cleanframe — simple pandas-based data cleanup

cleanframe detects common column types (email, phone, date, name) and automatically normalizes them with simple rule-based fixers and validators.

Quick start

1. Install dependencies

python -m pip install pandas

2. Run unit tests

pip install pytest
pytest -q

3. Use the library from a script

python - <<'PY'
import sys
sys.path.insert(0, 'datacleaner')
import pandas as pd
from cleanframe import fix

df = pd.DataFrame({
    'email': ['AASHAY@GMAIL.COM', 'vansh@EXAMPLE.com', 'bad'],
    'phone': ['800-123-4567', '(800) 333-4444', 'nope'],
    'date': ['2023-13-01', '2020-02-29', 'invalid'],
    'name': ['aashay', 'VANSH', '3rd street'],
})
cleaned = fix(df)
print(cleaned)
PY

Usage note

• To clean a CSV file, call cleanframe.fix from a Python script and write the resulting DataFrame to disk.

Removed files and tools

• This repo was cleaned to remove several helper/experimental scripts. Use pytest for tests and import cleanframe.fix in your Python code to perform cleaning.

Design notes

• Detection: heuristics based on validators and pattern matching
• Fixing: deterministic normalization, no external APIs or ML
• Invalid values become NaN after cleaning (safe for later processing)