lidge-jun/pptx_original

PPTX Skill

Quick Reference

| Task | Guide | |------|-------| | Read/analyze content | python -m markitdown presentation.pptx | | Edit or create from template | Read editing.md | | Create from scratch | Read pptxgenjs.md |

---

Reading Content

python -m markitdown presentation.pptx        # text extraction
python scripts/thumbnail.py presentation.pptx  # visual overview
python scripts/office/unpack.py presentation.pptx unpacked/  # raw XML

---

Editing Workflow

Read editing.md for full details.

1. Analyze template with thumbnail.py
2. Unpack → manipulate slides → edit content → clean → pack

---

Creating from Scratch

Read pptxgenjs.md for full details. Use when no template or reference presentation is available.

---

Design Principles

Before Starting

• Pick a bold, content-informed color palette: if swapping your colors into a different presentation would still "work," your choices aren't specific enough
• Dominance over equality: one color dominates (60–70%), 1–2 supporting tones, one sharp accent
• Dark/light contrast: dark backgrounds for title + conclusion slides, light for content ("sandwich" structure)
• Commit to a visual motif: pick one distinctive element and repeat it (rounded image frames, icons in colored circles, thick single-side borders)

Color Palettes

| Theme | Primary | Secondary | Accent | |-------|---------|-----------|--------| | Midnight Executive | 1E2761 (navy) | CADCFC (ice blue) | FFFFFF | | Forest & Moss | 2C5F2D (forest) | 97BC62 (moss) | F5F5F5 | | Coral Energy | F96167 (coral) | F9E795 (gold) | 2F3C7E | | Warm Terracotta | B85042 (terracotta) | E7E8D1 (sand) | A7BEAE | | Ocean Gradient | 065A82 (deep blue) | 1C7293 (teal) | 21295C | | Charcoal Minimal | 36454F (charcoal) | F2F2F2 (off-white) | 212121 | | Teal Trust | 028090 (teal) | 00A896 (seafoam) | 02C39A | | Berry & Cream | 6D2E46 (berry) | A26769 (dusty rose) | ECE2D0 | | Sage Calm | 84B59F (sage) | 69A297 (eucalyptus) | 50808E | | Cherry Bold | 990011 (cherry) | FCF6F5 (off-white) | 2F3C7E |

Per-Slide Guidelines

Every slide needs a visual element — image, chart, icon, or shape.

Layout options:

• Two-column (text left, illustration right)
• Icon + text rows (icon in colored circle, bold header, description below)
• 2×2 or 2×3 grid with content blocks
• Half-bleed image with content overlay

Data display:

• Large stat callouts (60–72pt numbers with small labels)
• Comparison columns (before/after, pros/cons)
• Timeline or process flow (numbered steps, arrows)

Typography

Choose a header font with personality paired with a clean body font:

| Header | Body | |--------|------| | Georgia | Calibri | | Arial Black | Arial | | Cambria | Calibri | | Trebuchet MS | Calibri | | Palatino | Garamond |

| Element | Size | |---------|------| | Slide title | 36–44pt bold | | Section header | 20–24pt bold | | Body text | 14–16pt | | Captions | 10–12pt muted |

Spacing

• 0.5" minimum margins
• 0.3–0.5" between content blocks
• Leave breathing room

Common Mistakes to Avoid

• Repeating the same layout across slides — vary columns, cards, and callouts
• Centering body text — left-align paragraphs and lists; center only titles
• Insufficient size contrast — titles need 36pt+ to stand out from 14–16pt body
• Defaulting to blue — pick colors reflecting the specific topic
• Inconsistent spacing — choose 0.3" or 0.5" gaps and use consistently
• Styling one slide but leaving the rest plain — commit fully or keep it simple throughout
• Text-only slides — add images, icons, charts, or visual elements
• Missing text box padding — when aligning lines/shapes with text edges, set margin: 0 or offset to account for padding
• Low-contrast elements — icons and text both need strong contrast against background
• Accent lines under titles — these read as AI-generated; use whitespace or background color instead

---

QA (Required)

Approach QA as a bug hunt. Your first render is almost never correct.

Content QA

python -m markitdown output.pptx
python -m markitdown output.pptx | grep -iE "xxxx|lorem|ipsum|this.*(page|slide).*layout"

Fix any leftover placeholder text before declaring success.

Visual QA

Use subagents for visual inspection — even for 2–3 slides. Fresh eyes catch what you miss after staring at code.

Convert slides to images (see below), then have the subagent inspect with this prompt:

Visually inspect these slides. Assume there are issues — find them.

Look for:
- Overlapping elements (text through shapes, stacked elements)
- Text overflow or cut off at edges/box boundaries
- Elements too close (< 0.3" gaps) or uneven spacing
- Insufficient margin from slide edges (< 0.5")
- Low-contrast text or icons
- Text boxes too narrow causing excessive wrapping
- Leftover placeholder content

For each slide, list issues found.

Verification Loop

1. Generate slides → convert to images → inspect
2. List issues found
3. Fix issues
4. Re-verify affected slides — one fix often creates another problem
5. Repeat until a full pass reveals no new issues

Complete at least one fix-and-verify cycle before declaring success.

---

Converting to Images

python scripts/office/soffice.py --headless --convert-to pdf output.pptx
pdftoppm -jpeg -r 150 output.pdf slide
# Creates slide-01.jpg, slide-02.jpg, etc.

# Re-render specific slide after fix:
pdftoppm -jpeg -r 150 -f N -l N output.pdf slide-fixed

---

Dependencies

• pip install "markitdown[pptx]" — text extraction
• pip install Pillow — thumbnail grids
• npm install -g pptxgenjs — creating from scratch
• LibreOffice (soffice) — PDF conversion (auto-configured via scripts/office/soffice.py)
• Poppler (pdftoppm) — PDF to images