Installation¶

Requirements¶

Python 3.9 or higher
pandas 1.5.0 or higher

Basic Installation¶

Install TracePipe from PyPI:

pip install tracepipe

Optional Dependencies¶

TracePipe has optional dependencies for additional features:

Arrow/Parquet SupportMemory ProfilingAll Features

pip install tracepipe[arrow]

Enables:

pd.read_parquet() tracking
df.to_parquet() with automatic column stripping
Optimized Arrow-based serialization

pip install tracepipe[memory]

Enables:

Memory usage statistics in tp.debug.inspect().stats()
Process memory tracking

pip install tracepipe[all]

Installs all optional dependencies.

Development Installation¶

For contributing to TracePipe:

git clone https://github.com/tracepipe/tracepipe.git
cd tracepipe

# Using uv (recommended)
uv sync --all-extras

# Or using pip
pip install -e ".[dev]"

Verify Installation¶

import tracepipe as tp
print(tp.__version__)  # Should print version number

# Quick test
import pandas as pd
tp.enable()
df = pd.DataFrame({"a": [1, 2, 3]})
df = df.dropna()
print(tp.check(df))
tp.disable()

Supported Pandas Versions¶

TracePipe is tested against:

pandas Version	Status
1.5.x	✅ Supported
2.0.x	✅ Supported
2.1.x	✅ Supported
2.2.x	✅ Supported

Troubleshooting¶

Import Error¶

If you get an import error, ensure pandas is installed:

pip install pandas>=1.5.0

Version Conflicts¶

If you have version conflicts, try creating a fresh virtual environment:

python -m venv .venv
source .venv/bin/activate  # Linux/Mac
# or: .venv\Scripts\activate  # Windows
pip install tracepipe