Parallel cli bulk conversions via threading

Initial parallel document conversion: creates a pool of N threads
defined by the setting 'parallel_conversions'. Each thread calls
convert() on a document.

dangerzone/container.py

@@ -254,6 +254,27 @@ def convert(
     return success
 
 
+def get_max_parallel_conversions() -> int:
+    n_cpu = 1
+    if platform.system() == "Linux":
+        # if on linux containers run natively
+        cpu_count = os.cpu_count()
+        if cpu_count is not None:
+            n_cpu = cpu_count
+
+    elif get_runtime_name() == "docker":
+        # For Windows and MacOS containers run in VM
+        # So we obtain the CPU count for the VM
+        n_cpu_str = subprocess.check_output(
+            [get_runtime(), "info", "--format", "{{.NCPU}}"],
+            text=True,
+            startupinfo=get_subprocess_startupinfo(),
+        )
+        n_cpu = int(n_cpu_str.strip())
+
+    return 2 * n_cpu + 1
+
+
 # From global_common:
 
 # def validate_convert_to_pixel_output(self, common, output):

dangerzone/logic.py

@@ -1,3 +1,4 @@
+import concurrent.futures
 import gzip
 import json
 import logging
@@ -11,7 +12,7 @@ from typing import Callable, List, Optional
 import appdirs
 import colorama
 
-from .container import convert
+from . import container
 from .document import Document
 from .settings import Settings
 from .util import get_resource_path
@@ -49,10 +50,8 @@ class DangerzoneCore(object):
     def convert_documents(
         self, ocr_lang: Optional[str], stdout_callback: Callable[[str], None]
     ) -> None:
-        all_successful = True
-
-        for document in self.documents:
-            success = convert(
+        def convert_doc(document: Document) -> None:
+            success = container.convert(
                 document.input_filename,
                 document.output_filename,
                 ocr_lang,
@@ -63,6 +62,10 @@ class DangerzoneCore(object):
             else:
                 document.mark_as_failed()
 
+        max_jobs = container.get_max_parallel_conversions()
+        with concurrent.futures.ThreadPoolExecutor(max_workers=max_jobs) as executor:
+            executor.map(convert_doc, self.documents)
+
     def get_safe_documents(self) -> List[Document]:
         return [doc for doc in self.documents if doc.is_safe()]