document-generator

Document Generator Skill

Quick Reference

Document Type	Read/Extract	Create	Edit
PDF	`markitdown` or `pdfplumber`	`reportlab`	Unpack → XML → pack
DOCX	`pandoc` or unpack	`docx-js`	Unpack → XML → pack
XLSX	`pandas` or `openpyxl`	`openpyxl`	`openpyxl`
PPTX	`markitdown`	`pptxgenjs`	Unpack → XML → pack

Common Workflows

1. Read/Extract Content

# Universal: markitdown for text extraction
python -m markitdown document.pdf
python -m markitdown document.docx
python -m markitdown document.pptx

PDF tables specifically

python -c "import pdfplumber; ..."

Excel data analysis

python -c "import pandas as pd; df = pd.read_excel('file.xlsx')"


### 2. Create New Document
- **PDF**: Use `reportlab` (Python) — see [references/pdf-basic.md](references/pdf-basic.md)
- **DOCX**: Use `docx-js` (JavaScript) — see [references/docx-create.md](references/docx-create.md)
- **XLSX**: Use `openpyxl` (Python) — see [references/xlsx-create.md](references/xlsx-create.md)
- **PPTX**: Use `pptxgenjs` (JavaScript) — see [references/pptx-create.md](references/pptx-create.md)

### 3. Edit Existing Document
All office formats follow: unpack → edit XML → pack

```bash
# Unpack any office file
python scripts/unpack.py input.ext unpacked/

# For PDFs: use pypdf directly
python -c "from pypdf import PdfReader, PdfWriter; ..."

# Pack back to office format
python scripts/pack.py unpacked/ output.ext

4. Convert Between Formats

# Office → PDF (via LibreOffice)
python scripts/soffice.py --headless --convert-to pdf input.docx

# PDF → Images
python scripts/pdf_to_images.py input.pdf output_dir/

Document-Specific Guides

PDF Operations

Basic operations (merge, split, rotate, extract): See references/pdf-basic.md
Forms (fillable fields): See references/pdf-forms.md
Advanced (rendering, JavaScript libraries): See references/pdf-advanced.md

Quick PDF commands:

# Check if fillable
python scripts/pdf_check.py form.pdf

# Extract field info
python scripts/pdf_extract_fields.py form.pdf fields.json

# Convert to images
python scripts/pdf_to_images.py document.pdf images/

DOCX Operations

Creation: See references/docx-create.md
Editing: See references/docx-edit.md
Critical rules: Always use DXA units, never percentage for tables; never unicode bullets

XLSX Operations

Creation & editing: See references/xlsx-create.md
Financial models: Zero formula errors required; use formulas not hardcoded values
Recalculation: Run python scripts/recalc.py file.xlsx after creating formulas

PPTX Operations

Create from scratch: See references/pptx-create.md
Edit template: See references/pptx-edit.md

Essential Scripts

Script	Purpose
`scripts/unpack.py`	Extract and pretty-print any office file
`scripts/pack.py`	Repack edited files
`scripts/soffice.py`	LibreOffice headless for conversions
`scripts/recalc.py`	Recalculate Excel formulas
`scripts/thumbnail.py`	Create PPTX thumbnail grid
`scripts/add_slide.py`	Add/duplicate slides in PPTX
`scripts/clean.py`	Remove orphaned PPTX files
`scripts/accept_changes.py`	Accept all tracked changes in DOCX
`scripts/validate.py`	Validate DOCX/PPTX structure
`scripts/pdf_check.py`	Check if PDF has fillable fields
`scripts/pdf_extract_fields.py`	Extract form field info
`scripts/pdf_to_images.py`	Convert PDF to images
`scripts/pdf_fill.py`	Fill PDF forms

Note: scripts/comment.py is not yet implemented (requires templates).

Quality Assurance

Always verify output:

Content QA: Extract text and check for missing content
```
python -m markitdown output.ext
```

Visual QA (for presentations):

python scripts/soffice.py --headless --convert-to pdf output.pptx
python scripts/pdf_to_images.py output.pdf images/

Formula QA (Excel):

python scripts/recalc.py output.xlsx
# Fix any #REF!, #DIV/0!, #VALUE! errors

Dependencies

Python: pip install pandas openpyxl reportlab pypdf pdfplumber "markitdown[all]"
JavaScript: npm install -g docx pptxgenjs
System: LibreOffice, Poppler (pdftoppm)