pptx-generation

name	pptx-generation
description	Generate academic PowerPoint presentations (.pptx) using python-pptx. Use this skill for making PPT, slides, presentations, 生成PPT, 做PPT, 写PPT, 幻灯片. Provides complete helper functions and templates. Preferred over scientific-slides and frontend-slides for all PPTX generation.

Academic Presentation Generation

Overview

Create publication-quality academic presentations (.pptx) for group meetings, thesis defenses, conference talks, and product introductions.

Templates

Group Meeting (10-15 slides)

1. Title slide (project name, date, presenter)
2. Background / Context (1-2 slides)
3. This Week's Progress (3-5 slides)
4. Results & Figures (2-3 slides)
5. Challenges / Questions (1 slide)
6. Next Steps (1 slide)

Thesis Defense (30-50 slides)

1. Title → 2. Outline → 3. Background (5-8) → 4. Aims (1-2) → 5-7. Aim chapters (8-12 each) → 8. Discussion (2-3) → 9. Conclusions (1-2) → 10. Future → 11. Acknowledgments → 12. Backup

Conference Talk (12-20 slides, 10-15 min)

1. Title → 2. Problem → 3. Key Question → 4. Approach → 5-9. Results → 10. Implications → 11. Acknowledgments

Product / Project Introduction (10-15 slides)

1. Title → 2. Problem → 3. Solution → 4. Architecture → 5-8. Core features → 9-10. Case studies → 11. Comparison → 12. Quick start → 13. Roadmap → 14. Summary

Design Principles

• One message per slide: two messages → two slides
• Visual hierarchy: title → key message → supporting data → source
• Minimal text: bullet points, not paragraphs
• Figure-first: lead with figures, explain verbally
• Consistent styling: same fonts, colors, layout throughout

Technical Implementation

Dependencies

pip install -q python-pptx Pillow

CRITICAL: Execution Pattern

ALWAYS write the ENTIRE script FIRST, then run it. NEVER send text before running the script.

The script MUST be self-contained. Install deps, define helpers, build slides, save — all in ONE bash call:

pip install -q python-pptx Pillow 2>/dev/null && python3 << 'PPTXEOF'
# === FULL SCRIPT HERE (see template below) ===
PPTXEOF

Color Themes

Academic palettes — pick ONE and use consistently:

THEMES = {
    "lancet": {
        "primary": (0x00, 0x46, 0x8B),
        "accent": (0xED, 0x00, 0x00),
        "accent2": (0x42, 0xB5, 0x40),
        "light": (0xE8, 0xEE, 0xF4),
        "text": (0x33, 0x33, 0x33),
        "muted": (0x88, 0x88, 0x88),
    },
    "npg": {
        "primary": (0x3C, 0x54, 0x88),
        "accent": (0xE6, 0x4B, 0x35),
        "accent2": (0x00, 0xA0, 0x87),
        "light": (0xEE, 0xF0, 0xF5),
        "text": (0x33, 0x33, 0x33),
        "muted": (0x84, 0x91, 0xB4),
    },
    "nejm": {
        "primary": (0x00, 0x72, 0xB5),
        : (, , ),
        : (, , ),
        : (, , ),
        : (, , ),
        : (, , ),
    },
}

Complete Template Script

Copy this entire template. Fill in the SLIDES list with actual content. Do not restructure the helpers — just fill in content.

from pptx import Presentation
from pptx.util import Inches, Pt, Emu
from pptx.enum.text import PP_ALIGN, MSO_ANCHOR
from pptx.dml.color import RGBColor
import os

# --- CONFIG ---
OUTPUT_DIR = os.path.expanduser("~/.scienceclaw/workspace")
os.makedirs(OUTPUT_DIR, exist_ok=True)
FILENAME = "presentation.pptx"  # <-- change this
OUTPUT_PATH = os.path.join(OUTPUT_DIR, FILENAME)

THEME = "lancet"  # lancet | npg | nejm
T = {
    "lancet": {"p":(0,70,139),"a":(237,0,0),"a2":(66,181,64),"l":(232,238,244),"t":(51,51,51),"m":(136,136,136)},
    "npg":    {"p":(60,84,136),"a":(230,75,53),"a2":(0,160,135),:(,,),:(,,),:(,,)},
    :   {:(,,),:(,,),:(,,),:(,,),:(,,),:(,,)},
}[THEME]

 ():  RGBColor(*T[key])

prs = Presentation()
prs.slide_width  = Inches()
prs.slide_height = Inches()
W, H = prs.slide_width, prs.slide_height

 ():
    s = slide.shapes.add_shape(, left, top, w, h)
    s.line.fill.background()
     fill: s.fill.solid(); s.fill.fore_color.rgb = RGBColor(*fill)
    : s.fill.background()
     s

 ():
    box = slide.shapes.add_textbox(left, top, w, h)
    box.text_frame.word_wrap = 
    p = box.text_frame.paragraphs[]
    p.text = text
    p.font.size = Pt(size)
    p.font.bold = bold
    p.font.color.rgb = rgb(color)
    p.alignment = align
    box.text_frame.paragraphs[].space_after = Pt()
     box

 ():
    box = slide.shapes.add_textbox(left, top, w, h)
    tf = box.text_frame
    tf.word_wrap = 
     i, item  (items):
        p = tf.paragraphs[]  i ==   tf.add_paragraph()
        p.text = item
        p.font.size = Pt(size)
        p.font.color.rgb = rgb(color)
        p.space_after = Pt()
        p.level = 
     box

 ():
    s = prs.slides.add_slide(prs.slide_layouts[])
    add_shape(s, , , W, H, fill=T[])
    add_shape(s, , Inches(), W, Inches(), fill=((,T[][]-), (,T[][]-), (,T[][]-)))
    add_text(s, Inches(), Inches(), Inches(), Inches(), title, size=, bold=, color=, align=PP_ALIGN.LEFT)
     subtitle:
        add_text(s, Inches(), Inches(), Inches(), Inches(), subtitle, size=, color=)
     presenter:
        add_text(s, Inches(), Inches(), Inches(), Inches(), presenter, size=, color=)
     s

 ():
    s = prs.slides.add_slide(prs.slide_layouts[])
    add_shape(s, , , W, H, fill=T[])
    add_shape(s, Inches(), Inches(), Inches(), Pt(), fill=T[])
    add_text(s, Inches(), Inches(), Inches(), Inches(), title, size=, bold=, color=)
     s

 ():
    s = prs.slides.add_slide(prs.slide_layouts[])
    add_shape(s, , , W, Inches(), fill=T[])
    add_text(s, Inches(), Inches(), Inches(), Inches(), title, size=, bold=, color=)
    add_shape(s, Inches(), Inches(), Inches(), Pt(), fill=T[])
    add_bullets(s, Inches(), Inches(), Inches(), Inches(), bullets, size=)
     note:
        add_text(s, Inches(), Inches(), Inches(), Inches(), note, size=, color=)
     s

 ():
    s = prs.slides.add_slide(prs.slide_layouts[])
    add_shape(s, , , W, Inches(), fill=T[])
    add_text(s, Inches(), Inches(), Inches(), Inches(), title, size=, bold=, color=)
    add_shape(s, Inches(), Inches(), Inches(), Pt(), fill=T[])
     left_title:
        add_text(s, Inches(), Inches(), Inches(), Inches(), left_title, size=, bold=, color=)
    add_bullets(s, Inches(), Inches(), Inches(), Inches(), left_items, size=)
     right_title:
        add_text(s, Inches(), Inches(), Inches(), Inches(), right_title, size=, bold=, color=)
    add_bullets(s, Inches(), Inches(), Inches(), Inches(), right_items, size=)
     s

 ():
    s = prs.slides.add_slide(prs.slide_layouts[])
    add_shape(s, , , W, Inches(), fill=T[])
    add_text(s, Inches(), Inches(), Inches(), Inches(), title, size=, bold=, color=)
    add_shape(s, Inches(), Inches(), Inches(), Pt(), fill=T[])
    cols = (headers)
    tbl = s.shapes.add_table((rows)+, cols, Inches(), Inches(), Inches(), Inches()).table
     j, h  (headers):
        c = tbl.cell(, j); c.text = h
         p  c.text_frame.paragraphs:
            p.font.size = Pt(); p.font.bold = ; p.font.color.rgb = RGBColor(,,)
        c.fill.solid(); c.fill.fore_color.rgb = rgb()
     i, row  (rows):
         j, val  (row):
            c = tbl.cell(i+, j); c.text = (val)
             p  c.text_frame.paragraphs:
                p.font.size = Pt(); p.font.color.rgb = rgb()
            c.fill.solid(); c.fill.fore_color.rgb = RGBColor(*(T[]  i%==  (,,)))
     s

 ():
    s = prs.slides.add_slide(prs.slide_layouts[])
    add_shape(s, , , W, H, fill=T[])
    add_text(s, Inches(), Inches(), Inches(), Inches(), title, size=, bold=, color=, align=PP_ALIGN.CENTER)
     items:
        add_bullets(s, Inches(), Inches(), Inches(), Inches(), items, size=, color=)
     s






title_slide(
    ,
    subtitle=,
    presenter=,
)

content_slide(, [
    ,
    ,
    ,
])



end_slide(, [])





prs.save(OUTPUT_PATH)
()

Output Convention

• Save to ~/.scienceclaw/workspace/<descriptive-name>.pptx
• Print the absolute path and slide count on success
• If the script fails, print the traceback — do NOT silently swallow errors

Error Handling

• If python-pptx fails to install: tell the user to run pip install python-pptx on the host
• If script errors: read traceback, fix, re-run (max 3 attempts)
• After success: verify the file exists and has non-zero size before telling the user

Language & Content Rules

• Match the user's language. If user writes in Chinese, all slide content must be Chinese.
• Keep bullet points to 3-5 per slide, 6-10 words per bullet.
• Use the academic palette consistently.
• The model decides which helper to use for each slide (title_slide, content_slide, two_col_slide, table_slide, section_slide, end_slide).