Update dependencies and improve configuration for better compatibility

Author: felix-schultz

.gitignore

@@ -9,6 +9,7 @@ lerna-debug.log*
 .venv
 __pycache__
 .cache
+cache
 
 node_modules
 dist

convert.py

@@ -11,6 +11,7 @@
 from typing import Annotated, Dict, List, Optional, Type
 
 import rich.table
+import torch
 import typer
 from docling_core.transforms.serializer.html import (
     HTMLDocSerializer,
@@ -83,6 +84,12 @@
 console = Console()
 err_console = Console(stderr=True)
 
+IS_MAC_ARM = sys.platform == "darwin" and platform.machine().lower() in {"arm64", "aarch64"}
+DEFAULT_ACCELERATOR_DEVICE = AcceleratorDevice.MPS if IS_MAC_ARM else AcceleratorDevice.AUTO
+
+print("MPS built:", torch.backends.mps.is_built())
+print("MPS available:", torch.backends.mps.is_available())
+
 ocr_factory_internal = get_ocr_factory(allow_external_plugins=False)
 ocr_engines_enum_internal = ocr_factory_internal.get_enum()
 
@@ -257,7 +264,7 @@ def convert_files(  # noqa: C901
     num_threads: Annotated[int, typer.Option(..., help="Number of threads")] = 4,
     device: Annotated[
         AcceleratorDevice, typer.Option(..., help="Accelerator device")
-    ] = AcceleratorDevice.AUTO,
+    ] = DEFAULT_ACCELERATOR_DEVICE,
     tempdir: Optional[str] = None
 ) -> List[ConversionResult]:
     log_format = "%(asctime)s\t%(levelname)s\t%(name)s: %(message)s"

flow-docling.spec

@@ -4,12 +4,14 @@ from PyInstaller.utils.hooks import collect_all
 datas = []
 binaries = []
 hiddenimports = []
-tmp_ret = collect_all('docling_parse')
-datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
 tmp_ret = collect_all('easyocr')
 datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
+tmp_ret = collect_all('transformers')
+datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
 tmp_ret = collect_all('docling')
 datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
+tmp_ret = collect_all('docling_parse')
+datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
 
 
 a = Analysis(
@@ -18,7 +20,7 @@ a = Analysis(
     binaries=binaries,
     datas=datas,
     hiddenimports=hiddenimports,
-    hookspath=['./hooks'],
+    hookspath=[],
     hooksconfig={},
     runtime_hooks=[],
     excludes=[],

main.py

@@ -18,7 +18,7 @@
 
 WEIGHTS_PATH = os.environ.get("HF_HOME", None)
 IMAGE_RESOLUTION_SCALE = 2.0
-LOCK_FILE = ".models_downloaded.lock"
+LOCK_FILE = ".models_downloaded.lock2"
 THREADS = None
 
 @asynccontextmanager

poetry.lock

@@ -10,8 +10,8 @@ authors = [
 
 
 [tool.poetry.dependencies]
-docling     = ">=2.40.0,<3.0.0"
-pyinstaller = ">=6.14.1,<7.0.0"
+docling     = ">=2.49.0,<3.0.0"
+pyinstaller = ">=6.15.0,<7.0.0"
 uvicorn     = ">=0.35.0,<0.40.0"
 fastapi     = ">=0.115.14,<0.120.0"
 einops      = ">=0.8.1,<1.0.0"