Merge pull request #114 from funstory-ai/feat/dual-pdf-options

feat: add dual PDF generation options (side-by-side and alternating pages)

README.md

@@ -110,6 +110,8 @@ uv run babeldoc --bing --files example.pdf --files example2.pdf
 - `--dual-translate-first`: Put translated pages first in dual PDF mode (default: original pages first)
 - `--disable-rich-text-translate`: Disable rich text translation (may help improve compatibility with some PDFs)
 - `--enhance-compatibility`: Enable all compatibility enhancement options (equivalent to --skip-clean --dual-translate-first --disable-rich-text-translate)
+- `--use-side-by-side-dual`: Use side-by-side mode for dual PDF (default). Original and translated pages are shown side by side on the same page.
+- `--use-alternating-pages-dual`: Use alternating pages mode for dual PDF. Original and translated pages are arranged in alternate order.
 
 > [!TIP]
 > - Both `--skip-clean` and `--dual-translate-first` may help improve compatibility with some PDF readers

babeldoc/__init__.py

@@ -1 +1 @@
-__version__ = "0.1.11"
+__version__ = "0.1.12"

babeldoc/const.py

@@ -2,7 +2,7 @@
 import subprocess
 from pathlib import Path
 
-__version__ = "0.1.11"
+__version__ = "0.1.12"
 
 CACHE_FOLDER = Path.home() / ".cache" / "babeldoc"
 

babeldoc/document_il/backend/pdf_creater.py

@@ -156,6 +156,100 @@ def _debug_render_rectangle(
         # Restore graphics state
         draw_op.append(b"Q\n")
 
+    def create_side_by_side_dual_pdf(
+        self,
+        original_pdf: pymupdf.Document,
+        translated_pdf: pymupdf.Document,
+        dual_out_path: str,
+        translation_config: TranslationConfig,
+    ) -> pymupdf.Document:
+        """Create a dual PDF with side-by-side pages (original and translation).
+
+        Args:
+            original_pdf: Original PDF document
+            translated_pdf: Translated PDF document
+            dual_out_path: Output path for the dual PDF
+            translation_config: Translation configuration
+
+        Returns:
+            The created dual PDF document
+        """
+        # Create a new PDF for side-by-side pages
+        dual = pymupdf.open()
+        page_count = min(original_pdf.page_count, translated_pdf.page_count)
+
+        for page_id in range(page_count):
+            # Get pages from both PDFs
+            orig_page = original_pdf[page_id]
+            trans_page = translated_pdf[page_id]
+
+            # Calculate total width and use max height
+            total_width = orig_page.rect.width + trans_page.rect.width
+            max_height = max(orig_page.rect.height, trans_page.rect.height)
+
+            # Create new page with combined width
+            dual_page = dual.new_page(width=total_width, height=max_height)
+
+            # Define rectangles for left and right sides
+            left_width = (
+                orig_page.rect.width
+                if not translation_config.dual_translate_first
+                else trans_page.rect.width
+            )
+            rect_left = pymupdf.Rect(0, 0, left_width, max_height)
+            rect_right = pymupdf.Rect(left_width, 0, total_width, max_height)
+
+            # Show pages according to dual_translate_first setting
+            if translation_config.dual_translate_first:
+                # Show translated page on left and original on right
+                rect_left, rect_right = rect_right, rect_left
+
+            # Show original page on left and translated on right (default)
+            dual_page.show_pdf_page(
+                rect_left,
+                original_pdf,
+                page_id,
+                keep_proportion=True,
+            )
+            dual_page.show_pdf_page(
+                rect_right,
+                translated_pdf,
+                page_id,
+                keep_proportion=True,
+            )
+
+        return dual
+
+    def create_alternating_pages_dual_pdf(
+        self,
+        original_pdf_path: str,
+        translated_pdf: pymupdf.Document,
+        translation_config: TranslationConfig,
+    ) -> pymupdf.Document:
+        """Create a dual PDF with alternating pages (original and translation).
+
+        Args:
+            original_pdf_path: Path to the original PDF
+            translated_pdf: Translated PDF document
+            translation_config: Translation configuration
+
+        Returns:
+            The created dual PDF document
+        """
+        # Open the original PDF and insert translated PDF
+        dual = pymupdf.open(original_pdf_path)
+        dual.insert_file(translated_pdf)
+
+        # Rearrange pages to alternate between original and translated
+        page_count = translated_pdf.page_count
+        for page_id in range(page_count):
+            if translation_config.dual_translate_first:
+                dual.move_page(page_count + page_id, page_id * 2)
+            else:
+                dual.move_page(page_count + page_id, page_id * 2 + 1)
+
+        return dual
+
     def write_debug_info(
         self,
         pdf: pymupdf.Document,
@@ -388,48 +482,28 @@ def write(self, translation_config: TranslationConfig) -> TranslateResult:
                 original_pdf = pymupdf.open(self.original_pdf_path)
                 translated_pdf = pdf
 
-                # Create a new PDF for side-by-side pages
-                dual = pymupdf.open()
-                page_count = min(original_pdf.page_count, translated_pdf.page_count)
-
-                for page_id in range(page_count):
-                    # Get pages from both PDFs
-                    orig_page = original_pdf[page_id]
-                    trans_page = translated_pdf[page_id]
-
-                    # Calculate total width and use max height
-                    total_width = orig_page.rect.width + trans_page.rect.width
-                    max_height = max(orig_page.rect.height, trans_page.rect.height)
-
-                    # Create new page with combined width
-                    dual_page = dual.new_page(width=total_width, height=max_height)
-
-                    # Define rectangles for left and right sides
-                    left_width = (
-                        orig_page.rect.width
-                        if not translation_config.dual_translate_first
-                        else trans_page.rect.width
-                    )
-                    rect_left = pymupdf.Rect(0, 0, left_width, max_height)
-                    rect_right = pymupdf.Rect(left_width, 0, total_width, max_height)
-
-                    # Show pages according to dual_translate_first setting
-                    if translation_config.dual_translate_first:
-                        # Show translated page on left and original on right
-                        rect_left, rect_right = rect_right, rect_left
+                # Choose between side-by-side and alternating pages format
+                # Default to side-by-side if not specified
+                use_side_by_side = getattr(
+                    translation_config,
+                    "use_side_by_side_dual",
+                    True,
+                )
 
-                    # Show original page on left and translated on right (default)
-                    dual_page.show_pdf_page(
-                        rect_left,
+                if use_side_by_side:
+                    # Create a dual PDF with side-by-side pages (original and translation)
+                    dual = self.create_side_by_side_dual_pdf(
                         original_pdf,
-                        page_id,
-                        keep_proportion=True,
+                        translated_pdf,
+                        dual_out_path,
+                        translation_config,
                     )
-                    dual_page.show_pdf_page(
-                        rect_right,
+                else:
+                    # Create a dual PDF with alternating pages (original and translation)
+                    dual = self.create_alternating_pages_dual_pdf(
+                        self.original_pdf_path,
                         translated_pdf,
-                        page_id,
-                        keep_proportion=True,
+                        translation_config,
                     )
 
                 if translation_config.debug:

babeldoc/main.py

@@ -20,7 +20,7 @@
 from babeldoc.translation_config import TranslationConfig
 
 logger = logging.getLogger(__name__)
-__version__ = "0.1.11"
+__version__ = "0.1.12"
 
 
 def create_parser():
@@ -176,6 +176,18 @@ def create_parser():
         action="store_true",
         help="Enable all compatibility enhancement options (equivalent to --skip-clean --dual-translate-first --disable-rich-text-translate)",
     )
+    translation_params.add_argument(
+        "--use-side-by-side-dual",
+        default=True,
+        action="store_true",
+        help="Use side-by-side mode for dual PDF (default). When enabled, original and translated pages are shown side by side.",
+    )
+    translation_params.add_argument(
+        "--use-alternating-pages-dual",
+        dest="use_side_by_side_dual",
+        action="store_false",
+        help="Use alternating pages mode for dual PDF. When enabled, original and translated pages are arranged in alternate order.",
+    )
     translation_params.add_argument(
         "--report-interval",
         type=float,
@@ -426,6 +438,7 @@ async def main():
             dual_translate_first=args.dual_translate_first,
             disable_rich_text_translate=args.disable_rich_text_translate,
             enhance_compatibility=args.enhance_compatibility,
+            use_side_by_side_dual=args.use_side_by_side_dual,
             report_interval=args.report_interval,
             min_text_length=args.min_text_length,
         )

babeldoc/translation_config.py

@@ -37,6 +37,7 @@ def __init__(
         enhance_compatibility: bool = False,  # 增强兼容性模式
         report_interval: float = 0.1,  # Progress report interval in seconds
         min_text_length: int = 5,  # Minimum text length to translate
+        use_side_by_side_dual: bool = True,  # 是否使用拼版式双语PDF（并排显示原文和译文）
     ):
         self.input_file = input_file
         self.translator = translator
@@ -62,6 +63,7 @@ def __init__(
         )
         self.report_interval = report_interval
         self.min_text_length = min_text_length
+        self.use_side_by_side_dual = use_side_by_side_dual
         if progress_monitor:
             if progress_monitor.cancel_event is None:
                 progress_monitor.cancel_event = threading.Event()

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "BabelDOC"
-version = "0.1.11"
+version = "0.1.12"
 description = "Yet Another Document Translator"
 license = "AGPL-3.0"
 readme = "README.md"
@@ -124,7 +124,7 @@ dev = [
 ]
 
 [bumpver]
-current_version = "0.1.11"
+current_version = "0.1.12"
 version_pattern = "MAJOR.MINOR.PATCH[.PYTAGNUM]"
 
 [bumpver.file_patterns]