From 650a7e594911ee838c0b7926efbe06b7f999d2ec Mon Sep 17 00:00:00 2001 From: myhloli Date: Fri, 3 Jul 2026 20:38:38 +0800 Subject: [PATCH] feat: unify PDFium page object boundary retrieval for compatibility --- mineru/utils/pdf_classify.py | 19 ++++++++++++++++++- pyproject.toml | 2 +- 2 files changed, 19 insertions(+), 2 deletions(-) diff --git a/mineru/utils/pdf_classify.py b/mineru/utils/pdf_classify.py index 8e23c915..20a2f3b8 100644 --- a/mineru/utils/pdf_classify.py +++ b/mineru/utils/pdf_classify.py @@ -773,6 +773,21 @@ def _resolve_pdf_object(obj): return obj +def _get_pdfium_page_object_bounds(page_object): + """兼容 pypdfium2 4.x/5.x，统一获取页面对象的边界坐标。""" + get_bounds = getattr(page_object, "get_bounds", None) + if callable(get_bounds): + return get_bounds() + + get_pos = getattr(page_object, "get_pos", None) + if callable(get_pos): + return get_pos() + + raise AttributeError( + "PDFium page object has neither get_bounds() nor get_pos()" + ) + + def get_high_image_coverage_ratio_pdfium(pdf_doc, page_indices): high_image_coverage_pages = 0 @@ -791,7 +806,9 @@ def get_high_image_coverage_ratio_pdfium(pdf_doc, page_indices): filter=[pdfium_c.FPDF_PAGEOBJ_IMAGE], max_depth=3 ): try: - left, bottom, right, top = page_object.get_pos() + left, bottom, right, top = _get_pdfium_page_object_bounds( + page_object + ) image_area += max(0.0, right - left) * max(0.0, top - bottom) finally: close_pdfium_child(page_object) diff --git a/pyproject.toml b/pyproject.toml index 27a9e565..8c77c623 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -38,7 +38,7 @@ dependencies = [ "requests", "httpx", "pillow>=11.0.0", - "pypdfium2>=4.30.0", + "pypdfium2>=4.30.0,<6.0.0", "pypdf>=5.6.0", "reportlab", "pdftext>=0.6.3,<0.8.0",