illuma-presentx 0.1.0

PPTX generation, QA verification, and template editing toolkit

illuma-presentx

PPTX generation, QA verification, and template editing toolkit for LLM-powered presentation workflows.

Installation

# Core (XML tools, thumbnail generation)
pip install git+https://github.com/illuma-ai/illuma-presentx.git

# With text extraction support
pip install "illuma-presentx[text] @ git+https://github.com/illuma-ai/illuma-presentx.git"

# With layout analysis support
pip install "illuma-presentx[layout] @ git+https://github.com/illuma-ai/illuma-presentx.git"

# All optional dependencies
pip install "illuma-presentx[all] @ git+https://github.com/illuma-ai/illuma-presentx.git"

Quick Start

CLI Usage

# Run the PPTX workflow runner
presentx --help

# Generate thumbnails from a presentation
python -m presentx.thumbnail input.pptx

# Unpack a PPTX for editing
python -m presentx.office.unpack input.pptx unpacked/

# Repack after editing
python -m presentx.office.pack unpacked/ output.pptx --original input.pptx

PPTXWorkflow Usage

from presentx import PPTXWorkflow

workflow = PPTXWorkflow(work_dir="/mnt/data")

# Unpack a template
workflow.unpack("template.pptx", "unpacked/")

# Generate thumbnails for analysis
workflow.thumbnail("template.pptx")

# Clean orphaned files after editing
workflow.clean("unpacked/")

# Repack into final output
workflow.pack("unpacked/", "output.pptx", original="template.pptx")

Architecture

src/presentx/
  __init__.py          # PPTXWorkflow, public API
  pptx_runner.py       # CLI entry point
  thumbnail.py         # Slide thumbnail grid generation
  add_slide.py         # Duplicate/create slides from layouts
  clean.py             # Remove orphaned slides, media, rels
  office/
    unpack.py          # Extract and pretty-print PPTX XML
    pack.py            # Validate, repair, and repack PPTX
    soffice.py         # LibreOffice wrapper for PDF conversion
    schemas/           # OOXML XSD schemas for validation
  js/
    merge_decks.cjs    # Multi-part deck merger (Node.js)
  docs/
    SKILL.md           # Main skill guide (creation, QA, design)
    pptxgenjs.md       # PptxGenJS API reference and patterns
    editing.md         # Template editing workflow

Available Tools

Tool	Purpose
thumbnail.py	Generate visual grid of slide thumbnails
unpack.py	Extract PPTX to directory with pretty-printed XML
pack.py	Repack directory into validated PPTX
add_slide.py	Duplicate a slide or create from a layout
clean.py	Remove orphaned slides, media, and relationships
soffice.py	Convert PPTX to PDF via LibreOffice
merge_decks.cjs	Merge multi-part PPTX files (Node.js)

Workflows

Create from Scratch (PptxGenJS)

Use when no template is available. See docs/pptxgenjs.md for the full PptxGenJS API reference including shapes, charts, icons, and common pitfalls.

Edit from Template

Use when modifying an existing presentation. See docs/editing.md for the unpack/edit/pack workflow, XML editing patterns, and formatting rules.

QA Verification

Every generated presentation should go through content QA (markitdown text check) and visual QA (slide-to-image conversion + inspection). See docs/SKILL.md for the full QA protocol.

Dependencies

Core: defusedxml, lxml, Pillow

Optional:

• markitdown[pptx] -- text extraction from PPTX files
• python-pptx -- layout analysis and programmatic slide manipulation

External (not pip-managed):

• Node.js + pptxgenjs -- creating presentations from scratch
• LibreOffice (soffice) -- PDF conversion for visual QA
• Poppler (pdftoppm) -- PDF to image conversion

License

MIT