Merge pull request #140 from funstory-ai/feat/support-xobject-in-scan…

…ned-detection feat(scanned detection): Add support for XObject in scanned file detection
funstory-ai · Mar 9, 2025 · 45abc72 · 45abc72
2 parents ba8665b + dc5251c
commit 45abc72
Show file tree

Hide file tree

Showing 5 changed files with 9 additions and 5 deletions.
diff --git a/babeldoc/__init__.py b/babeldoc/__init__.py
@@ -1 +1 @@
-__version__ = "0.1.29"
+__version__ = "0.1.30"
diff --git a/babeldoc/const.py b/babeldoc/const.py
@@ -2,7 +2,7 @@
 import subprocess
 from pathlib import Path
 
-__version__ = "0.1.29"
+__version__ = "0.1.30"
 
 CACHE_FOLDER = Path.home() / ".cache" / "babeldoc"
 

diff --git a/babeldoc/document_il/midend/detect_scanned_file.py b/babeldoc/document_il/midend/detect_scanned_file.py
@@ -113,6 +113,10 @@ def detect_page_is_scanned(page: il_version_1.Page, pdf: pymupdf.Document) -> bo
         pdf.update_object(new_xref, "<<>>")
         pdf.update_stream(new_xref, page.base_operations.value.encode("utf-8"))
         pdf[page.page_number].set_contents(new_xref)
+
+        for xobj in page.pdf_xobject:
+            pdf.update_stream(xobj.xref_id, xobj.base_operations.value.encode("utf-8"))
+
         after_page_image = pdf[page.page_number].get_pixmap()
         after_page_image = np.frombuffer(after_page_image.samples, np.uint8).reshape(
             after_page_image.height,

diff --git a/babeldoc/main.py b/babeldoc/main.py
@@ -25,7 +25,7 @@
 from babeldoc.translation_config import WatermarkOutputMode
 
 logger = logging.getLogger(__name__)
-__version__ = "0.1.29"
+__version__ = "0.1.30"
 
 
 def create_parser():

diff --git a/pyproject.toml b/pyproject.toml
@@ -1,6 +1,6 @@
 [project]
 name = "BabelDOC"
-version = "0.1.29"
+version = "0.1.30"
 description = "Yet Another Document Translator"
 license = "AGPL-3.0"
 readme = "README.md"
@@ -133,7 +133,7 @@ pythonpath = [".", "src"]
 testpaths = ["tests"]
 
 [bumpver]
-current_version = "0.1.29"
+current_version = "0.1.30"
 version_pattern = "MAJOR.MINOR.PATCH[.PYTAGNUM]"
 
 [bumpver.file_patterns]