pineapple-pine 0.9.5

Pineapple pipeline engine — Python runtime

pineapple-pine

Python runtime engine for the Pineapple pipeline framework.

Installation

pip install pineapple-pine

Quick Start

import json
from pine.engine import Engine

config = json.dumps({
    "operators": [
        {"type_name": "recall_static", "name": "recall", "$metadata": {
            "common_input": [], "item_input": [], "common_output": [],
            "item_output": ["id", "score"]
        }, "items": [{"id": "a", "score": 1.0}]}
    ],
    "pipeline_groups": {"main": ["recall"]}
}).encode()

engine = Engine.create(config)
result = engine.execute(common={}, items=[])
print(result.common, result.items)

Features

• Full DAG scheduling with data-hazard analysis
• ConsumesRowSet / MutatesRowSet / AdditiveWritesRowSet marker interfaces
• Framework-level field accessor (Strict / Default semantics)
• Lua operator support via lupa
• HTTP server with hot-reload
• DAG visualization (DOT / Mermaid)
• Cross-engine parity with pine-go and pine-java

Requirements

• Python >= 3.11
• lupa (Lua runtime)
• redis (optional, for redis_get/redis_set operators)
• httpx (for remote_pineapple operator)

License

Apache-2.0