When your data grows beyond a toy example,
pprintstarts to break.
You see cut-off arrays. Flattened hierarchies. Lost context.
And worst of all? Hidden bugs buried under “pretty” formatting.
That’s why I created SetPrint — a Python library that shows structure, not just values.
🔍 What this article covers
- ✅ Side-by-side comparisons:
pprint/setprint
- ✅ Real-world examples: image data, confusion matrices
- ✅ Benchmarks + 5 must-know tips for structured debugging
🧪 Quick Colab Demo
Want to see it in action? Try this demo notebook — no install needed.
(To use Colab, a Google account is required.)
🔗 Try this notebook on Google Colab
1. Visual Comparison — pprint vs setprint
data = {
"users": [
{"name": "Alice", "scores": np.array([95, 88, 76])},
{"name": "Bob", "scores": np.array([72, 85, 90])}
],
"meta": {"created": "2025-04-23", "version": 1.2}
}
== pprint ==
{'meta': {'created': '2025-04-23', 'version': 1.2},
'users': [{'name': 'Alice', 'scores': array([95, 88, 76])},
{'name': 'Bob', 'scores': array([72, 85, 90])}]}
== setprint ==
keep_settings
['y', 'y', 'y', 'y']
--------------------------------------------------------
◆dict
├── users:►list
│ ├───── -------. ◆dict
│ │ ├───────── name : Alice
│ │ └───────── scores:>ndarray
│ │ ├─────── 95
│ │ ├─────── 88
│ │ └─────── 76
│ └───── -------. ◆dict
│ ├───────── name : Bob
│ └───────── scores:>ndarray
│ ├─────── 72
│ ├─────── 85
│ └─────── 90
└── meta :◆dict
├───── created:2025-04-23
└───── version: 1.2
--------------------------------------------------------
Even at a glance, the structure is crystal clear.
2. Image Arrays? No Problem
rgb = np.random.randint(0, 255, size=(3, 3, 3))
set_collection ( route='SLIM', y_axis=False, keep_settings={1: 'yf', 2: 'f'} )
keep_settings
['yf', 'f', 'f']
-------------------------------------------------------------------------------------------------
>ndarray
├──── >ndarray [ >ndarray [ 255 0 4 ] >ndarray [ 255 85 0 ] >ndarray [ 255 170 0 ] ]
├──── >ndarray [ >ndarray [ 170 255 0 ] >ndarray [ 85 255 0 ] >ndarray [ 0 255 4 ] ]
└──── >ndarray [ >ndarray [ 0 170 255 ] >ndarray [ 0 85 255 ] >ndarray [ 4 0 255 ] ]
-------------------------------------------------------------------------------------------------
You don’t just see the values—you see the hierarchy.
3. Confusion Matrices as Text-Based Heatmaps
cm = np.array([[50, 2, 0, 0], [3, 45, 1, 0], [0, 4, 60, 5], [0, 0, 6, 70]])
set_collection ( route='SLIM', y_axis=False, keep_settings={1:'y', 2:'x'} )
keep_settings
['y', 'x']
------------------------------
>ndarray
├──── >ndarray ┬──┬──┬──┐
│ 50 2 0 0
├──── >ndarray ┬──┬──┬──┐
│ 3 45 1 0
├──── >ndarray ┬──┬──┬──┐
│ 0 4 60 5
└──── >ndarray ┬──┬──┬──┐
0 0 6 70
------------------------------
4. Benchmark: Visibility vs Speed
| Library | Runtime (ms) | Structure Clarity |
|---|---|---|
| pprint | 1.5 | ★☆☆☆☆ |
| rich.pretty | 3.2 | ★★☆☆☆ |
| setprint | 4.8 | ★★★★★ |
A few ms extra for full visibility? Totally worth it.
5. 5 Must-Know Tips for Using SetPrint
- Use
keep_settings = {1: 'x', 3: 'yf', 4: 'f'}for full control over dimensions
- Show vertical guides with
y_axis=True
- Output is a list of lines → easy to write to file
- Handles mixed types:
dict+list+ndarray? No problem
- Switch between styles like
'SLIM','BOLD', or even'HALF'
6. One Function, Infinite Views
from setprint import SetPrint
list_data = SetPrint(data)
formatted = list_data.set_collection(
keep_settings={1:'x', 2:'yf', 3:'f'},
route='SLIM', y_axis=True
)
for line in formatted:
print(line)
🚀 Try it Now
pip install setprint
🎯 If you found it useful, please consider giving a ⭐ on GitHub!
🐛 Bug reports, 💡 feature requests, and 📬 pull requests are all welcome!
📎 GitHub: mtur2007/SetPrint
📘 PyPI: setprint
🔍 Colab Demo: Try on Colab
🔗 Also available on Medium!
💬 I'd love to hear your feedback on either platform. ✨
If you’ve ever said “I just want to see the damn structure” — SetPrint is for you.
(Thank you for reading! This is my very first article introducing SetPrint to the international audience. I'd love to hear your feedback and thoughts.)