Update analyze.py

analyze.py

@@ -11,7 +11,8 @@ Row = dict[str, Any]
 T = TypeVar("T")
 BATCH_SIZE = 1
 MAX_TEXT_LENGTH = 500
-batch_analyzer: Optional[BatchAnalyzerEngine] = None
+analyzer = AnalyzerEngine()
+batch_analyzer = BatchAnalyzerEngine(analyzer)
 
 
 class PresidioEntity(TypedDict):
@@ -121,16 +122,13 @@ def analyze(
 def presidio_scan_entities(
     rows: Iterable[Row], scanned_columns: list[str], columns_descriptions: list[str]
 ) -> Iterable[PresidioEntity]:
-    global batch_analyzer
     cache: dict[str, list[RecognizerResult]] = {}
-    if batch_analyzer is None:
-        batch_analyser = BatchAnalyzerEngine(AnalyzerEngine())
     rows_with_scanned_columns_only = (
         {column_name: get_strings(row[column_name])[:MAX_TEXT_LENGTH] for column_name in scanned_columns} for row in rows
     )
     for indices, batch in batched(rows_with_scanned_columns_only, BATCH_SIZE, with_indices=True):
         yield from analyze(
-            batch_analyzer=batch_analyser,
+            batch_analyzer=batch_analyzer,
             batch=batch,
             indices=indices,
             scanned_columns=scanned_columns,