Packages
Tower is a CLI and Python runtime for deploying Python data apps, pipelines, and AI agents to production with built-in orchestration, analytical storage, and multi-tenant APIs.
Towncrier generates changelogs from individual news fragment files rather than a single shared log, reducing merge conflicts and focusing on end-user-relevant changes.
tox automates testing across multiple Python environments and versions, managing virtual environments and running test suites with configurable test tools to verify package builds and compatibility.
Install it if you maintain a Python package or work on projects that require multi-environment testing.
A tox plugin that orchestrates testing of Ansible collections across multiple Python versions and Ansible releases, using ansible-test, pytest, and tox to manage environments.
A tox plugin that manages Docker container lifecycle during test runs, allowing you to spin up and configure containers (databases, services, etc.) as part of your test environment.
tox-gh automatically selects and runs the appropriate tox environments for each Python version in a GitHub Actions matrix, enabling parallel test execution across multiple workers.
Install it if you use tox for local testing and want to run the same environments in GitHub Actions without manual synchronization.
A tox plugin that automatically detects and runs the appropriate test environments on GitHub Actions based on the Python version in each job, with built-in log grouping support.
Install it if you run tox on GitHub Actions with multiple Python versions.
Formats tox.toml files with consistent style and comment preservation, applying opinionated rules similar to black for Python code.
tox-uv replaces virtualenv and pip with uv in tox environments, accelerating test environment creation and dependency installation while maintaining tox's test orchestration interface.
Install it if you want faster test environments and are comfortable accepting uv's behavior (including any bugs) as a trade-off for speed.
tox-uv-bare is a tox plugin that replaces virtualenv and pip with uv for test environment management, without bundling a uv binary—you provide your own uv installation.
A Python client library for interacting with TP-Link Omada SDN controllers, providing programmatic access to device management, site configuration, and network status.
tpu-inference is a hardware plugin for vLLM that enables high-performance serving of large language models on Google TPUs, supporting both PyTorch and JAX model definitions through a unified backend.
However, it is strictly tied to TPU infrastructure—it cannot run on CPU or GPU systems.
CLI tool that detects Cloud TPU devices and reads runtime metrics from libtpu, including memory usage and duty cycle, in both static snapshot and live streaming modes.
Wraps any iterable to display a real-time progress bar in the terminal or Jupyter notebook, showing iteration count, elapsed time, and estimated time remaining.
Adds a progress bar to joblib.Parallel execution using tqdm, displaying real-time completion status as parallel jobs run.
However, the unclear license status (CC-BY-SA-4.0 is not a standard software license) and aging maintenance cycle (362 days since last release) warrant caution in…
Converts tqdm progress bars to logging-friendly messages in headless environments, automatically switching from interactive terminal output to structured log lines when stdout is unavailable.
Displays multiple tqdm progress bars from worker processes in a multiprocessing pool, with logging and aggregate progress monitoring routed through the main process.
Automatically captures and displays local variables in Python tracebacks, eliminating manual debug logging while keeping exception context readable.
Install it if you regularly spend time reconstructing variable state from exceptions.
Provides a backport of Python's traceback module to older Python versions, allowing consistent traceback formatting and manipulation across Python 2.6+ and Python 3.2+.
Install only if you are maintaining legacy code that must run on Python 2.6–3.4 and cannot upgrade—otherwise, use the standard library traceback module.
A drop-in replacement for Python's traceback module that includes local and global variable values in exception stack traces.
Traceloop SDK instruments LLM applications to capture execution traces and send them to Traceloop or your own observability backend, built on OpenTelemetry.
TraceML tracks metrics, parameters, artifacts, and data references for machine learning experiments, with integrations for Keras, PyTorch, TensorFlow, Fastai, PyTorch Lightning, and HuggingFace, plus offline mode for tracking without an API.
Formats Python exceptions and tracebacks into readable HTML, JSON, and terminal output with variable inspection, working in Jupyter, FastAPI, Sanic, and standard Python scripts.
Install it if you spend time debugging Python code in notebooks, web frameworks, or terminals.
Trackio is a local-first experiment tracking library that logs metrics, media, and configs to a SQLite database with a web dashboard, and optionally syncs to Hugging Face Spaces for sharing.
Provides Python access to over 300,000 economic indicators, exchange rates, stock indexes, bond yields, and commodity prices from Trading Economics' databases, with support for XML, CSV, and JSON export.
Install it if you need programmatic access to Trading Economics' economic and market data; the main prerequisite is obtaining an API key from Trading Economics.
Query TradingView's screener API to retrieve market data across stocks, crypto, options, forex, futures, and other instruments with support for technical indicators, fundamental metrics, and multiple timeframes.
Install it if you need programmatic access to TradingView market data; be aware that real-time data requires authentication and that unauthenticated queries return…
Retrieves technical analysis indicators and trading signals from TradingView for financial symbols without requiring Selenium or a browser.
Trafaret validates and transforms foreign data (JSON, configs, user input) against schemas you define, with support for custom error messages and async checking.
Trafilatura extracts main text, metadata, and structured content from web pages and HTML, converting raw HTML into clean, usable data in multiple output formats.
Adds HTTP 402 Payment Required protocol support to AI agents and APIs, enabling them to charge for services using blockchain-based payments through the D402 standard.
However, verify the license terms before production use, confirm that Python 3.12+ is available in your environment, and understand that the 25 runtime dependencies…
Parses source code into a queryable graph of functions, classes, calls, and semantic annotations using tree-sitter and rustworkx for security and code-structure analysis.
Walk filesystem paths while respecting .gitignore rules, and execute functions on found files using a process pool for parallelism.
Install it if you need .gitignore-aware file discovery with built-in parallelism; skip it if you need active maintenance or ongoing feature development.
Trainer is a PyTorch model training framework that handles the training loop, optimization, and experiment logging for machine learning models with support for mixed precision, distributed training, and callback customization.
Not recommended if you require active maintenance or support for the latest PyTorch features.
Traitlets provides typed attributes for Python classes with automatic validation, coercion, change notifications, and dynamic defaults—a pure-Python alternative to the Enthought traits library.
Install it if you need typed, validated attributes with change notifications—particularly if you are building configuration systems, data models, or interactive…
Provides custom trait types for scientific computing, extending the traitlets validation framework with domain-specific attribute types.
However, the aging maintenance status (296 days since last release) means you should verify that the trait types you need are already implemented and stable; expect…
Parses track names and filenames to extract language, codec, and accessibility metadata (SDH, closed captions, forced subtitles, commentary) for audio and subtitle streams.
Install it if you work with video files that need track-level metadata enrichment or language variant detection.
Trame is a Python web framework for building interactive visual analytics applications that combine 3D visualization, data analysis, and web UI components into a single browser-based or desktop application.
trame-client is the browser-side runtime for trame applications, handling client-server communication, state synchronization, dynamic component loading, and template rendering.
Provides shared helper functions, decorators, and utility classes for trame packages to use across server, client, and widget contexts without duplicating code.
Provides Vue.js utility widgets and helper components that extend trame's widget library for building interactive web applications.