Required Software on Teradata Vantage
Based on the connection to Vantage, the teradataml package requires the following minimum software versions be installed on Vantage:
Teradata Vantage with database release 16.20.16.01 or later or VantageCloud Lake.
Required Software on Client
The teradataml package requires a minimum version of the following software be installed on the client:
- Python >= 3.8 (For IBM PowerPC Python >= 3.9)
- Operating Systems:
- Windows OS: Windows 7
- macOS: os x 10.9
- Linux: Ubuntu 16.04, CentOS 7.0, RHEL 7.1 with gcc 5.5, or SLES 12
The teradataml package only supports 64-bit version of these operating systems. - Dependent packages:
- teradatasql 17.10.00.11
- teradatasqlalchemy 20.0.0.02
- pandas 0.22.00
- psutil
- requests >=2.25.1
- scikit-learn >= 0.24.2
- IPython >= 8.10.0
- imbalanced-learn >= 0.8.0
- pyjwt >= 2.8.0
- cryptography >= 42.0.5
Compatible Development Environment
The following third-party Integrated Development Environment (IDE) and data science user interface (UI) tool is compatible with Teradata Package for Python:
- Jupyter
- Spyder
Optional Dependency Installation
To use the the following Teradata Package for Python features, you need to manually install the corresponding optional dependencies.
| Feature | Required Packages |
|---|---|
| AutoML Family functions | imbalanced-learn, scikit-learn |
| OpenSourceML | scikit-learn, lightgbm |
| Visualization | matplotlib, seaborn |
| Teradataml EDA UI | teradatamlwidgets |
Install any optional feature dependencies using:
pip install teradataml[<feature_name>]
Where <feature_name> can be one of:
- automl – Installs dependencies for AutoML capabilitiesExample:
pip install teradataml[autonml]
- openml – Installs dependencies for OpenSourceML featuresExample:
pip install teradataml[openml]
- visualization – Installs visualization libraries
Example:
pip install teradataml[visualization]
- eda-ui – Installs UI components for EDA workflows
Example:
pip install teradataml[eda-ui]