sayou-assembler 0.1.2

Assembler components for the Sayou Data Platform

sayou-assembler

sayou-assembler is the graph construction engine of the Sayou Data Platform. It takes standardized nodes from sayou-wrapper and assembles them into a logical In-Memory Knowledge Graph.

It is responsible for interpreting relationships (like Parent-Child), generating reverse edges, and ensuring the graph's topological integrity before it is persisted by sayou-loader.

Philosophy

"Structure before Storage." A database is just a container. The value of a Knowledge Graph comes from its structure. sayou-assembler focuses on the logic of connecting dots (Nodes) to create meaning (Context), independent of the underlying database technology (Neo4j, SQL, NetworkX).

🚀 Key Features

• In-Memory Graph Model: Manages GraphNode and GraphEdge objects efficiently in Python.
• Hierarchy Strategy: Automatically builds tree-like structures using parent_id attributes.
• Bi-directional Linking: Automatically generates reverse edges (e.g., hasChild from hasParent) to enable bidirectional graph traversal.
• Strategy Pattern: Supports pluggable assembly strategies (Hierarchy, Semantic, Chronological).

📦 Installation

pip install sayou-assembler

⚡ Quickstart

The AssemblerPipeline creates the graph using a specified strategy (default: hierarchy).

import json
from sayou.assembler.pipeline import AssemblerPipeline

def run_demo():
    # 1. Load Standard Nodes (Output from sayou-wrapper)
    input_file = "wrapped_nodes.json"
    with open(input_file, "r", encoding="utf-8") as f:
        wrapper_output = json.load(f)

    # 2. Initialize Pipeline
    # Uses 'HierarchyBuilder' to link parent-child nodes
    pipeline = AssemblerPipeline(strategy="hierarchy")
    pipeline.initialize()

    # 3. Run Assembly
    kg_result = pipeline.run(wrapper_output)

    # 4. Inspect Graph
    print(f"Nodes: {kg_result['summary']['node_count']}")
    print(f"Edges: {kg_result['summary']['edge_count']}")
    
    for edge in kg_result['edges'][:3]:
        print(f"{edge['source']} --[{edge['type']}]--> {edge['target']}")

if __name__ == "__main__":
    run_demo()

🤝 Contributing

Contributions for new assembly strategies (e.g., constructing graphs based on keyword co-occurrence) are welcome.

📜 License

Apache 2.0 License © 2025 Sayouzone