Merge branch 'main' into speed_improvements

Author: vladd-bit

ocr_service/app/app.py

@@ -4,7 +4,7 @@
 import subprocess
 import time
 from threading import Event, Thread
-from typing import Any
+from typing import Any, cast
 
 import psutil
 from fastapi import FastAPI
@@ -30,19 +30,23 @@ def start_office_server(port_num: str) -> dict[str, Any]:
     uno_port = str(int(port_num) + 1000)  # e.g. XML-RPC 9900, UNO 10900
     user_installation = f"{settings.TMP_FILE_DIR}/lo_profile_{port_num}"
 
+    lo_python = cast(str, settings.LIBRE_OFFICE_PYTHON_PATH)
+    lo_exec = cast(str, settings.LIBRE_OFFICE_EXEC_PATH)
+
     loffice_process: dict[str, Any] = {
         "process": subprocess.Popen(
             args=[
-                settings.LIBRE_OFFICE_PYTHON_PATH,
+                lo_python,
                 "-m",
                 "unoserver.server",
                 "--interface", settings.LIBRE_OFFICE_NETWORK_INTERFACE,
                 "--uno-interface", settings.LIBRE_OFFICE_NETWORK_INTERFACE,
-                "--executable", settings.LIBRE_OFFICE_EXEC_PATH,
+                "--executable", lo_exec,
                 "--port", port_num,
                 "--uno-port", uno_port,
-                "--user-installation", user_installation
-            ],
+                "--user-installation", user_installation,
+               # "--logfile", f"loffice_{port_num}.log"
+            ], # type: ignore
             cwd=settings.TMP_FILE_DIR,
             close_fds=True,
             shell=False

ocr_service/processor/converter.py

@@ -9,7 +9,7 @@
 from io import BytesIO
 from subprocess import PIPE, Popen
 from threading import Timer
-from typing import Any
+from typing import Any, cast
 
 import pypdfium2 as pdfium
 from bs4 import BeautifulSoup
@@ -181,9 +181,10 @@ def _preprocess_doc(self, stream: bytes, file_name: str) -> bytes:
             for port_num, loffice_process in self.loffice_process_list.items():
                 if loffice_process["used"] is False:
                     used_port_num = str(port_num)
+                    lo_python = cast(str, settings.LIBRE_OFFICE_PYTHON_PATH)
                     converter_bootstrap = "from unoserver.client import converter_main; converter_main()"
                     _args = [
-                        settings.LIBRE_OFFICE_PYTHON_PATH,
+                        lo_python,
                         "-c",
                         converter_bootstrap,
                         doc_file_path,

ocr_service/settings.py

@@ -1,3 +1,5 @@
+# mypy: disable-error-code=prop-decorator
+
 import ast
 import logging
 import multiprocessing
@@ -210,4 +212,4 @@ def LIBRE_OFFICE_LISTENER_PORT_RANGE(self) -> range:
             return range(start, end)
         return range(self.DEFAULT_LIBRE_OFFICE_SERVER_PORT, self.LIBRE_OFFICE_PORT_CAP)
 
-settings = Settings() # type: ignore[call-arg]
+settings = Settings() # pyright: ignore[reportCallIssue]