YQL-19747 Extract NormalizeName from yql/.../core/ast

We need the `NYql::NormalizeName` at `yql/.../sql/v1/complete`, but do not want to depend on the entire `yql/.../core/ast` module, so moving it to `yql/.../core/sql_types` module.

We need it to find equivalent names for a proper candidates ranking as `MULTI_AGGREGATE_BY` and `MultiAggregateBy` are equivalent and their frequencies should be merged. Also we can apply this function to `NameRequest::prefix` and suggest `MULTI_AGGREGATE_BY` on `prefix.StartsWith("multia")`.

- Related to https://github.com/ydb-platform/ydb/issues/9056
- Related to https://github.com/vityaman/ydb/issues/21

---

Pull Request resolved: https://github.com/ytsaurus/ytsaurus/pull/1204
commit_hash:02f3ef316bbca818e22aafe88edfde8e4eb214c3

9 files changed, 89 insertions, 65 deletions

diff --git a/yql/essentials/ast/ya.make b/yql/essentials/ast/ya.make
index 58c21a1d7f1..a0a003cb2eb 100644
--- a/yql/essentials/ast/ya.make
+++ b/yql/essentials/ast/ya.make
@@ -37,6 +37,7 @@ PEERDIR(
     yql/essentials/utils
     yql/essentials/utils/fetch
     yql/essentials/core/issue
+    yql/essentials/core/sql_types
     yql/essentials/core/url_lister/interface
     yql/essentials/parser/pg_catalog
 )
diff --git a/yql/essentials/ast/yql_expr.cpp b/yql/essentials/ast/yql_expr.cpp
index 21d0b9a446c..fb60a08476b 100644
--- a/yql/essentials/ast/yql_expr.cpp
+++ b/yql/essentials/ast/yql_expr.cpp
@@ -3830,60 +3830,6 @@ const TTypeAnnotationNode& RemoveOptionality(const TTypeAnnotationNode& type) {
     return ETypeAnnotationKind::Optional == type.GetKind() ? *type.Cast<TOptionalExprType>()->GetItemType() : type;
 }
 
-TMaybe<TIssue> NormalizeName(TPosition position, TString& name) {
-    const ui32 inputLength = name.length();
-    ui32 startCharPos = 0;
-    ui32 totalSkipped = 0;
-    bool atStart = true;
-    bool justSkippedUnderscore = false;
-    for (ui32 i = 0; i < inputLength; ++i) {
-        const char c = name.at(i);
-        if (c == '_') {
-            if (!atStart) {
-                if (justSkippedUnderscore) {
-                    return TIssue(position, TStringBuilder() << "\"" << name << "\" looks weird, has multiple consecutive underscores");
-                }
-                justSkippedUnderscore = true;
-                ++totalSkipped;
-                continue;
-            } else {
-                ++startCharPos;
-            }
-        }
-        else {
-            atStart = false;
-            justSkippedUnderscore = false;
-        }
-    }
-
-    if (totalSkipped >= 5) {
-        return TIssue(position, TStringBuilder() << "\"" << name << "\" looks weird, has multiple consecutive underscores");
-    }
-
-    ui32 outPos = startCharPos;
-    for (ui32 i = startCharPos; i < inputLength; i++) {
-        const char c = name.at(i);
-        if (c == '_') {
-            continue;
-        } else {
-            name[outPos] = AsciiToLower(c);
-            ++outPos;
-        }
-    }
-
-    name.resize(outPos);
-    Y_ABORT_UNLESS(inputLength - outPos == totalSkipped);
-
-    return Nothing();
-}
-
-TString NormalizeName(const TStringBuf& name) {
-    TString result(name);
-    TMaybe<TIssue> error = NormalizeName(TPosition(), result);
-    YQL_ENSURE(error.Empty(), "" << error->GetMessage());
-    return result;
-}
-
 void TDefaultTypeAnnotationVisitor::Visit(const TUnitExprType& type) {
     Y_UNUSED(type);
 }
diff --git a/yql/essentials/ast/yql_expr.h b/yql/essentials/ast/yql_expr.h
index 40cd7bb2d2b..722d3f6be51 100644
--- a/yql/essentials/ast/yql_expr.h
+++ b/yql/essentials/ast/yql_expr.h
@@ -9,6 +9,7 @@
 #include "yql_pos_handle.h"
 
 #include <yql/essentials/core/url_lister/interface/url_lister_manager.h>
+#include <yql/essentials/core/sql_types/normalize_name.h>
 #include <yql/essentials/utils/yql_panic.h>
 #include <yql/essentials/public/issue/yql_issue_manager.h>
 #include <yql/essentials/public/udf/udf_data_type.h>
@@ -2954,9 +2955,6 @@ const TTypeAnnotationNode& GetSeqItemType(const TTypeAnnotationNode& seq);
 
 const TTypeAnnotationNode& RemoveOptionality(const TTypeAnnotationNode& type);
 
-TMaybe<TIssue> NormalizeName(TPosition position, TString& name);
-TString NormalizeName(const TStringBuf& name);
-
 } // namespace NYql
 
 template<>
diff --git a/yql/essentials/core/sql_types/normalize_name.cpp b/yql/essentials/core/sql_types/normalize_name.cpp
new file mode 100644
index 00000000000..5b050503d36
--- /dev/null
+++ b/yql/essentials/core/sql_types/normalize_name.cpp
@@ -0,0 +1,63 @@
+#include "normalize_name.h"
+
+#include <yql/essentials/utils/yql_panic.h>
+
+#include <util/string/builder.h>
+#include <util/string/ascii.h>
+
+namespace NYql {
+
+    TMaybe<TIssue> NormalizeName(TPosition position, TString& name) {
+        const ui32 inputLength = name.length();
+        ui32 startCharPos = 0;
+        ui32 totalSkipped = 0;
+        bool atStart = true;
+        bool justSkippedUnderscore = false;
+        for (ui32 i = 0; i < inputLength; ++i) {
+            const char c = name.at(i);
+            if (c == '_') {
+                if (!atStart) {
+                    if (justSkippedUnderscore) {
+                        return TIssue(position, TStringBuilder() << "\"" << name << "\" looks weird, has multiple consecutive underscores");
+                    }
+                    justSkippedUnderscore = true;
+                    ++totalSkipped;
+                    continue;
+                } else {
+                    ++startCharPos;
+                }
+            } else {
+                atStart = false;
+                justSkippedUnderscore = false;
+            }
+        }
+
+        if (totalSkipped >= 5) {
+            return TIssue(position, TStringBuilder() << "\"" << name << "\" looks weird, has multiple consecutive underscores");
+        }
+
+        ui32 outPos = startCharPos;
+        for (ui32 i = startCharPos; i < inputLength; i++) {
+            const char c = name.at(i);
+            if (c == '_') {
+                continue;
+            } else {
+                name[outPos] = AsciiToLower(c);
+                ++outPos;
+            }
+        }
+
+        name.resize(outPos);
+        Y_ABORT_UNLESS(inputLength - outPos == totalSkipped);
+
+        return Nothing();
+    }
+
+    TString NormalizeName(const TStringBuf& name) {
+        TString result(name);
+        TMaybe<TIssue> error = NormalizeName(TPosition(), result);
+        YQL_ENSURE(error.Empty(), "" << error->GetMessage());
+        return result;
+    }
+
+} // namespace NYql
diff --git a/yql/essentials/core/sql_types/normalize_name.h b/yql/essentials/core/sql_types/normalize_name.h
new file mode 100644
index 00000000000..94d1e337680
--- /dev/null
+++ b/yql/essentials/core/sql_types/normalize_name.h
@@ -0,0 +1,13 @@
+#pragma once
+
+#include <yql/essentials/core/issue/yql_issue.h>
+
+#include <util/generic/string.h>
+#include <util/generic/fwd.h>
+
+namespace NYql {
+
+    TMaybe<TIssue> NormalizeName(TPosition position, TString& name);
+    TString NormalizeName(const TStringBuf& name);
+
+} // namespace NYql
diff --git a/yql/essentials/core/ut/yql_expr_type_annotation_ut.cpp b/yql/essentials/core/sql_types/normalize_name_ut.cpp
index 09ce7fd838b..ddd9b24d1e5 100644
--- a/yql/essentials/core/ut/yql_expr_type_annotation_ut.cpp
+++ b/yql/essentials/core/sql_types/normalize_name_ut.cpp
@@ -1,9 +1,8 @@
-#include "yql_expr_type_annotation.h"
+#include "normalize_name.h"
 
 #include <library/cpp/testing/unittest/registar.h>
 
-
-namespace NYql {
+using namespace NYql;
 
 Y_UNIT_TEST_SUITE(Misc) {
     Y_UNIT_TEST(NormalizeName) {
@@ -48,6 +47,4 @@ Y_UNIT_TEST_SUITE(Misc) {
         CheckIssues("a_B_c_d_e_f", "a_B_c_d_e_f");
         CheckIssues("_a_B_c_d_e_f", "_a_B_c_d_e_f");
     }
-}
-
-} // namespace NYql
+} // Y_UNIT_TEST_SUITE(Misc)
diff --git a/yql/essentials/core/sql_types/ut/ya.make b/yql/essentials/core/sql_types/ut/ya.make
index 11a1b7ddfb1..363e4dc4a44 100644
--- a/yql/essentials/core/sql_types/ut/ya.make
+++ b/yql/essentials/core/sql_types/ut/ya.make
@@ -1,6 +1,8 @@
 UNITTEST_FOR(yql/essentials/core/sql_types)
+
 SRCS(
     match_recognize_ut.cpp
+    normalize_name_ut.cpp
 )
 
 PEERDIR(
diff --git a/yql/essentials/core/sql_types/ya.make b/yql/essentials/core/sql_types/ya.make
index 88eb76fb511..fae32738bd3 100644
--- a/yql/essentials/core/sql_types/ya.make
+++ b/yql/essentials/core/sql_types/ya.make
@@ -4,13 +4,18 @@ SRCS(
     block.h
     match_recognize.h
     match_recognize.cpp
+    normalize_name.cpp
     simple_types.h
     simple_types.cpp
     yql_atom_enums.h
     yql_callable_names.h
 )
-GENERATE_ENUM_SERIALIZATION(match_recognize.h)
 
+PEERDIR(
+    yql/essentials/core/issue
+)
+
+GENERATE_ENUM_SERIALIZATION(match_recognize.h)
 GENERATE_ENUM_SERIALIZATION(yql_atom_enums.h)
 
 END()
diff --git a/yql/essentials/core/ut/ya.make b/yql/essentials/core/ut/ya.make
index e3a03ad340b..5fdab55b08e 100644
--- a/yql/essentials/core/ut/ya.make
+++ b/yql/essentials/core/ut/ya.make
@@ -4,7 +4,6 @@ SRCS(
     yql_column_order_ut.cpp
     yql_expr_constraint_ut.cpp
     yql_expr_optimize_ut.cpp
-    yql_expr_type_annotation_ut.cpp
     yql_library_compiler_ut.cpp
     yql_opt_utils_ut.cpp
     yql_udf_index_ut.cpp