Code/python/pdfcreator/main.py

from pypdf import PdfReader
from io import BytesIO
from reportlab.lib.pagesizes import LETTER
from reportlab.pdfgen import canvas
from pypdf import PdfReader, PdfWriter, PageObject

# ----------- CONFIG -------------
PDF_INPUTS = [
    {"file": "pdfcreator/input.pdf", "sections": ["1", "2.2", "3"]},
    {"file": "pdfcreator/input2.pdf", "sections": ["3", "4"]},
]
OUTPUT_PDF = "pdfcreator/combined_sections.pdf"


# Cropping ratios
HEADER_CROP = 0.1  # top of first page of section
FOOTER_CROP = 0.0  # bottom of pages
# --------------------------------

# ----- HELPER FUNCTIONS ---------


def build_outline_tree(reader):
    def _build(outline):
        tree = []
        for item in outline:
            if isinstance(item, list):
                tree[-1]["children"] = _build(item)
            else:
                tree.append({
                    "title": item.title.strip(),
                    "page": reader.get_destination_page_number(item),
                    "children": []
                })
        return tree
    return _build(reader.outline)


def find_section(nodes, title):
    for node in nodes:
        if node["title"] == title or node["title"].startswith(title + " "):
            return node
        found = find_section(node["children"], title)
        if found:
            return found
    return None


def collect_subtree_pages(node, pages=None):
    if pages is None:
        pages = []
    pages.append(node["page"])
    for child in node["children"]:
        collect_subtree_pages(child, pages)
    return pages


def flatten_outline_pages(nodes, pages=None):
    if pages is None:
        pages = []
    for node in nodes:
        pages.append(node["page"])
        flatten_outline_pages(node["children"], pages)
    return pages


def find_end_page(target_node, outline_tree, total_pages):
    subtree_pages = collect_subtree_pages(target_node)
    last_section_page = max(subtree_pages)
    all_outline_pages = flatten_outline_pages(outline_tree)
    all_outline_pages = sorted(set(all_outline_pages))
    for page in all_outline_pages:
        if page > last_section_page:
            return page
    return total_pages


def crop_page(page, top_ratio=0.0, bottom_ratio=0.0):
    llx, lly, urx, ury = page.mediabox
    height = ury - lly
    new_lly = lly + height * bottom_ratio
    new_ury = ury - height * top_ratio
    if new_ury <= new_lly:
        raise ValueError("Invalid crop ratios: page height would be negative")
    page.cropbox.lower_left = (llx, new_lly)
    page.cropbox.upper_right = (urx, new_ury)


def normalize_page_size(page, reference_box):
    """
    Force page MediaBox and CropBox to match reference.
    """
    page.mediabox.lower_left = reference_box.lower_left
    page.mediabox.upper_right = reference_box.upper_right

    page.cropbox.lower_left = reference_box.lower_left
    page.cropbox.upper_right = reference_box.upper_right

# --------------------------------


# --------- MAIN PROCESS ----------
writer = PdfWriter()
toc_entries = []  # To build TOC later
current_page_index = 0

for pdf_info in PDF_INPUTS:
    file_path = pdf_info["file"]
    sections_to_extract = pdf_info["sections"]

    reader = PdfReader(file_path)
    outline_tree = build_outline_tree(reader)
    total_pages = len(reader.pages)

    for section_title in sections_to_extract:
        target = find_section(outline_tree, section_title)
        if not target:
            print(f"[WARN] Section '{section_title}' not found in {file_path}")
            continue

        start_page = target["page"]
        end_page = find_end_page(target, outline_tree, total_pages)

        REFERENCE_BOX = None
        # Add pages to combined PDF
        for i, p in enumerate(range(start_page, end_page)):
            page = reader.pages[p]

            # Crop first page header+footer
            if i == 0:
                crop_page(page, top_ratio=HEADER_CROP,
                          bottom_ratio=FOOTER_CROP)
            else:
                crop_page(page, top_ratio=HEADER_CROP,
                          bottom_ratio=FOOTER_CROP)
                # crop_page(page, bottom_ratio=FOOTER_CROP)

            if REFERENCE_BOX is None:
                # Make a copy, not a reference
                REFERENCE_BOX = (
                    page.cropbox.lower_left,
                    page.cropbox.upper_right
                )
            # Step 3: Normalize page size
            page.mediabox.lower_left = REFERENCE_BOX[0]
            page.mediabox.upper_right = REFERENCE_BOX[1]
            page.cropbox.lower_left = REFERENCE_BOX[0]
            page.cropbox.upper_right = REFERENCE_BOX[1]

            writer.add_page(page)

        # Track TOC
        toc_entries.append({
            "title": f"{section_title} ({file_path})",
            "page": current_page_index + 1  # 1-based page number
        })
        current_page_index += (end_page - start_page)

# --------- ADD TOC PAGE(S) ----------


def create_toc_pdf(toc_entries):
    packet = BytesIO()
    c = canvas.Canvas(packet, pagesize=LETTER)
    c.setFont("Helvetica-Bold", 16)
    c.drawString(50, 750, "Table of Contents")
    c.setFont("Helvetica", 12)
    y = 720
    for entry in toc_entries:
        text = f"{entry['title']} .... {entry['page']}"
        c.drawString(50, y, text)
        y -= 20
        if y < 50:
            c.showPage()
            y = 750
    c.save()
    packet.seek(0)
    return PdfReader(packet)


toc_pdf = create_toc_pdf(toc_entries)

# Combine TOC + extracted sections
final_writer = PdfWriter()

# TOC first
for page in toc_pdf.pages:
    final_writer.add_page(page)

# Then extracted content
for page in writer.pages:
    final_writer.add_page(page)

# Save
with open(OUTPUT_PDF, "wb") as f:
    final_writer.write(f)


# --------- WRITE OUTPUT -----------
with open(OUTPUT_PDF, "wb") as f:
    final_writer.write(f)

print(f"[INFO] Combined PDF written to {OUTPUT_PDF} with TOC.")