sn-da-non-spreadsheet-analysis

1# Document Analysis Skill — Word / PDF / PPT
2 
3End-to-end workflow for Word, PDF, and PPT document parsing. Each format has
4specific parsing pitfalls — follow the format-specific sub-skill exactly.
5 
6---
7 
8## Workflow
9 
10### Step 0 — Identify file type and input scope
11 
12```python
13import os
14 
15input_path = "/mnt/data/..."  # from user
16 
17# Detect single file vs directory (multi-file scenario)
18if os.path.isdir(input_path):
19    all_files = [
20        os.path.join(input_path, f)
21        for f in os.listdir(input_path)
22        if f.lower().endswith(('.docx', '.doc', '.pdf', '.pptx', '.ppt'))
23    ]
24    print(f"Found {len(all_files)} documents: {all_files}")
25else:
26    all_files = [input_path]
27 
28# Route by extension
29ext = os.path.splitext(all_files[0])[-1].lower()
30print(f"File type: {ext}")
31```
32 
33> **Critical rule**: When `input_path` is a directory OR the user says "这些文件" / "所有文档",
34> process **every file** and aggregate. Never stop at the first file.
35 
36---
37 
38### Step 1 — Load sub-skill by format
39 
40| Extension | Sub-skill to load |
41|-----------|------------------|
42| `.docx` / `.doc` | `capability/word-analysis/SKILL.md` |
43| `.pdf` | `capability/pdf-analysis/SKILL.md` |
44| `.pptx` / `.ppt` | `capability/ppt-analysis/SKILL.md` |
45 
46```
47read_file(path="<skills_root>/sn-da-non-spreadsheet-analysis/capability/<format>-analysis/SKILL.md")
48```
49 
50Load **only the sub-skill you need** — do not load all three at once.
51 
52---
53 
54### Step 2 — Parse and extract
55…