Refactor logging record thread local storage

src/ocrmypdf/_pipelines/common.py

@@ -5,7 +5,6 @@ import json
 import logging
 import logging.handlers
 import os
-import threading
 from collections.abc import Sequence
 from dataclasses import dataclass
 from pathlib import Path
@@ -48,21 +47,17 @@ from ocrmypdf.pdfa import file_claims_pdfa
 
 log = logging.getLogger(__name__)
 
-tls = threading.local()
-tls.pageno = None
 
+def set_logging_tls(tls):
+    old_factory = logging.getLogRecordFactory()
 
-old_factory = logging.getLogRecordFactory()
+    def wrapper(*args, **kwargs):
+        record = old_factory(*args, **kwargs)
+        if hasattr(tls, 'pageno'):
+            record.pageno = tls.pageno
+        return record
 
-
-def record_factory(*args, **kwargs):
-    record = old_factory(*args, **kwargs)
-    if hasattr(tls, 'pageno'):
-        record.pageno = tls.pageno
-    return record
-
-
-logging.setLogRecordFactory(record_factory)
+    logging.setLogRecordFactory(wrapper)
 
 
 class PageResult(NamedTuple):

src/ocrmypdf/_pipelines/hocr_to_ocr_pdf.py

@@ -30,6 +30,7 @@ from ocrmypdf._pipelines.common import (
     HOCRResult,
     post_process,
     report_output_pdf,
+    set_logging_tls,
     setup_pipeline,
     worker_init,
 )
@@ -37,21 +38,11 @@ from ocrmypdf._plugin_manager import OcrmypdfPluginManager
 
 log = logging.getLogger(__name__)
 
+
 tls = threading.local()
 tls.pageno = None
 
-
-old_factory = logging.getLogRecordFactory()
-
-
-def record_factory(*args, **kwargs):
-    record = old_factory(*args, **kwargs)
-    if hasattr(tls, 'pageno'):
-        record.pageno = tls.pageno
-    return record
-
-
-logging.setLogRecordFactory(record_factory)
+set_logging_tls(tls)
 
 
 def exec_hocrtransform_sync(page_context: PageContext) -> HOCRResult:

src/ocrmypdf/_pipelines/pdf_to_hocr.py

@@ -27,6 +27,7 @@ from ocrmypdf._pipeline import (
 from ocrmypdf._pipelines.common import (
     HOCRResult,
     process_page,
+    set_logging_tls,
     setup_pipeline,
     worker_init,
 )
@@ -37,22 +38,10 @@ from ocrmypdf._validation import (
 
 log = logging.getLogger(__name__)
 
-
 tls = threading.local()
 tls.pageno = None
 
-
-old_factory = logging.getLogRecordFactory()
-
-
-def record_factory(*args, **kwargs):
-    record = old_factory(*args, **kwargs)
-    if hasattr(tls, 'pageno'):
-        record.pageno = tls.pageno
-    return record
-
-
-logging.setLogRecordFactory(record_factory)
+set_logging_tls(tls)
 
 
 def exec_page_hocr_sync(page_context: PageContext) -> HOCRResult:

src/ocrmypdf/_pipelines/standard.py

@@ -40,6 +40,7 @@ from ocrmypdf._pipelines.common import (
     post_process,
     process_page,
     report_output_pdf,
+    set_logging_tls,
     setup_pipeline,
     worker_init,
 )
@@ -59,18 +60,7 @@ log = logging.getLogger(__name__)
 tls = threading.local()
 tls.pageno = None
 
-
-old_factory = logging.getLogRecordFactory()
-
-
-def record_factory(*args, **kwargs):
-    record = old_factory(*args, **kwargs)
-    if hasattr(tls, 'pageno'):
-        record.pageno = tls.pageno
-    return record
-
-
-logging.setLogRecordFactory(record_factory)
+set_logging_tls(tls)
 
 
 def _image_to_ocr_text(