diff options
| -rw-r--r-- | .gitignore | 1 | ||||
| -rwxr-xr-x | kjandoc | 184 |
2 files changed, 185 insertions, 0 deletions
diff --git a/.gitignore b/.gitignore new file mode 100644 index 0000000..1d17dae --- /dev/null +++ b/.gitignore @@ -0,0 +1 @@ +.venv @@ -0,0 +1,184 @@ +#!/usr/bin/env python3 + +# merge pptx files using libreoffice slide rendering + python-pptx reconstruction. +# creative approach: render each slide as a high-res image via libreoffice, +# then stitch them into one pptx. this preserves 100% of the visual formatting +# since we're working with exact raster snapshots of each slide. + +import os +import sys +import glob +import shutil +import subprocess +import tempfile +from pathlib import Path +from PIL import Image +from pptx import Presentation +from pptx.util import Emu + + +DPI = 300 # high-res export + + +def pptx_to_images(pptx_path, output_dir): + # use libreoffice to convert pptx -> pdf, then pdf -> images via pdftoppm. + # fallback: libreoffice direct png export if pdftoppm unavailable. + pptx_path = os.path.abspath(pptx_path) + pdf_path = os.path.join(output_dir, Path(pptx_path).stem + '.pdf') + + # step 1: pptx -> pdf via libreoffice (preserves all formatting) + subprocess.run([ + 'libreoffice', '--headless', '--convert-to', 'pdf', + '--outdir', output_dir, pptx_path + ], check=True, capture_output=True) + + if not os.path.exists(pdf_path): + raise RuntimeError(f"libreoffice failed to produce {pdf_path}") + + # step 2: pdf -> png images + # try pdftoppm first (from poppler-utils, higher quality) + image_prefix = os.path.join(output_dir, 'slide') + + if shutil.which('pdftoppm'): + subprocess.run([ + 'pdftoppm', '-png', '-r', str(DPI), + pdf_path, image_prefix + ], check=True, capture_output=True) + else: + # fallback: use libreoffice to export as images directly + # this works but pdftoppm gives better quality + subprocess.run([ + 'libreoffice', '--headless', '--convert-to', 'png', + '--outdir', output_dir, pdf_path + ], check=True, capture_output=True) + + # collect and sort image files + images = sorted(glob.glob(os.path.join(output_dir, 'slide-*.png'))) + if not images: + images = sorted(glob.glob(os.path.join(output_dir, '*.png'))) + + if not images: + raise RuntimeError("no slide images produced") + + return images + + +def images_to_pptx(image_groups, output_path, slide_width_emu=9144000, slide_height_emu=6858000): + # build a pptx from slide images, one image per slide filling the entire area. + prs = Presentation() + prs.slide_width = slide_width_emu + prs.slide_height = slide_height_emu + + # use blank layout (index 6 is typically blank) + blank_layout = None + for layout in prs.slide_layouts: + if layout.name == 'Blank': + blank_layout = layout + break + if blank_layout is None: + blank_layout = prs.slide_layouts[6] if len(prs.slide_layouts) > 6 else prs.slide_layouts[0] + + total = 0 + for label, images in image_groups: + for img_path in images: + slide = prs.slides.add_slide(blank_layout) + + # remove any placeholder shapes from blank layout + for ph in list(slide.placeholders): + sp = ph._element + sp.getparent().remove(sp) + + # add image covering the full slide + slide.shapes.add_picture( + img_path, + left=0, + top=0, + width=slide_width_emu, + height=slide_height_emu + ) + total += 1 + print(f" [{label}] {len(images)} slides added") + + prs.save(output_path) + return total + + +def merge_presentations(*pptx_files, output='merged.pptx'): + # merge pptx files with zero formatting loss via image rendering. + if not pptx_files: + raise ValueError("no files provided") + + image_groups = [] + + for pptx_file in pptx_files: + label = Path(pptx_file).stem + print(f"[*] rendering: {label}") + + tmpdir = tempfile.mkdtemp(prefix=f'pptx_merge_{label}_') + try: + images = pptx_to_images(pptx_file, tmpdir) + image_groups.append((label, images)) + print(f" -> {len(images)} slides rendered at {DPI} DPI") + except Exception as e: + print(f" [!] error rendering {label}: {e}", file=sys.stderr) + raise + + # get slide dimensions from first presentation + first_prs = Presentation(pptx_files[0]) + sw = first_prs.slide_width + sh = first_prs.slide_height + + print(f"[*] building merged presentation...") + total = images_to_pptx(image_groups, output, sw, sh) + print(f"[+] merged {total} slides from {len(pptx_files)} presentations -> {output}") + + # cleanup temp dirs + for label, images in image_groups: + if images: + tmpdir = os.path.dirname(images[0]) + shutil.rmtree(tmpdir, ignore_errors=True) + + return output + + +if __name__ == "__main__": + import argparse + + parser = argparse.ArgumentParser( + description='merge pptx files with perfect formatting preservation via rendering', + usage='%(prog)s input1.pptx [input2.pptx ...] -o output.pptx' + ) + parser.add_argument( + 'inputs', + metavar='INPUT', + nargs='+', + help='input pptx files to merge' + ) + parser.add_argument( + '-o', '--output', + default='merged.pptx', + help='output filename (default: merged.pptx)' + ) + parser.add_argument( + '--dpi', + type=int, + default=DPI, + help=f'rendering DPI for slide images (default: {DPI})' + ) + + args = parser.parse_args() + + # validate input files exist + for f in args.inputs: + if not os.path.exists(f): + print(f"[!] error: file not found: {f}", file=sys.stderr) + sys.exit(1) + if not f.lower().endswith('.pptx'): + print(f"[!] error: not a pptx file: {f}", file=sys.stderr) + sys.exit(1) + + # update global DPI if specified + if args.dpi != DPI: + globals()['DPI'] = args.dpi + + merge_presentations(*args.inputs, output=args.output)
\ No newline at end of file |