From b1da09f141fe1c9af0aa6c7e3b4cc8b39b5c31ed Mon Sep 17 00:00:00 2001
From: "James R. Barlow" <james@purplerock.ca>
Date: Sun, 31 Jan 2021 02:21:03 -0800
Subject: [PATCH] Add plugin for setting logging console

So that we are not tied to tqdm.
---
 docs/plugins.rst                            |  2 ++
 src/ocrmypdf/__main__.py                    |  5 +++-
 src/ocrmypdf/_pipeline.py                   |  7 ++++--
 src/ocrmypdf/api.py                         | 26 ++++++++++++++-------
 src/ocrmypdf/builtin_plugins/concurrency.py |  6 +++++
 src/ocrmypdf/optimize.py                    |  2 +-
 src/ocrmypdf/pluginspec.py                  |  6 +++++
 7 files changed, 42 insertions(+), 12 deletions(-)

diff --git a/docs/plugins.rst b/docs/plugins.rst
index fb97f7f9..f1c3037c 100644
--- a/docs/plugins.rst
+++ b/docs/plugins.rst
@@ -136,6 +136,8 @@ Execution and progress reporting
 .. autoclass: ocrmypdf.pluginspec.Executor
     :members:
 
+.. autofunction:: ocrmypdf.pluginspec.get_logging_console
+
 .. autofunction:: ocrmypdf.pluginspec.get_executor
 
 .. autofunction:: ocrmypdf.pluginspec.get_progress_bar
diff --git a/src/ocrmypdf/__main__.py b/src/ocrmypdf/__main__.py
index 3c897b72..1046a50c 100755
--- a/src/ocrmypdf/__main__.py
+++ b/src/ocrmypdf/__main__.py
@@ -47,7 +47,10 @@ def run(args=None):
         verbosity = Verbosity.quiet
         options.progress_bar = False
     configure_logging(
-        verbosity, progress_bar_friendly=options.progress_bar, manage_root_logger=True
+        verbosity,
+        progress_bar_friendly=options.progress_bar,
+        manage_root_logger=True,
+        plugin_manager=plugin_manager,
     )
     log.debug('ocrmypdf %s', __version__)
     try:
diff --git a/src/ocrmypdf/_pipeline.py b/src/ocrmypdf/_pipeline.py
index d4c171d8..30ab9751 100644
--- a/src/ocrmypdf/_pipeline.py
+++ b/src/ocrmypdf/_pipeline.py
@@ -19,7 +19,6 @@ import img2pdf
 import pikepdf
 from pikepdf.models.metadata import encode_pdf_date
 from PIL import Image, ImageColor, ImageDraw
-from tqdm import tqdm
 
 from ocrmypdf import leptonica
 from ocrmypdf._concurrent import Executor
@@ -726,7 +725,11 @@ def convert_to_pdfa(input_pdf: Path, input_ps_stub: Path, context: PdfContext):
         output_file=output_file,
         compression=options.pdfa_image_compression,
         pdfa_part=options.output_type[-1],  # is pdfa-1, pdfa-2, or pdfa-3
-        progressbar_class=tqdm if options.progress_bar else None,
+        progressbar_class=(
+            context.plugin_manager.hook.get_progress_bar()
+            if options.progress_bar
+            else None
+        ),
     )
 
     return output_file
diff --git a/src/ocrmypdf/api.py b/src/ocrmypdf/api.py
index 9a37cad6..60de98fe 100644
--- a/src/ocrmypdf/api.py
+++ b/src/ocrmypdf/api.py
@@ -15,7 +15,10 @@ from pathlib import Path
 from typing import AnyStr, BinaryIO, Iterable, Optional, Union
 from warnings import warn
 
-from ocrmypdf._logging import PageNumberFilter, TqdmConsole
+from ocrmypdf._logging import (  # pylint: disable=unused-import
+    PageNumberFilter,
+    TqdmConsole,
+)
 from ocrmypdf._plugin_manager import get_plugin_manager
 from ocrmypdf._sync import run_pipeline
 from ocrmypdf._validation import check_options
@@ -47,6 +50,7 @@ def configure_logging(
     verbosity: Verbosity,
     progress_bar_friendly: bool = True,
     manage_root_logger: bool = False,
+    plugin_manager=None,
 ):
     """Set up logging.
 
@@ -74,12 +78,13 @@ def configure_logging(
     their own debug logging.
 
     Args:
-        verbosity (Verbosity): Verbosity level.
-        progress_bar_friendly (bool): Install the TqdmConsole log handler, which is
+        verbosity: Verbosity level.
+        progress_bar_friendly: Install the TqdmConsole log handler, which is
             compatible with the tqdm progress bar; without this log messages will
-            overwrite the progress bar
-        manage_root_logger (bool): Configure the process's root logger, to ensure
+            overwrite the progress bar.
+        manage_root_logger: Configure the process's root logger, to ensure
             all log output is sent through
+        plugin_manager: The plugin manager.
 
     Returns:
         The toplevel logger for ocrmypdf (or the root logger, if we are managing it).
@@ -90,8 +95,8 @@ def configure_logging(
     log = logging.getLogger(prefix)
     log.setLevel(logging.DEBUG)
 
-    if progress_bar_friendly:
-        console = logging.StreamHandler(stream=TqdmConsole(sys.stderr))
+    if plugin_manager and progress_bar_friendly:
+        console = plugin_manager.hook.get_logging_console()
     else:
         console = logging.StreamHandler(stream=sys.stderr)
 
@@ -245,6 +250,7 @@ def ocr(  # pylint: disable=unused-argument
     user_patterns: os.PathLike = None,
     fast_web_view: float = None,
     plugins: Iterable[StrPath] = None,
+    plugin_manager=None,
     keep_temporary_files: bool = None,
     progress_bar: bool = None,
     **kwargs,
@@ -296,6 +302,9 @@ def ocr(  # pylint: disable=unused-argument
     Returns:
         :class:`ocrmypdf.ExitCode`
     """
+    if plugins and plugin_manager:
+        raise ValueError("plugins= and plugin_manager are mutually exclusive")
+
     if not plugins:
         plugins = []
     elif isinstance(plugins, (str, Path)):
@@ -315,7 +324,8 @@ def ocr(  # pylint: disable=unused-argument
         # they might install different plugins, and generally speaking we have areas
         # of code that use global state.
 
-        plugin_manager = get_plugin_manager(plugins)
+        if not plugin_manager:
+            plugin_manager = get_plugin_manager(plugins)
         plugin_manager.hook.add_options(parser=parser)  # pylint: disable=no-member
 
         if 'verbose' in kwargs:
diff --git a/src/ocrmypdf/builtin_plugins/concurrency.py b/src/ocrmypdf/builtin_plugins/concurrency.py
index df6d42c6..5bfc05b3 100644
--- a/src/ocrmypdf/builtin_plugins/concurrency.py
+++ b/src/ocrmypdf/builtin_plugins/concurrency.py
@@ -27,6 +27,7 @@ from typing import Callable, Iterable, Optional, Union
 from tqdm import tqdm
 
 from ocrmypdf import Executor, hookimpl
+from ocrmypdf._logging import TqdmConsole
 from ocrmypdf.exceptions import InputFileError
 
 Queue = Union[multiprocessing.Queue, queue.Queue]
@@ -161,3 +162,8 @@ def get_executor():
 @hookimpl
 def get_progress_bar():
     return tqdm
+
+
+@hookimpl
+def get_logging_console():
+    return logging.StreamHandler(stream=TqdmConsole(sys.stderr))
diff --git a/src/ocrmypdf/optimize.py b/src/ocrmypdf/optimize.py
index 112a26c3..7acb9241 100644
--- a/src/ocrmypdf/optimize.py
+++ b/src/ocrmypdf/optimize.py
@@ -423,7 +423,7 @@ def transcode_jpegs(
         pbar.update()
 
     executor(
-        use_threads=True,
+        use_threads=True,  # Processes are significantly slower at this task
         max_workers=options.jobs,
         tqdm_kwargs=dict(
             desc="JPEGs",
diff --git a/src/ocrmypdf/pluginspec.py b/src/ocrmypdf/pluginspec.py
index cf68fd77..ee4e8dc4 100644
--- a/src/ocrmypdf/pluginspec.py
+++ b/src/ocrmypdf/pluginspec.py
@@ -8,6 +8,7 @@
 from abc import ABC, abstractmethod, abstractstaticmethod
 from argparse import ArgumentParser, Namespace
 from collections import namedtuple
+from logging import Handler
 from pathlib import Path
 from typing import TYPE_CHECKING, AbstractSet, Callable, Iterable, List, Optional
 
@@ -27,6 +28,11 @@ hookspec = pluggy.HookspecMarker('ocrmypdf')
 # pylint: disable=unused-argument
 
 
+@hookspec
+def get_logging_console() -> Handler:
+    """Returns a logging handler. Should be configured to handle progress bars."""
+
+
 @hookspec
 def add_options(parser: ArgumentParser) -> None:
     """Allows the plugin to add its own command line and API arguments.