chore: implement StrExtractOp and StrRepeatOp

Author: chelsea-lin

bigframes/core/compile/sqlglot/expressions/unary_compiler.py

@@ -182,6 +182,13 @@ def _(op: ops.StrContainsRegexOp, expr: TypedExpr) -> sge.Expression:
     return sge.RegexpLike(this=expr.expr, expression=sge.convert(op.pat))
 
 
+@UNARY_OP_REGISTRATION.register(ops.StrExtractOp)
+def _(op: ops.StrExtractOp, expr: TypedExpr) -> sge.Expression:
+    return sge.RegexpExtract(
+        this=expr.expr, expression=sge.convert(op.pat), group=sge.convert(op.n)
+    )
+
+
 @UNARY_OP_REGISTRATION.register(ops.StrFindOp)
 def _(op: ops.StrFindOp, expr: TypedExpr) -> sge.Expression:
     # INSTR is 1-based, so we need to adjust the start position.
@@ -211,6 +218,10 @@ def _(op: ops.StrContainsOp, expr: TypedExpr) -> sge.Expression:
     return sge.Like(this=expr.expr, expression=sge.convert(f"%{op.pat}%"))
 
 
+@UNARY_OP_REGISTRATION.register(ops.StrRepeatOp)
+def _(op: ops.StrRepeatOp, expr: TypedExpr) -> sge.Expression:
+    return sge.Repeat(this=expr.expr, times=sge.convert(op.repeats))
+
 @UNARY_OP_REGISTRATION.register(ops.date_op)
 def _(op: ops.base_ops.UnaryOp, expr: TypedExpr) -> sge.Expression:
     return sge.Date(this=expr.expr)

tests/unit/core/compile/sqlglot/expressions/snapshots/test_unary_compiler/test_str_extract/out.sql

@@ -0,0 +1,13 @@
+WITH `bfcte_0` AS (
+  SELECT
+    `string_col` AS `bfcol_0`
+  FROM `bigframes-dev`.`sqlglot_test`.`scalar_types`
+), `bfcte_1` AS (
+  SELECT
+    *,
+    REGEXP_EXTRACT(`bfcol_0`, '([a-z]*)') AS `bfcol_1`
+  FROM `bfcte_0`
+)
+SELECT
+  `bfcol_1` AS `string_col`
+FROM `bfcte_1`

tests/unit/core/compile/sqlglot/expressions/snapshots/test_unary_compiler/test_str_repeat/out.sql

@@ -0,0 +1,13 @@
+WITH `bfcte_0` AS (
+  SELECT
+    `string_col` AS `bfcol_0`
+  FROM `bigframes-dev`.`sqlglot_test`.`scalar_types`
+), `bfcte_1` AS (
+  SELECT
+    *,
+    REPEAT(`bfcol_0`, 2) AS `bfcol_1`
+  FROM `bfcte_0`
+)
+SELECT
+  `bfcol_1` AS `string_col`
+FROM `bfcte_1`

tests/unit/core/compile/sqlglot/expressions/test_unary_compiler.py

@@ -524,6 +524,19 @@ def test_str_contains_regex(scalar_types_df: bpd.DataFrame, snapshot):
     snapshot.assert_match(sql, "out.sql")
 
 
+def test_str_extract(scalar_types_df: bpd.DataFrame, snapshot):
+    bf_df = scalar_types_df[["string_col"]]
+    sql = _apply_unary_op(bf_df, ops.StrExtractOp(r"([a-z]*)", 1), "string_col")
+
+    snapshot.assert_match(sql, "out.sql")
+
+
+def test_str_repeat(scalar_types_df: bpd.DataFrame, snapshot):
+    bf_df = scalar_types_df[["string_col"]]
+    sql = _apply_unary_op(bf_df, ops.StrRepeatOp(2), "string_col")
+    snapshot.assert_match(sql, "out.sql")
+
+
 def test_str_find(scalar_types_df: bpd.DataFrame, snapshot):
     bf_df = scalar_types_df[["string_col"]]
     sql = _apply_unary_op(bf_df, ops.StrFindOp("e", start=None, end=None), "string_col")