Carry pre-onData state into the fresh runner

After making onData() return a fresh runner, calls made before onData
(addCheck, addAnalyzer, restrictToColumns, withLowCardinality,
addConstraintRules) silently dropped the configured state. Copy the
relevant lists/flags into the new instance so call order doesn't matter.

Author: chenliu0831

pydeequ/v2/profiles.py

@@ -95,16 +95,23 @@ def onData(
         table: Optional[str] = None,
         dataframe: "Optional[DataFrame]" = None,
     ) -> "ColumnProfilerRunner":
-        """Return a fresh runner bound to the given data (keyword-only)."""
+        """Return a fresh runner bound to the given data (keyword-only).
+
+        Configuration set before ``onData`` (``restrictToColumns``,
+        ``withLowCardinalityHistogramThreshold``) is carried into the new runner.
+        """
         if table is not None and dataframe is not None:
             raise ValueError("Provide either 'table' or 'dataframe', not both")
         if table is not None:
-            bound = self._engine.for_table(table)
+            engine = self._engine.for_table(table)
         elif dataframe is not None:
-            bound = self._engine.for_dataframe(dataframe)
+            engine = self._engine.for_dataframe(dataframe)
         else:
             raise ValueError("Must provide either 'table' or 'dataframe'")
-        return ColumnProfilerRunner(bound)
+        bound = ColumnProfilerRunner(engine)
+        bound._restrict_to_columns = self._restrict_to_columns
+        bound._low_cardinality_threshold = self._low_cardinality_threshold
+        return bound
 
     def restrictToColumns(self, columns: Sequence[str]) -> "ColumnProfilerRunner":
         self._restrict_to_columns = columns

pydeequ/v2/suggestions.py

@@ -91,16 +91,23 @@ def onData(
         table: Optional[str] = None,
         dataframe: "Optional[DataFrame]" = None,
     ) -> "ConstraintSuggestionRunner":
-        """Return a fresh runner bound to the given data (keyword-only)."""
+        """Return a fresh runner bound to the given data (keyword-only).
+
+        Rules and column restrictions set before ``onData`` are carried
+        into the new runner.
+        """
         if table is not None and dataframe is not None:
             raise ValueError("Provide either 'table' or 'dataframe', not both")
         if table is not None:
-            bound = self._engine.for_table(table)
+            engine = self._engine.for_table(table)
         elif dataframe is not None:
-            bound = self._engine.for_dataframe(dataframe)
+            engine = self._engine.for_dataframe(dataframe)
         else:
             raise ValueError("Must provide either 'table' or 'dataframe'")
-        return ConstraintSuggestionRunner(bound)
+        bound = ConstraintSuggestionRunner(engine)
+        bound._rules = list(self._rules)
+        bound._restrict_to_columns = self._restrict_to_columns
+        return bound
 
     def addConstraintRules(self, rules: Rules) -> "ConstraintSuggestionRunner":
         self._rules.append(rules)

pydeequ/v2/verification.py

@@ -105,10 +105,17 @@ def onData(
         table: Optional[str] = None,
         dataframe: "Optional[DataFrame]" = None,
     ) -> "VerificationSuite":
-        """Return a fresh suite bound to the given data (keyword-only)."""
-        return VerificationSuite(
+        """Return a fresh suite bound to the given data (keyword-only).
+
+        Checks added before ``onData`` are carried into the new suite, so
+        the call order ``Suite(engine).addCheck(...).onData(...).run()``
+        produces the same result as ``Suite(engine).onData(...).addCheck(...).run()``.
+        """
+        bound = VerificationSuite(
             _bind_engine(self._engine, table=table, dataframe=dataframe)
         )
+        bound._checks = list(self._checks)
+        return bound
 
     def addCheck(self, check: Check) -> "VerificationSuite":
         self._checks.append(check)
@@ -145,10 +152,15 @@ def onData(
         table: Optional[str] = None,
         dataframe: "Optional[DataFrame]" = None,
     ) -> "AnalysisRunner":
-        """Return a fresh runner bound to the given data (keyword-only)."""
-        return AnalysisRunner(
+        """Return a fresh runner bound to the given data (keyword-only).
+
+        Analyzers added before ``onData`` are carried into the new runner.
+        """
+        bound = AnalysisRunner(
             _bind_engine(self._engine, table=table, dataframe=dataframe)
         )
+        bound._analyzers = list(self._analyzers)
+        return bound
 
     def addAnalyzer(self, analyzer: _ConnectAnalyzer) -> "AnalysisRunner":
         self._analyzers.append(analyzer)