From 6d7ee987213990d610de423f7fc64d400f522cc3 Mon Sep 17 00:00:00 2001
From: "James R. Barlow" <jim@purplerock.ca>
Date: Tue, 13 Mar 2018 08:54:52 -0700
Subject: [PATCH] Force Tesseract 4 to be single threaded

Gives better performance (throughput basis) than the existing solution
and scales better on powerful boxes.
---
 ocrmypdf/__main__.py | 7 +++++++
 ocrmypdf/pipeline.py | 6 ------
 2 files changed, 7 insertions(+), 6 deletions(-)

diff --git a/ocrmypdf/__main__.py b/ocrmypdf/__main__.py
index 94412453..c3a006da 100755
--- a/ocrmypdf/__main__.py
+++ b/ocrmypdf/__main__.py
@@ -730,6 +730,13 @@ def run_pipeline():
     # options.input_file, options.pdf_renderer are already bound.)
     if not options.jobs:
         options.jobs = available_cpu_count()
+
+    # Performance is improved by setting Tesseract to single threaded. In tests
+    # this gives better throughput than letting a smaller number of Tesseract
+    # jobs run multithreaded.
+    if tesseract.v4():
+        os.environ.setdefault('OMP_THREAD_LIMIT', '1')
+
     try:
         work_folder = mkdtemp(prefix="com.github.ocrmypdf.")
         options.history_file = os.path.join(
diff --git a/ocrmypdf/pipeline.py b/ocrmypdf/pipeline.py
index afda0f34..6202f01b 100644
--- a/ocrmypdf/pipeline.py
+++ b/ocrmypdf/pipeline.py
@@ -1105,8 +1105,6 @@ def build_pipeline(options, work_folder, log, context):
         extras=[log, context])
     task_ocr_tesseract_hocr.graphviz(fillcolor='"#00cc66"')
     task_ocr_tesseract_hocr.active_if(options.pdf_renderer == 'hocr')
-    if tesseract.v4():
-        task_ocr_tesseract_hocr.jobs_limit(2)  # Uses multi-core on its own
 
     task_select_visible_page_image = main_pipeline.collate(
         task_func=select_visible_page_image,
@@ -1159,8 +1157,6 @@ def build_pipeline(options, work_folder, log, context):
         extras=[log, context])
     task_ocr_tesseract_textonly_pdf.graphviz(fillcolor='"#ff69b4"')
     task_ocr_tesseract_textonly_pdf.active_if(options.pdf_renderer == 'sandwich')
-    if tesseract.v4():
-        task_ocr_tesseract_textonly_pdf.jobs_limit(2)
 
     task_combine_layers = main_pipeline.collate(
         task_func=combine_layers,
@@ -1184,8 +1180,6 @@ def build_pipeline(options, work_folder, log, context):
         extras=[log, context])
     task_ocr_tesseract_and_render_pdf.graphviz(fillcolor='"#66ccff"')
     task_ocr_tesseract_and_render_pdf.active_if(options.pdf_renderer == 'tesseract')
-    if tesseract.v4():
-        task_ocr_tesseract_and_render_pdf.jobs_limit(2)  # Uses multi-core
 
     # PDF/A
     task_generate_postscript_stub = main_pipeline.transform(