mineru-next-dev 4.0.0.dev20260625__py3-none-any.whl

mineru/__init__.py

@@ -0,0 +1,35 @@
+# Copyright (c) Opendatalab. All rights reserved.
+from .parser import (  # noqa: F401
+    DocumentParser,
+    DocxParser,
+    HtmlParser,
+    MinerUApiParser,
+    ParseResult,
+    PdfFlashParser,
+    PdfHybridParser,
+    PdfPipelineParser,
+    PdfVlmParser,
+    PptxParser,
+    XlsxParser,
+    parse,
+)
+from .types import Block, Line, PageInfo, Span  # noqa: F401
+
+__all__ = [
+    "Block",
+    "DocumentParser",
+    "DocxParser",
+    "HtmlParser",
+    "Line",
+    "MinerUApiParser",
+    "PageInfo",
+    "ParseResult",
+    "PdfFlashParser",
+    "PdfHybridParser",
+    "PdfPipelineParser",
+    "PdfVlmParser",
+    "PptxParser",
+    "Span",
+    "XlsxParser",
+    "parse",
+]

mineru/backend/__init__.py

@@ -0,0 +1 @@
+# Copyright (c) Opendatalab. All rights reserved.

mineru/backend/flash/__init__.py

@@ -0,0 +1 @@
+# Flash tier: CPU-only lightweight PDF text extraction.

mineru/backend/flash/pdf_extractor.py

@@ -0,0 +1,48 @@
+"""Flash tier PDF extractor — CPU-only text extraction via pypdfium2."""
+
+from __future__ import annotations
+
+import pypdfium2
+
+from ...utils.pdfium_guard import close_pdfium_child, close_pdfium_document, open_pdfium_document, pdfium_guard
+
+
+def extract_pages_text(filepath: str, start_page: int = 0, end_page: int | None = None) -> list[str]:
+    """Extract plain text from each PDF page, preserving empty pages."""
+
+    pdf = open_pdfium_document(pypdfium2.PdfDocument, filepath)
+    try:
+        with pdfium_guard():
+            total = len(pdf)
+        end = total if end_page is None else min(end_page, total)
+
+        pages: list[str] = []
+        for i in range(start_page, end):
+            page = None
+            text_page = None
+            try:
+                with pdfium_guard():
+                    page = pdf[i]
+                    text_page = page.get_textpage()
+                    pages.append(text_page.get_text_range() or "")
+            finally:
+                close_pdfium_child(text_page)
+                close_pdfium_child(page)
+        return pages
+    finally:
+        close_pdfium_document(pdf)
+
+
+def extract_text(filepath: str, start_page: int = 0, end_page: int | None = None) -> str:
+    """Extract plain text from a PDF file using pypdfium2.
+
+    Args:
+        filepath: Path to the PDF file.
+        start_page: 0-based start page (inclusive).
+        end_page: 0-based end page (exclusive).  None means all pages.
+
+    Returns:
+        Extracted text as a string, one page per `\\n\\n`.
+    """
+
+    return "\n\n".join(text for text in extract_pages_text(filepath, start_page, end_page) if text)

mineru/backend/hybrid/__init__.py

@@ -0,0 +1 @@
+# Copyright (c) Opendatalab. All rights reserved.