Weavelet

A library for managing function calls in specific order, to avoid code repeating. Similar to Apache Hamilton but simpler without 3rd party dependencies if not needed.

Features

• Simple DAG orchestration: Build execution graphs based on function signatures
• Decorator-based: Register functions as nodes using the @node() decorator
• Cross-module support: Nodes can be spread across multiple files
• Parameter injection: Automatic dependency resolution and parameter passing
• Result caching: Avoid recomputing expensive operations
• Pure Python: No external dependencies for core functionality

Quick Start

Installation

pip install weavelet

Basic Usage

Create nodes in any Python file:

# a_nodes.py
from weavelet import node

@node()
def auth_token(api_key: str) -> str:
    return f"Bearer {api_key}"

@node()
def user_profile(auth_token: str, user_id: int) -> dict:
    return {"id": user_id, "name": "Ala"}

@node()
def make_greeting(user_profile: dict) -> str:
    return f"Hello, {user_profile['name']}!"

Execute the DAG:

# main.py
from weavelet import Weave
import a_nodes

weave = Weave()
weave.include(a_nodes)
weave.params.add("api_key", "xyz")
weave.params.add("user_id", 128)

result = weave.run("make_greeting")  # -> "Hello, Ala!"
print(result)

How it Works

Weavelet automatically builds a DAG based on function signatures:

• Function parameters that match other node names become dependencies
• Parameters not matching node names are injected from weave.params
• Execution order is determined through topological sorting
• Results are cached to avoid redundant computation

Requirements

• Python >=3.12
• No external dependencies for core functionality

Development

Uses uv as package manager with TDD approach.