Metadata-Version: 2.3
Name: gotaglio
Version: 0.2.2
Summary: Framework for creating and running experiment pipelines
License: MIT
Keywords: pipeline,experiments,llm,framework
Author: Mike Hopcroft
Author-email: mhopcroft@gmail.com
Requires-Python: >=3.12
Classifier: Development Status :: 3 - Alpha
Classifier: Intended Audience :: Developers
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Topic :: Software Development :: Libraries :: Python Modules
Requires-Dist: azure-ai-inference (>=1.0.0b7,<2.0.0)
Requires-Dist: gitpython (>=3.1.44,<4.0.0)
Requires-Dist: glom (>=24.11.0,<25.0.0)
Requires-Dist: jinja2 (>=3.1.5,<4.0.0)
Requires-Dist: nest-asyncio (>=1.6.0,<2.0.0)
Requires-Dist: numpy (>=2.2.1,<3.0.0)
Requires-Dist: openai (>=1.59.7,<2.0.0)
Requires-Dist: pydantic (>=2.11.7,<3.0.0)
Requires-Dist: pyparsing (>=3.2.1,<4.0.0)
Requires-Dist: pyyaml (>=6.0.2,<7.0.0)
Requires-Dist: rich (>=13.9.4,<14.0.0)
Requires-Dist: scipy (>=1.15.1,<2.0.0)
Requires-Dist: tiktoken (>=0.9.0,<0.10.0)
Project-URL: Homepage, https://github.com/MikeHopcroft/gotaglio
Project-URL: Repository, https://github.com/MikeHopcroft/gotaglio
Description-Content-Type: text/markdown

# GoTaglio

`GoTaglio` is a lightweight python toolbox for creating ML pipelines for model evaluation and case labeling. Its goal is to accelerate the Applied Science inner-loop by allowing principled experimentation to start informally on an engineer's machine in minutes, while producing learnings and artifacts that scale through production.

`GoTaglio` is designed to be very low friction. It is kind of like a thumb drive, loaded with power tools, that will work in any Python environment.
* It does not require significant cloud infrastructure deployment. All that is needed are model endpoints and credentials to access them.
* It can be used in cloud environments like [AzureML](https://azure.microsoft.com/en-us/products/machine-learning) or with frameworks like [mlflow](https://mlflow.org/).
* Pipeline code can be incorporated into production systems.

`GoTaglio` includes the following key elements:
* Ability to rapidly define and run end-to-end ML pipelines.
* Automatic logging and organization of information about runs.
* The ability to rerun an earlier experiment with small changes introduced on the command-line.
* Structured logging to facilitate run analysis, comparing runs and tracking key metrics over time as the pipeline evolves.
* A python library that can be accessed from [Jupyter notebooks](https://jupyter.org/).
* A command-line tool to simplify common operations.
* [COMING SOON] A web-based tool for oragnizing and labeling cases.

## Try GoTaglio

GoTaglio comes with [several samples](documentation/samples.md) that run out-of-the-box with included LLM mocks or your LLM endpoints.

## Learn GoTaglio

Get an overview of key [GoTaglio concepts](documentation/concepts.md) such as
* configuration merging
* models
* pipelines
* structured logging

## Use GoTaglio

Learn how to [incorporate GoTaglio into your process](documentation/usage.md) as
* a command-line tool
* a [Jupyter notebook](https://jupyter.org/) enhancement
* a python library

