BabelDOC 0.2.17__tar.gz → 0.2.19__tar.gz

{babeldoc-0.2.17 → babeldoc-0.2.19}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: BabelDOC
-Version: 0.2.17
+Version: 0.2.19
 Summary: Yet Another Document Translator
 Project-URL: Homepage, https://github.com/funstory-ai/BabelDOC
 Project-URL: Issues, https://github.com/funstory-ai/BabelDOC/issues

babeldoc-0.2.19/babeldoc/__init__.py

@@ -0,0 +1 @@
+__version__ = "0.2.19"

{babeldoc-0.2.17 → babeldoc-0.2.19}/babeldoc/assets/assets.py

@@ -328,8 +328,11 @@ async def async_warmup():
     _ = encoding_for_model("gpt-4o")
     async with httpx.AsyncClient() as client:
         onnx_task = asyncio.create_task(get_doclayout_onnx_model_path_async(client))
+        onnx_task2 = asyncio.create_task(
+            get_table_detection_rapidocr_model_path_async(client)
+        )
         font_tasks = asyncio.create_task(download_all_fonts_async(client))
-        await asyncio.gather(onnx_task, font_tasks)
+        await asyncio.gather(onnx_task, onnx_task2, font_tasks)
 
 
 def warmup():
@@ -360,6 +363,8 @@ def generate_all_assets_file_list():
             "name": "doclayout_yolo_docstructbench_imgsz1024.onnx",
             "sha3_256": DOCLAYOUT_YOLO_DOCSTRUCTBENCH_IMGSZ1024ONNX_SHA3_256,
         },
+    )
+    result["models"].append(
         {
             "name": "ch_PP-OCRv4_det_infer.onnx",
             "sha3_256": TABLE_DETECTION_RAPIDOCR_MODEL_SHA3_256,

{babeldoc-0.2.17 → babeldoc-0.2.19}/babeldoc/const.py

@@ -3,7 +3,7 @@ import shutil
 import subprocess
 from pathlib import Path
 
-__version__ = "0.2.17"
+__version__ = "0.2.19"
 
 CACHE_FOLDER = Path.home() / ".cache" / "babeldoc"
 

{babeldoc-0.2.17 → babeldoc-0.2.19}/babeldoc/document_il/frontend/il_creater.py

@@ -18,6 +18,7 @@ from pdfminer.pdftypes import resolve1 as pdftypes_resolve1
 from pdfminer.psparser import PSLiteral
 
 from babeldoc.document_il import il_version_1
+from babeldoc.document_il.utils.style_helper import YELLOW
 from babeldoc.translation_config import TranslationConfig
 
 
@@ -568,6 +569,15 @@ class ILCreater:
 
         self.current_page.pdf_character.append(pdf_char)
 
+        if self.translation_config.show_char_box:
+            self.current_page.pdf_rectangle.append(
+                il_version_1.PdfRectangle(
+                    box=pdf_char.visual_bbox.box,
+                    graphic_state=YELLOW,
+                    debug_info=True,
+                )
+            )
+
     def create_il(self):
         pages = [
             page

{babeldoc-0.2.17 → babeldoc-0.2.19}/babeldoc/document_il/midend/styles_and_formulas.py

@@ -531,6 +531,8 @@ class StylesAndFormulas:
             r"|GlosaMath.+"
             r"|URWPalladioL.+"
             r"|CMSS.+"
+            r"|Arial.*"
+            r"|TimesNewRoman.*"
             r")$"
         )
         if self.translation_config.formular_font_pattern:

{babeldoc-0.2.17 → babeldoc-0.2.19}/babeldoc/high_level.py

@@ -504,6 +504,7 @@ def do_translate(
         logger.info(
             f"finish translate: {original_pdf_path}, cost: {finish_time - start_time} s",
         )
+        result.original_pdf_path = translation_config.input_file
         pm.translate_done(result)
         return result
 
@@ -661,21 +662,28 @@ def _do_translate_single(
 
     pdf_creater = PDFCreater(temp_pdf_path, docs, translation_config)
     result = pdf_creater.write(translation_config)
+    try:
+        if mono_watermark_first_page_doc_bytes:
+            mono_watermark_pdf = merge_watermark_doc(
+                result.mono_pdf_path,
+                mono_watermark_first_page_doc_bytes,
+                translation_config,
+            )
+            result.mono_pdf_path = mono_watermark_pdf
+    except Exception:
+        result.mono_pdf_path = result.no_watermark_mono_pdf_path
+    try:
+        if dual_watermark_first_page_doc_bytes:
+            dual_watermark_pdf = merge_watermark_doc(
+                result.dual_pdf_path,
+                dual_watermark_first_page_doc_bytes,
+                translation_config,
+            )
+            result.dual_pdf_path = dual_watermark_pdf
+    except Exception:
+        result.dual_pdf_path = result.no_watermark_dual_pdf_path
 
-    if mono_watermark_first_page_doc_bytes:
-        mono_watermark_pdf = merge_watermark_doc(
-            result.mono_pdf_path,
-            mono_watermark_first_page_doc_bytes,
-            translation_config,
-        )
-        result.mono_pdf_path = mono_watermark_pdf
-    if dual_watermark_first_page_doc_bytes:
-        dual_watermark_pdf = merge_watermark_doc(
-            result.dual_pdf_path,
-            dual_watermark_first_page_doc_bytes,
-            translation_config,
-        )
-        result.dual_pdf_path = dual_watermark_pdf
+    result.original_pdf_path = translation_config.input_file
 
     return result
 

{babeldoc-0.2.17 → babeldoc-0.2.19}/babeldoc/main.py

@@ -23,7 +23,7 @@ from babeldoc.translation_config import TranslationConfig
 from babeldoc.translation_config import WatermarkOutputMode
 
 logger = logging.getLogger(__name__)
-__version__ = "0.2.17"
+__version__ = "0.2.19"
 
 
 def create_parser():
@@ -198,6 +198,12 @@ def create_parser():
         default=False,
         help="Translate table text (experimental)",
     )
+    translation_group.add_argument(
+        "--show-char-box",
+        action="store_true",
+        default=False,
+        help="Show character box (debug only)",
+    )
     # service option argument group
     service_group = translation_group.add_mutually_exclusive_group()
     service_group.add_argument(
@@ -363,6 +369,7 @@ async def main():
             watermark_output_mode=watermark_output_mode,
             split_strategy=split_strategy,
             table_model=table_model,
+            show_char_box=args.show_char_box,
         )
 
         # Create progress handler

{babeldoc-0.2.17 → babeldoc-0.2.19}/babeldoc/translation_config.py

@@ -66,6 +66,7 @@ class TranslationConfig:
         # Add split-related parameters
         split_strategy: BaseSplitStrategy | None = None,
         table_model=None,
+        show_char_box: bool = False,
     ):
         self.translator = translator
 
@@ -144,6 +145,7 @@ class TranslationConfig:
         self._part_output_dirs: dict[int, Path] = {}
 
         self.table_model = table_model
+        self.show_char_box = show_char_box
 
     def parse_pages(self, pages_str: str | None) -> list[tuple[int, int]] | None:
         """解析页码字符串，返回页码范围列表

{babeldoc-0.2.17 → babeldoc-0.2.19}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "BabelDOC"
-version = "0.2.17"
+version = "0.2.19"
 description = "Yet Another Document Translator"
 license = "AGPL-3.0"
 readme = "README.md"
@@ -148,7 +148,7 @@ pythonpath = [".", "src"]
 testpaths = ["tests"]
 
 [bumpver]
-current_version = "0.2.17"
+current_version = "0.2.19"
 version_pattern = "MAJOR.MINOR.PATCH[.PYTAGNUM]"
 
 [bumpver.file_patterns]

{babeldoc-0.2.17 → babeldoc-0.2.19}/uv.lock

@@ -48,7 +48,7 @@ wheels = [
 
 [[package]]
 name = "babeldoc"
-version = "0.2.17"
+version = "0.2.19"
 source = { editable = "." }
 dependencies = [
     { name = "bitstring" },

babeldoc-0.2.17/babeldoc/__init__.py

@@ -1 +0,0 @@
-__version__ = "0.2.17"