TraceFrame

Track and detect suspicious pandas dataframe transformations.

TraceFrame helps developers debug silent dataframe corruption, unexpected scaling issues, and transformation anomalies.

Features

• Track dataframe mutations
• Detect suspicious numeric shifts
• Generate readable reports
• Lightweight and local-first
• Simple API

---

Installation

pip install traceframe

---

Quick Start

import pandas as pd

from traceframe import watch

df = pd.DataFrame(
    {
        "salary": [50000, 60000, 70000]
    }
)

df = watch(df)

df["salary"] = df["salary"] / 1000

df.trace_report()

---

Example Output

[traceframe warning]
- Large scale shift detected in 'salary'

TraceFrame Report

---

Development

Clone repository:

git clone YOUR_REPO_URL

Install dependencies:

uv sync

Run tests:

uv run pytest

---

License

MIT