BabelDOC 0.3.16__tar.gz → 0.3.18__tar.gz

{babeldoc-0.3.16 → babeldoc-0.3.18}/.github/workflows/docs.yml

@@ -17,7 +17,7 @@ jobs:
           git config user.name github-actions[bot]
           git config user.email 41898282+github-actions[bot]@users.noreply.github.com
       - name: Setup uv with Python 3.12
-        uses: astral-sh/setup-uv@0c5e2b8115b80b4c7c5ddf6ffdd634974642d182 # v5.4.1
+        uses: astral-sh/setup-uv@d4b2f3b6ecc6e67c4457f6d3e41ec42d3d0fcb86 # v5.4.2
         with:
           python-version: "3.12"
           enable-cache: true

babeldoc-0.3.18/.github/workflows/lint.yml

@@ -0,0 +1,17 @@
+name: Lint Code
+permissions:
+  contents: read
+  pull-requests: write
+on: [push]
+
+jobs:
+  lint:
+    strategy:
+      fail-fast: false
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - name: Ruff
+        uses: astral-sh/ruff-action@v3
+      - name: AutoCorrect
+        uses: huacnlee/autocorrect-action@main

babeldoc-0.3.18/.github/workflows/pr-lint.yml

@@ -0,0 +1,46 @@
+name: Lint Code and Review Dog Report
+
+on: [pull_request]
+permissions:
+  contents: read
+  pull-requests: write
+jobs:
+  ruff:
+    name: runner / ruff
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      
+      - name: Install Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: '3.11'
+          
+      - name: Install ruff
+        run: pip install ruff
+        
+      - name: Install reviewdog
+        uses: reviewdog/action-setup@e04ffabe3898a0af8d0fb1af00c188831c4b5893 # v1.3.2
+        with:
+          reviewdog_version: latest
+          
+      - name: Run ruff with reviewdog
+        env:
+          REVIEWDOG_GITHUB_API_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        run: |
+          ruff check . --output-format=rdjson | reviewdog -f=rdjson -reporter=github-pr-review -fail-on-error
+          
+  autocorrect:
+    name: runner / autocorrect
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - name: AutoCorrect
+        uses: huacnlee/autocorrect-action@bf91ab3904c2908dd8e71312a8a83ed1eb632997 # v2.13.3
+      - name: Report ReviewDog
+        if: failure()
+        uses: huacnlee/autocorrect-action@bf91ab3904c2908dd8e71312a8a83ed1eb632997 # v2.13.3
+        env:
+          REVIEWDOG_GITHUB_API_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          reviewdog: true

{babeldoc-0.3.16 → babeldoc-0.3.18}/.github/workflows/publish-to-pypi.yml

@@ -35,7 +35,7 @@ jobs:
           token: ${{ secrets.GITHUB_TOKEN }}
           
       - name: Setup uv with Python 3.12
-        uses: astral-sh/setup-uv@0c5e2b8115b80b4c7c5ddf6ffdd634974642d182 # v5.4.1
+        uses: astral-sh/setup-uv@d4b2f3b6ecc6e67c4457f6d3e41ec42d3d0fcb86 # v5.4.2
         with:
           python-version: "3.12"
           enable-cache: true

{babeldoc-0.3.16 → babeldoc-0.3.18}/.github/workflows/test.yml

@@ -28,7 +28,7 @@ jobs:
           path: ~/.cache/babeldoc
           key: babeldoc-assets-${{ hashFiles('babeldoc/assets/embedding_assets_metadata.py') }}
       - name: Setup uv with Python ${{ matrix.python-version }}
-        uses: astral-sh/setup-uv@0c5e2b8115b80b4c7c5ddf6ffdd634974642d182 # v5.4.1
+        uses: astral-sh/setup-uv@d4b2f3b6ecc6e67c4457f6d3e41ec42d3d0fcb86 # v5.4.2
         with:
           python-version: ${{ matrix.python-version }}
           enable-cache: true

{babeldoc-0.3.16 → babeldoc-0.3.18}/.gitignore

@@ -12,6 +12,7 @@ web/dist
 web/dist-ssr
 web/*.local
 
+memray*
 **/*.so
 *.pdf
 *.docx

{babeldoc-0.3.16 → babeldoc-0.3.18}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: BabelDOC
-Version: 0.3.16
+Version: 0.3.18
 Summary: Yet Another Document Translator
 Project-URL: Homepage, https://github.com/funstory-ai/BabelDOC
 Project-URL: Issues, https://github.com/funstory-ai/BabelDOC/issues
@@ -191,6 +191,7 @@ uv run babeldoc --files example.pdf --files example2.pdf --openai --openai-model
 - `--no-watermark`: [DEPRECATED] Use --watermark-output-mode=no_watermark instead.
 - `--translate-table-text`: Translate table text (experimental, default: False)
 - `--skip-scanned-detection`: Skip scanned document detection (default: False). When using split translation, only the first part performs detection if not skipped.
+- `--ocr-workaround`: Use OCR workaround (default: False). When enabled, the tool will use OCR to detect text and fill background for scanned PDF.
 
 > [!TIP]
 > - Both `--skip-clean` and `--dual-translate-first` may help improve compatibility with some PDF readers
@@ -199,6 +200,7 @@ uv run babeldoc --files example.pdf --files example2.pdf --openai --openai-model
 > - If you encounter any compatibility issues, try using `--enhance-compatibility` first
 > - Use `--max-pages-per-part` for large documents to split them into smaller parts for translation and automatically merge them back.
 > - Use `--skip-scanned-detection` to speed up processing when you know your document is not a scanned PDF.
+> - Use `--ocr-workaround` to fill background for scanned PDF. (Current assumption: background is pure white, text is pure black, this option will also auto enable `--skip-scanned-detection`)
 
 ### Translation Service Options
 

{babeldoc-0.3.16 → babeldoc-0.3.18}/README.md

@@ -145,6 +145,7 @@ uv run babeldoc --files example.pdf --files example2.pdf --openai --openai-model
 - `--no-watermark`: [DEPRECATED] Use --watermark-output-mode=no_watermark instead.
 - `--translate-table-text`: Translate table text (experimental, default: False)
 - `--skip-scanned-detection`: Skip scanned document detection (default: False). When using split translation, only the first part performs detection if not skipped.
+- `--ocr-workaround`: Use OCR workaround (default: False). When enabled, the tool will use OCR to detect text and fill background for scanned PDF.
 
 > [!TIP]
 > - Both `--skip-clean` and `--dual-translate-first` may help improve compatibility with some PDF readers
@@ -153,6 +154,7 @@ uv run babeldoc --files example.pdf --files example2.pdf --openai --openai-model
 > - If you encounter any compatibility issues, try using `--enhance-compatibility` first
 > - Use `--max-pages-per-part` for large documents to split them into smaller parts for translation and automatically merge them back.
 > - Use `--skip-scanned-detection` to speed up processing when you know your document is not a scanned PDF.
+> - Use `--ocr-workaround` to fill background for scanned PDF. (Current assumption: background is pure white, text is pure black, this option will also auto enable `--skip-scanned-detection`)
 
 ### Translation Service Options
 

babeldoc-0.3.18/babeldoc/__init__.py

@@ -0,0 +1 @@
+__version__ = "0.3.18"

{babeldoc-0.3.16 → babeldoc-0.3.18}/babeldoc/const.py

@@ -3,7 +3,7 @@ import shutil
 import subprocess
 from pathlib import Path
 
-__version__ = "0.3.16"
+__version__ = "0.3.18"
 
 CACHE_FOLDER = Path.home() / ".cache" / "babeldoc"
 

{babeldoc-0.3.16 → babeldoc-0.3.18}/babeldoc/converter.py

@@ -92,7 +92,15 @@ class PDFConverterEx(PDFConverter):
         textwidth = font.char_width(cid)
         textdisp = font.char_disp(cid)
 
-        font_id = self.il_creater.current_page_font_name_id_map.get(font.xobj_id, None)
+        if not hasattr(font, "xobj_id"):
+            log.debug(
+                f"Font {font.fontname} does not have xobj_id attribute.",
+            )
+            font_id = "UNKNOW"
+        else:
+            font_id = self.il_creater.current_page_font_name_id_map.get(
+                font.xobj_id, None
+            )
 
         item = AWLTChar(
             matrix,

{babeldoc-0.3.16 → babeldoc-0.3.18}/babeldoc/document_il/backend/pdf_creater.py

@@ -31,11 +31,11 @@ def _subset_fonts_process(pdf_path, output_path):
         pdf = pymupdf.open(pdf_path)
         pdf.subset_fonts(fallback=False)
         pdf.save(output_path)
-        # 返回0表示成功
+        # 返回 0 表示成功
         os._exit(0)
     except Exception as e:
         logger.error(f"Error in font subsetting subprocess: {e}")
-        # 返回1表示失败
+        # 返回 1 表示失败
         os._exit(1)
 
 
@@ -69,11 +69,11 @@ def _save_pdf_clean_process(
             deflate_fonts=deflate_fonts,
             linear=linear,
         )
-        # 返回0表示成功
+        # 返回 0 表示成功
         os._exit(0)
     except Exception as e:
         logger.error(f"Error in save PDF with clean=True subprocess: {e}")
-        # 返回1表示失败
+        # 返回 1 表示失败
         os._exit(1)
 
 
@@ -185,21 +185,25 @@ class PDFCreater:
         except Exception:
             return set()
 
-    def _debug_render_rectangle(
+    def _render_rectangle(
         self,
         draw_op: BitStream,
         rectangle: il_version_1.PdfRectangle,
+        line_width: float = 1,
     ):
-        """Draw a debug rectangle in PDF for visualization purposes.
+        """Draw a rectangle in PDF for visualization purposes.
 
         Args:
             draw_op: BitStream to append PDF drawing operations
             rectangle: Rectangle object containing position information
+            line_width: Line width
         """
         x1 = rectangle.box.x
         y1 = rectangle.box.y
         x2 = rectangle.box.x2
         y2 = rectangle.box.y2
+        width = x2 - x1
+        height = y2 - y1
         # Save graphics state
         draw_op.append(b"q ")
 
@@ -207,17 +211,13 @@ class PDFCreater:
         draw_op.append(
             rectangle.graphic_state.passthrough_per_char_instruction.encode(),
         )  # Green stroke
-        draw_op.append(b" 1 w ")  # Line width
-
-        # Draw four lines manually
-        # Bottom line
-        draw_op.append(f"{x1} {y1} m {x2} {y1} l S ".encode())
-        # Right line
-        draw_op.append(f"{x2} {y1} m {x2} {y2} l S ".encode())
-        # Top line
-        draw_op.append(f"{x2} {y2} m {x1} {y2} l S ".encode())
-        # Left line
-        draw_op.append(f"{x1} {y2} m {x1} {y1} l S ".encode())
+        if line_width > 0:
+            draw_op.append(f" {line_width} w ".encode())  # Line width
+        draw_op.append(f"{x1} {y1} {width} {height} re ".encode())
+        if rectangle.fill_background:
+            draw_op.append(b" f ")
+        else:
+            draw_op.append(b" S ")
 
         # Restore graphics state
         draw_op.append(b"Q\n")
@@ -410,7 +410,7 @@ class PDFCreater:
             for rect in page.pdf_rectangle:
                 if not rect.debug_info:
                     continue
-                self._debug_render_rectangle(page_op, rect)
+                self._render_rectangle(page_op, rect)
             draw_op = page_op
             # Since this is a draw instruction container,
             # no additional information is needed
@@ -711,7 +711,17 @@ class PDFCreater:
                     # 然后添加段落中的字符
                     for paragraph in page.pdf_paragraph:
                         chars.extend(self.render_paragraph_to_char(paragraph))
-
+                    for rect in page.pdf_rectangle:
+                        if (
+                            translation_config.ocr_workaround
+                            and not rect.debug_info
+                            and rect.fill_background
+                        ):
+                            if rect.xobj_id in xobj_available_fonts:
+                                draw_op = xobj_draw_ops[rect.xobj_id]
+                            else:
+                                draw_op = page_op
+                            self._render_rectangle(draw_op, rect, line_width=0.1)
                     # 渲染所有字符
                     for char in chars:
                         if char.char_unicode == "\n":
@@ -773,7 +783,9 @@ class PDFCreater:
                             )
                         # pdf.update_stream(xobj.xref_id, b'')
                     for rect in page.pdf_rectangle:
-                        self._debug_render_rectangle(page_op, rect)
+                        if translation_config.debug and rect.debug_info:
+                            self._render_rectangle(page_op, rect)
+
                     draw_op = page_op
                     op_container = pdf.get_new_xref()
                     # Since this is a draw instruction container,
@@ -783,6 +795,9 @@ class PDFCreater:
                     pdf[page.page_number].set_contents(op_container)
                     pbar.advance()
             translation_config.raise_if_cancelled()
+            gc_level = 1
+            if self.translation_config.ocr_workaround:
+                gc_level = 4
             with self.translation_config.progress_monitor.stage_start(
                 SUBSET_FONT_STAGE_NAME,
                 1,
@@ -811,7 +826,7 @@ class PDFCreater:
                         pdf,
                         mono_out_path,
                         translation_config,
-                        garbage=1,
+                        garbage=gc_level,
                         deflate=True,
                         clean=not translation_config.skip_clean,
                         deflate_fonts=True,
@@ -864,7 +879,7 @@ class PDFCreater:
                         dual,
                         dual_out_path,
                         translation_config,
-                        garbage=1,
+                        garbage=gc_level,
                         deflate=True,
                         clean=not translation_config.skip_clean,
                         deflate_fonts=True,

{babeldoc-0.3.16 → babeldoc-0.3.18}/babeldoc/document_il/frontend/il_creater.py

@@ -19,6 +19,7 @@ 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 BLACK
 from babeldoc.document_il.utils.style_helper import YELLOW
 from babeldoc.translation_config import TranslationConfig
 
@@ -478,6 +479,7 @@ class ILCreater:
         self.current_page_font_name_id_map = {}
         self.current_page_font_char_bounding_box_map = {}
         self.mupdf_font_map: dict[int, pymupdf.Font] = {}
+        self.graphic_state_pool = {}
 
     def on_finish(self):
         self.progress.__exit__(None, None, None)
@@ -796,6 +798,21 @@ class ILCreater:
             f"{arg} {op}" for op, arg in gs.passthrough_instruction
         )
 
+        # 可能会影响部分 graphic state 准确度。不过 BabelDOC 仅使用 passthrough_per_char_instruction
+        # 所以应该是没啥影响
+        # 但是池化 graphic state 后可以减少内存占用
+        if (
+            graphic_state.passthrough_per_char_instruction
+            not in self.graphic_state_pool
+        ):
+            self.graphic_state_pool[graphic_state.passthrough_per_char_instruction] = (
+                graphic_state
+            )
+        else:
+            graphic_state = self.graphic_state_pool[
+                graphic_state.passthrough_per_char_instruction
+            ]
+
         return graphic_state
 
     def on_lt_char(self, char: LTChar):
@@ -887,6 +904,8 @@ class ILCreater:
             xobj_id=char.xobj_id,
             visual_bbox=visual_bbox,
         )
+        if self.translation_config.ocr_workaround:
+            pdf_char.pdf_style.graphic_state = BLACK
         if pdf_style.font_size == 0.0:
             logger.warning(
                 "Font size is 0.0 for character %s. Skip it.",

{babeldoc-0.3.16 → babeldoc-0.3.18}/babeldoc/document_il/il_version_1.py

@@ -364,6 +364,18 @@ class PdfRectangle:
             "type": "Attribute",
         },
     )
+    fill_background: bool | None = field(
+        default=None,
+        metadata={
+            "type": "Attribute",
+        },
+    )
+    xobj_id: int | None = field(
+        default=None,
+        metadata={
+            "type": "Attribute",
+        },
+    )
 
 
 @dataclass
@@ -756,6 +768,12 @@ class PdfParagraph:
             "type": "Attribute",
         },
     )
+    layout_id: int | None = field(
+        default=None,
+        metadata={
+            "type": "Attribute",
+        },
+    )
 
 
 @dataclass

{babeldoc-0.3.16 → babeldoc-0.3.18}/babeldoc/document_il/il_version_1.rnc

@@ -115,6 +115,7 @@ PDFParagraph =
     attribute FirstLineIndent { xsd:boolean }?,
     attribute debug_id { xsd:string }?,
     attribute layout_label { xsd:string }?,
+    attribute layout_id { xsd:int }?,
     Box,
     PDFStyle,
     PDFParagraphComposition*
@@ -148,5 +149,7 @@ PDFRectangle =
   element pdfRectangle {
     Box,
     GraphicState,
-    attribute debug_info { xsd:boolean }?
+    attribute debug_info { xsd:boolean }?,
+    attribute fill_background { xsd:boolean }?,
+    attribute xobjId { xsd:int }?
   }

{babeldoc-0.3.16 → babeldoc-0.3.18}/babeldoc/document_il/midend/paragraph_finder.py

@@ -8,9 +8,11 @@ from babeldoc.document_il import PdfCharacter
 from babeldoc.document_il import PdfLine
 from babeldoc.document_il import PdfParagraph
 from babeldoc.document_il import PdfParagraphComposition
+from babeldoc.document_il import PdfRectangle
 from babeldoc.document_il.utils.layout_helper import Layout
 from babeldoc.document_il.utils.layout_helper import add_space_dummy_chars
 from babeldoc.document_il.utils.layout_helper import get_char_unicode_string
+from babeldoc.document_il.utils.style_helper import WHITE
 from babeldoc.translation_config import TranslationConfig
 
 logger = logging.getLogger(__name__)
@@ -35,6 +37,41 @@ class ParagraphFinder:
     def __init__(self, translation_config: TranslationConfig):
         self.translation_config = translation_config
 
+    def add_text_fill_background(self, page: Page):
+        layout_map = {layout.id: layout for layout in page.page_layout}
+        for paragraph in page.pdf_paragraph:
+            layout_id = paragraph.layout_id
+            if layout_id is None:
+                continue
+            layout = layout_map[layout_id]
+            if paragraph.box is None:
+                continue
+            x1, y1, x2, y2 = (
+                paragraph.box.x,
+                paragraph.box.y,
+                paragraph.box.x2,
+                paragraph.box.y2,
+            )
+            layout_box = layout.box
+            if layout_box.x < x1:
+                x1 = layout_box.x
+            if layout_box.y < y1:
+                y1 = layout_box.y
+            if layout_box.x2 > x2:
+                x2 = layout_box.x2
+            if layout_box.y2 > y2:
+                y2 = layout_box.y2
+            assert x2 > x1 and y2 > y1
+            page.pdf_rectangle.append(
+                PdfRectangle(
+                    box=Box(x1, y1, x2, y2),
+                    fill_background=True,
+                    graphic_state=WHITE,
+                    debug_info=False,
+                    xobj_id=paragraph.xobj_id,
+                )
+            )
+
     def update_paragraph_data(self, paragraph: PdfParagraph, update_unicode=False):
         if not paragraph.pdf_paragraph_composition:
             return
@@ -126,6 +163,12 @@ class ParagraphFinder:
         for paragraph in paragraphs:
             self.update_paragraph_data(paragraph, update_unicode=True)
 
+        if self.translation_config.ocr_workaround:
+            self.add_text_fill_background(page)
+            # since this is ocr file,
+            # image characters are not needed
+            page.pdf_character = []
+
     def is_isolated_formula(self, char: PdfCharacter):
         return char.char_unicode in (
             "(cid:122)",
@@ -176,6 +219,7 @@ class ParagraphFinder:
                     if current_paragraph is None:
                         current_paragraph = PdfParagraph(
                             pdf_paragraph_composition=[line],
+                            layout_id=char_layout.id,
                             debug_id=generate_base58_id(),
                             layout_label=char_layout.name
                             if not current_layout
@@ -193,7 +237,7 @@ class ParagraphFinder:
             is_small_char = char_area < median_char_area * 0.1
 
             # 检查是否需要开始新段落
-            # 如果字符面积小于中位数面积的10%且当前段落已有字符，则跳过新段落检测
+            # 如果字符面积小于中位数面积的 10% 且当前段落已有字符，则跳过新段落检测
             if not (is_small_char and current_line_chars) and (
                 current_layout is None
                 or char_layout.id != current_layout.id
@@ -214,6 +258,7 @@ class ParagraphFinder:
                     else:
                         current_paragraph = PdfParagraph(
                             pdf_paragraph_composition=[line],
+                            layout_id=current_layout.id,
                             debug_id=generate_base58_id(),
                             layout_label=current_layout.name,
                         )
@@ -231,6 +276,7 @@ class ParagraphFinder:
             if current_paragraph is None:
                 current_paragraph = PdfParagraph(
                     pdf_paragraph_composition=[line],
+                    layout_id=current_layout.id,
                     debug_id=generate_base58_id(),
                     layout_label=current_layout.name,
                 )
@@ -420,6 +466,7 @@ class ParagraphFinder:
                         unicode="",
                         debug_id=generate_base58_id(),
                         layout_label=paragraph.layout_label,
+                        layout_id=paragraph.layout_id,
                     )
                     # 更新原段落
                     paragraph.pdf_paragraph_composition = (
@@ -449,6 +496,7 @@ class ParagraphFinder:
                         unicode="",
                         debug_id=generate_base58_id(),
                         layout_label=paragraph.layout_label,
+                        layout_id=paragraph.layout_id,
                     )
                     # 更新原段落
                     paragraph.pdf_paragraph_composition = (

{babeldoc-0.3.16 → babeldoc-0.3.18}/babeldoc/document_il/midend/styles_and_formulas.py

@@ -395,7 +395,7 @@ class StylesAndFormulas:
                             right_line = comp.pdf_line
                             break
 
-                # 计算与左右文本的y轴交集
+                # 计算与左右文本的 y 轴交集
                 left_intersection = 0
                 right_intersection = 0
 
@@ -413,7 +413,7 @@ class StylesAndFormulas:
                     if intersection_end > intersection_start:
                         right_intersection = intersection_end - intersection_start
 
-                # 如果有两个文本段落，将交集较小的设为none
+                # 如果有两个文本段落，将交集较小的设为 none
                 if left_line and right_line:
                     if left_intersection < right_intersection:
                         left_line = None
@@ -556,6 +556,9 @@ class StylesAndFormulas:
             r"|.*DJ5Exchange.*"
             r"|.*Times-Italic.*"
             r"|.*Times-BoldItalic.*"
+            r"|.*CronosPro.*"
+            r"|.*ACaslon.*"
+            r"|.*Frutiger.*"
             r")$"
         )
         if self.translation_config.formular_font_pattern:

{babeldoc-0.3.16 → babeldoc-0.3.18}/babeldoc/document_il/midend/typesetting.py

@@ -221,8 +221,8 @@ class TypesettingUnit:
                 # 连接号
                 "～",  # 全角波浪号
                 "-",  # 连字符减号
-                "–",  # 短破折号(EN DASH)
-                "—",  # 长破折号(EM DASH)
+                "–",  # 短破折号 (EN DASH)
+                "—",  # 长破折号 (EM DASH)
                 # 间隔号
                 "·",  # 中间点
                 "・",  # 片假名中间点
@@ -767,7 +767,7 @@ class Typesetting:
         line_spacing = 1.7  # 初始行距为 1.7
         min_scale = 0.1  # 最小缩放因子
         min_line_spacing = 1.4  # 最小行距
-        expand_space_flag = 0  # 0: 未扩展, 1: 已向下扩展, 2: 已向右扩展
+        expand_space_flag = 0  # 0: 未扩展，1: 已向下扩展，2: 已向右扩展
 
         while scale >= min_scale:
             # 尝试布局排版单元

{babeldoc-0.3.16 → babeldoc-0.3.18}/babeldoc/document_il/utils/style_helper.py

@@ -25,6 +25,10 @@ def create_pdf_style(r, g, b, font_id="china-ss", font_size=6):
     )
 
 
+BLACK = il_version_1.GraphicState(passthrough_per_char_instruction="0 g 0 G")
+
+WHITE = il_version_1.GraphicState(passthrough_per_char_instruction="1 g 1 G")
+
 # Generate all color styles
 RED = il_version_1.GraphicState(
     passthrough_per_char_instruction="1.0000000000 0.2313725490 0.1882352941 rg "

{babeldoc-0.3.16 → babeldoc-0.3.18}/babeldoc/docvision/rpc_doclayout.py

@@ -43,10 +43,10 @@ def encode_image(image) -> bytes:
 
 
 @retry(
-    stop=stop_after_attempt(3),  # 最多重试3次
+    stop=stop_after_attempt(3),  # 最多重试 3 次
     wait=wait_exponential(
         multiplier=1, min=1, max=10
-    ),  # 指数退避策略，初始1秒，最大10秒
+    ),  # 指数退避策略，初始 1 秒，最大 10 秒
     retry=retry_if_exception_type((httpx.HTTPError, Exception)),  # 针对哪些异常重试
     before_sleep=lambda retry_state: logger.warning(
         f"Request failed, retrying in {retry_state.next_action.sleep} seconds... "

{babeldoc-0.3.16 → babeldoc-0.3.18}/babeldoc/high_level.py

@@ -586,6 +586,7 @@ def _do_translate_single(
     logger.debug(f"finish parse il from {temp_pdf_path}")
     docs = il_creater.create_il()
     logger.debug(f"finish create il from {temp_pdf_path}")
+    del il_creater
     if translation_config.debug:
         xml_converter.write_json(
             docs,
@@ -657,6 +658,7 @@ def _do_translate_single(
         il_translator = ILTranslator(translate_engine, translation_config)
 
     il_translator.translate(docs)
+    del il_translator
     logger.debug(f"finish ILTranslator from {temp_pdf_path}")
     if translation_config.debug:
         xml_converter.write_json(

{babeldoc-0.3.16 → babeldoc-0.3.18}/babeldoc/main.py

@@ -23,7 +23,7 @@ from babeldoc.translation_config import TranslationConfig
 from babeldoc.translation_config import WatermarkOutputMode
 
 logger = logging.getLogger(__name__)
-__version__ = "0.3.16"
+__version__ = "0.3.18"
 
 
 def create_parser():
@@ -210,6 +210,12 @@ def create_parser():
         default=False,
         help="Skip scanned document detection (speeds up processing for non-scanned documents)",
     )
+    translation_group.add_argument(
+        "--ocr-workaround",
+        action="store_true",
+        default=False,
+        help="Add text fill background (experimental)",
+    )
     # service option argument group
     service_group = translation_group.add_mutually_exclusive_group()
     service_group.add_argument(
@@ -377,6 +383,7 @@ async def main():
             table_model=table_model,
             show_char_box=args.show_char_box,
             skip_scanned_detection=args.skip_scanned_detection,
+            ocr_workaround=args.ocr_workaround,
         )
 
         # Create progress handler