update proto based on code review

Author: chenliu0831

pydeequ/v2/analyzers.py

@@ -250,16 +250,26 @@ def __repr__(self) -> str:
 
 @dataclass
 class Compliance(_ConnectAnalyzer):
-    """Fraction of rows satisfying a SQL predicate, named by `instance`."""
+    """
+    Fraction of rows satisfying a SQL predicate, named by `instance`.
+
+    `columns` should list every column name referenced in `predicate`. Deequ
+    uses it for an upfront precondition check so missing-column mistakes
+    surface as a clear DeequAnalysisException rather than a confusing
+    Spark SQL error during execution. Empty list disables the check.
+    """
 
     instance: str
     predicate: str
     where: Optional[str] = None
+    columns: Sequence[str] = ()
 
     def to_proto(self) -> proto.Analyzer:
         msg = proto.Analyzer()
         msg.compliance.instance = self.instance
         msg.compliance.predicate = self.predicate
+        if self.columns:
+            msg.compliance.columns.extend(self.columns)
         _set_where(msg, self.where)
         return msg
 

pydeequ/v2/proto/analysis_pb2.py

@@ -59,12 +59,14 @@ class HistogramSpec(_message.Message):
     def __init__(self, column: _Optional[str] = ..., max_detail_bins: _Optional[int] = ...) -> None: ...
 
 class ComplianceAnalyzerSpec(_message.Message):
-    __slots__ = ("instance", "predicate")
+    __slots__ = ("instance", "predicate", "columns")
     INSTANCE_FIELD_NUMBER: _ClassVar[int]
     PREDICATE_FIELD_NUMBER: _ClassVar[int]
+    COLUMNS_FIELD_NUMBER: _ClassVar[int]
     instance: str
     predicate: str
-    def __init__(self, instance: _Optional[str] = ..., predicate: _Optional[str] = ...) -> None: ...
+    columns: _containers.RepeatedScalarFieldContainer[str]
+    def __init__(self, instance: _Optional[str] = ..., predicate: _Optional[str] = ..., columns: _Optional[_Iterable[str]] = ...) -> None: ...
 
 class PatternMatchSpec(_message.Message):
     __slots__ = ("column", "pattern")