Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Add support for word boundaries \b and \B #5479

Merged
merged 10 commits into from
Jun 6, 2022
Merged

Add support for word boundaries \b and \B #5479

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
9594627
Enable \b and \B
anthony-chang May 12, 2022
85ef037
Add unit test for the fallback to CPU in split mode
anthony-chang May 12, 2022
2d6b7a4
Use double backslack for integration tests
anthony-chang May 16, 2022
bd7f1a0
Fix integration tests generating unwanted chars in input
anthony-chang May 17, 2022
05c2189
Merge branch 'branch-22.06' of github.com:nvidia/spark-rapids into su…
anthony-chang May 25, 2022
53ab633
Merge branch 'branch-22.06' of github.com:nvidia/spark-rapids into su…
anthony-chang May 27, 2022
07f3111
[WIP] save
anthony-chang May 27, 2022
a7d1f16
Fall back to CPU for $\B
anthony-chang May 27, 2022
89df884
Also fallback for $\b
anthony-chang May 28, 2022
9edea9e
Fix build
anthony-chang May 30, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 14 additions & 0 deletions integration_tests/src/main/python/string_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -836,6 +836,20 @@ def test_regexp_extract_idx_0():
'regexp_extract(a, "^([a-d]*)[0-9]*([a-d]*)\\z", 0)'),
conf=_regexp_conf)

def test_word_boundaries():
gen = mk_str_gen('([abc]{1,3}[\r\n\t \f]{0,2}[123]){1,5}')
assert_gpu_and_cpu_are_equal_collect(
lambda spark: debug_df(unary_op_df(spark, gen).selectExpr('a',
'rlike(a, "\\\\b")',
'rlike(a, "\\\\B")',
'rlike(a, "\\\\b\\\\B")',
'regexp_extract(a, "([a-d]+)\\\\b([e-h]+)", 1)',
'regexp_extract(a, "([a-d]+)\\\\B", 1)',
'regexp_replace(a, "\\\\b", "")',
'regexp_replace(a, "\\\\B", "")',
)),
conf=_regexp_conf)

def test_regexp_hexadecimal_digits():
gen = mk_str_gen(
'[abcd]\\\\x00\\\\x7f\\\\x80\\\\xff\\\\x{10ffff}\\\\x{00eeee}[\\\\xa0-\\\\xb0][abcd]')
Expand Down
6 changes: 3 additions & 3 deletions sql-plugin/src/main/scala/com/nvidia/spark/rapids/RegexParser.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -643,9 +643,9 @@ class CudfRegexTranspiler(mode: RegexMode) {
case 'W' =>
// see https://github.com/NVIDIA/spark-rapids/issues/4475
throw new RegexUnsupportedException("non-word class \\W is not supported")
case 'b' | 'B' =>
// see https://github.com/NVIDIA/spark-rapids/issues/4517
throw new RegexUnsupportedException("word boundaries are not supported")
case 'b' | 'B' if mode == RegexSplitMode =>
// see https://github.com/NVIDIA/spark-rapids/issues/5478
throw new RegexUnsupportedException("word boundaries are not supported in split mode")
case 'A' if mode == RegexSplitMode =>
throw new RegexUnsupportedException("string anchor \\A is not supported in split mode")
case 'Z' if mode == RegexSplitMode || mode == RegexReplaceMode =>
Expand Down
9 changes: 8 additions & 1 deletion tests/src/test/scala/com/nvidia/spark/rapids/RegularExpressionTranspilerSuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -175,7 +175,7 @@ class RegularExpressionTranspilerSuite extends FunSuite with Arm {
assertCpuGpuMatchesRegexpFind(patterns, Seq("", "\u0007", "a\u0007b",
"\u0007\u003f\u007f", "\u0080", "a\u00fe\u00ffb", "ab\ueeeecd"))
}

test("string anchors - find") {
val patterns = Seq("\\Atest", "\\A+test", "\\A{1}test", "\\A{1,}test",
"(\\A)+test", "(\\A){1}test", "(\\A){1,}test", "test\\z")
Expand Down Expand Up @@ -245,6 +245,13 @@ class RegularExpressionTranspilerSuite extends FunSuite with Arm {
assertCpuGpuMatchesRegexpFind(patterns, inputs)
}

test ("word boundaries will fall back to CPU - split") {
val patterns = Seq("\\b", "\\B")
patterns.foreach(pattern =>
assertUnsupported(pattern, RegexSplitMode, "word boundaries are not supported in split mode")
)
}

test("whitespace boundaries - replace") {
assertCpuGpuMatchesRegexpReplace(
Seq("\\s", "\\S"),
Expand Down