pdf_split.py

"""Split a PDF into two contiguous halves by page (PyMuPDF)."""
from __future__ import annotations

from dataclasses import dataclass


@dataclass(frozen=True)
class PdfSplitMeta:
    page_range_a: str  # e.g. "1-50" (1-based, inclusive)
    page_range_b: str
    total_pages: int


def split_pdf_two_halves(src_path: str, dest_a: str, dest_b: str) -> tuple[PdfSplitMeta | None, str]:
    """
    First half = pages 1 .. floor(n/2), second half = rest (at least one page each requires n >= 2).
    Returns (meta, "") on success or (None, error_message).
    """
    try:
        import fitz  # PyMuPDF
    except ImportError:
        return None, "PyMuPDF (fitz) is not installed."

    doc = None
    try:
        doc = fitz.open(src_path)
        n = len(doc)
        if n < 2:
            return None, "This PDF has only one page; split requires at least two pages. Try compressing or splitting the file manually."

        mid = n // 2  # pages 0..mid-1 in part A, mid..n-1 in part B
        out_a = fitz.open()
        try:
            out_a.insert_pdf(doc, from_page=0, to_page=mid - 1)
            out_a.save(dest_a)
        finally:
            out_a.close()

        out_b = fitz.open()
        try:
            out_b.insert_pdf(doc, from_page=mid, to_page=n - 1)
            out_b.save(dest_b)
        finally:
            out_b.close()

        meta = PdfSplitMeta(
            page_range_a=f"1-{mid}",
            page_range_b=f"{mid + 1}-{n}",
            total_pages=n,
        )
        return meta, ""
    except Exception as e:
        return None, f"PDF split failed: {e}"
    finally:
        if doc is not None:
            try:
                doc.close()
            except Exception:
                pass