Merge pull request #408 from DigitalSlideArchive/limit-ocr-concurrency

The default thread pool concurrency is too much for ocr

.trivyignore

@@ -1,5 +1,12 @@
 # Accept these; revisit as needed
-CVE-2024-26597
-CVE-2024-26865
+CVE-2023-52433
+CVE-2024-26642
+CVE-2024-26643
+CVE-2024-26800
 CVE-2024-26828
-CVE-2024-26585
+CVE-2024-26865
+CVE-2024-26921
+CVE-2024-26923
+CVE-2024-26924
+CVE-2024-26925
+CVE-2024-27397

docs/INSTALL.rst

@@ -36,7 +36,7 @@ Hardware Requirements
 
 The recommended hardware is 32 GBytes of memory on a system with at least 4 cores.  If you are using OCR, it is highly recommended to have an NVidia Cuda-capable GPU.
 
-The minimum hardware is 4 GBytes of memory on a system with at least 2 cores (12 GBytes if using iSyntax files).
+The minimum hardware is 4 GBytes of memory on a system with at least 2 cores, 8 GBytes is using OCR, and 12 GBytes if using iSyntax files.
 
 The speed of your storage greatly affects the speed of the system.  SSD drives will for import, local storage, and export are recommended.
 

wsi_deid/jobs.py

@@ -1,4 +1,5 @@
 import concurrent.futures
+import os
 
 from girder import logger
 from girder.models.item import Item
@@ -227,7 +228,8 @@ def associate_unfiled_images(job):  # noqa
         # for itemId in itemIds:
         #     label_text = get_label_text_for_item(itemId, job)
         label_text_list = []
-        with concurrent.futures.ThreadPoolExecutor() as executor:
+        with concurrent.futures.ThreadPoolExecutor(
+                max_workers=max(1, os.cpu_count() // 2)) as executor:
             futures = []
             for itemId in itemIds:
                 futures.append(executor.submit(get_label_text_for_item, itemId, job))