---
id: tfx-bsl
version: "1.21.0"
license: Apache 2.0
license_treatment: permissive
maintenance: active
---

# tfx-bsl — tfx_bsl (TFX Basic Shared Libraries) contains libraries shared by many TFX (TensorFlow eXtended) libraries and components.

License: permissive · Maintenance: active · Downloads: 159.4K/mo

## What it is and what it does

tfx_bsl is a foundational library that packages shared code used across the TensorFlow eXtended ecosystem. It provides utilities, data structures, and helper functions that multiple TFX components—such as TensorFlow Data Validation (TFDV), TensorFlow Model Analysis (TFMA), and TensorFlow Transform (TFT)—rely on to avoid duplication and maintain consistency.

The package is designed primarily as an internal dependency rather than a direct user-facing library. Its public API (exported from tfx_bsl/public submodules) is intended for TFX pipeline authors and component developers, while other internal APIs carry no backward-compatibility guarantee. It depends on a substantial stack including apache-beam for distributed processing, tensorflow for computation, tensorflow-metadata for schema handling, and pyarrow for columnar data, making it a heavyweight but essential component of the TFX infrastructure.

Use it for:

- Building TFX pipelines that use TFDV, TFMA, or TFT, where tfx_bsl is installed as a transitive dependency
- Developing custom TFX components that need access to shared utilities from the tfx_bsl/public API
- Working with TensorFlow data validation and transformation workflows that rely on tfx_bsl's internal infrastructure
- Ensuring version compatibility across multiple TFX libraries by pinning tfx_bsl to match minor versions (e.g., 0.14.* for tensorflow_data_validation 0.14.*)

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

tfx_bsl provides shared libraries and utilities used internally by TensorFlow eXtended (TFX) components and standalone TFX libraries like TFDV, TFMA, and TFT.

Yes, but typically as an indirect dependency. Install directly only if you are building custom TFX components or need specific utilities from tfx_bsl/public. The package is actively maintained, has no known vulnerabilities, and uses a permissive license. Medium install friction is acceptable given its role in the TFX ecosystem. If you are using any TFX library, it will be installed automatically.

## Install

pip install tfx-bsl
uv add tfx-bsl
poetry add tfx-bsl

## Installing tfx-bsl

Before you install: Medium install friction due to compiled wheels for specific Python versions (3.10–3.13) and a large dependency tree including apache-beam, tensorflow, and tensorflow-metadata. Actively maintained with recent releases.

License in practice: Licensed under Apache 2.0 (permissive), allowing use in commercial and open-source projects with minimal restrictions.

Quickstart:

pip install tfx-bsl

import tfx_bsl
# Typically used as a dependency of other TFX libraries rather than directly

Requires Python 3.10 or later; pre-built wheels available for macOS 11+ (ARM64) and Linux (manylinux_2_27+). Direct use is uncommon; intended primarily as a transitive dependency of TFX components.

Verify before relying:

- Specific public API surface and stability guarantees beyond the note that only tfx_bsl/public symbols are intended for direct use
- Performance characteristics or scalability limits when used with large-scale data pipelines
- Compatibility matrix details for untested dependency combinations mentioned in the documentation

## Package facts

- License: Apache 2.0 (permissive)
- Python support: supports_current
- Install friction: medium
- Maintenance: active
- Downloads: 159.4K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

tfx shared libraries, tensorflow extended utilities, tfx component dependencies, tfx data validation support, tensorflow pipeline infrastructure, tensorflow-ecosystem, data-pipeline-infrastructure

[View on SkillFed](https://skillfed.io/packages/tfx-bsl) · [View on PyPI](https://pypi.org/project/tfx-bsl/)