Pyteryx/AGENTS.md

Project Overview

Pyteryx is a python-first reimplementation of the Alteryx tool. The intent is to keep the same features of obserability and a rich DAG based user interface, but take advantage of the ease of deploying python.

Pyteryx should have a near identical look-and-feel of Alteryx, but have the option of exporting a workflow (yxmd) as a python script.

Alteryx Tools Documentation / UI Examples

https://help.alteryx.com/current/en/designer/tools.html

Alteryx Tool Logic / Conversion code

./alteryx-to-python-migration-strategy-main This migration system helps organizations transition from Alteryx's visual workflow platform to Python-based data processing pipelines. It automatically converts Alteryx workflow XML files (.yxmd) into equivalent Python code using pandas, numpy, and other standard data science libraries.

./alteryx-to-python-migration-strategy-main/migration_toolkit.py This file contains conversion logic that could be used to create an Alteryx execution engine that can run within the Pyteryx app.

Back End

This is a Python project using uv for dependency management and environment setup.

Agent instructions

When interacting with this project, AI agents should adhere to the following guidelines:

• Always use uv for dependency management and environment activation. Avoid using pip directly.
• Do not manually edit pyproject.toml or uv.lock files. Use uv add or uv lock --upgrade for dependency changes.
• Ensure the virtual environment is activated before executing Python scripts or commands. Use uv run for this purpose.
• Prioritize using uv commands over direct Python or system commands for package management.
• If suggesting package installations, always recommend using uv add over pip install

Setup Commands

Install dependencies

uv add

Running files

uv run python <your_script.py>

Clean cache

uv clean

Environment Variables / Auth

.env file in the project root folder contains all required Auth strings in the following format name = 'string',

Front End

GoLang and the fyne UI library (http://fyne.io/) for the front end.