a*)']:
+        assert_gpu_fallback_collect(
+            lambda spark, pattern=pattern: unary_op_df(spark, gen).selectExpr(
+                f'a rlike "{pattern}"'),
+            'RLike',
+            conf=_regexp_conf)
+
 def test_regexp_extract_all_idx_zero():
     gen = mk_str_gen('[abcd]{0,3}[0-9]{0,3}-[0-9]{0,3}[abcd]{1,3}')
     assert_gpu_and_cpu_are_equal_collect(

tests/src/test/scala/com/nvidia/spark/rapids/RegularExpressionParserSuite.scala

@@ -248,6 +248,15 @@ class RegularExpressionParserSuite extends AnyFunSuite {
         RegexRepetition(RegexChar('a'), SimpleQuantifier('?')))), None))))
   }
 
+  test("group not starting with ? is a capturing group") {
+    assert(parse("(=a)") === RegexSequence(ListBuffer(
+      RegexGroup(true, RegexSequence(ListBuffer(
+        RegexChar('='), RegexChar('a'))), None)))
+    assert(parse("(!a)") === RegexSequence(ListBuffer(
+      RegexGroup(true, RegexSequence(ListBuffer(
+        RegexChar('!'), RegexChar('a'))), None)))
+  }
+
   test("complex expression") {
     val ast = parse(
       "^" +            // start of line

tests/src/test/scala/com/nvidia/spark/rapids/RegularExpressionTranspilerSuite.scala

@@ -159,13 +159,13 @@ class RegularExpressionTranspilerSuite extends AnyFunSuite {
   }
 
   test("cuDF does not support positive or negative lookahead") {
-    val negPatterns = Seq("a(!b)", "a(!b)c?")
+    val negPatterns = Seq("a(?!b)", "a(?!b)c?")
     negPatterns.foreach(pattern =>
       assertUnsupported(pattern, RegexFindMode,
         "Negative lookahead groups are not supported")
     )
 
-    val posPatterns = Seq("a(=b)", "a(=b)c?")
+    val posPatterns = Seq("a(?=b)", "a(?=b)c?")
     posPatterns.foreach(pattern =>
       assertUnsupported(pattern, RegexFindMode,
         "Positive lookahead groups are not supported")