ppt-analysis

1# PPT Analysis — .pptx / .ppt
2 
3## Environment
4 
5```python
6from pptx import Presentation
7from pptx.util import Inches
8import os, subprocess, json
9 
10# python-pptx is available
11# For .ppt (old binary format): convert via libreoffice
12def load_pptx(path):
13    if path.lower().endswith('.ppt'):
14        import subprocess
15        out_dir = os.path.dirname(path)
16        subprocess.run(
17            ['libreoffice', '--headless', '--convert-to', 'pptx', '--outdir', out_dir, path],
18            check=True, capture_output=True
19        )
20        path = path.rsplit('.', 1)[0] + '.pptx'
21    return Presentation(path), path
22```
23 
24---
25 
26## Core Method 1: Full Text Extraction (ALL slides)
27 
28```python
29def extract_all_slides_text(pptx_path):
30    """
31    Extract text from every slide: text frames, tables, chart titles.
32    For slides with no extractable text, flag them for image captioning.
33    """
34    prs, _ = load_pptx(pptx_path)
35    slides_data = []
36 
37    for slide_num, slide in enumerate(prs.slides, start=1):
38        slide_texts = []
39        has_text = False
40 
41        for shape in slide.shapes:
42            # Text frame (most common)
43            if shape.has_text_frame:
44                for para in shape.text_frame.paragraphs:
45                    text = para.text.strip()
46                    if text:
47                        slide_texts.append(text)
48                        has_text = True
49 
50            # Table
51            if shape.has_table:
52                tbl = shape.table
53…