07 Knowledge Graph Extraction¶

Source: notebooks/07-knowledge-graph-extraction-graphistry-v0-1-1-e1.ipynb

Notebook focus¶

This page is a cell-by-cell walkthrough of the notebook, explaining the intent of each step and showing the exact code executed.

Cell-by-cell walkthrough¶

Cell 1 (Markdown)¶

07 Knowledge Graph Extraction¶

Build knowledge graphs from structured entity/relationship data using GraphistryBuilders.

What you will learn: - Model entities with id/label dicts - Define typed relationships (source, target, type) - Generate node and edge DataFrames for visualization

Requirements: llamatelemetry installed. No GPU required for graph building.

Cell 2 (Markdown)¶

1) Install¶

Cell 3 (Code)¶

Summary: Installs required dependencies and runtime tools.

!pip -q install git+https://github.com/llamatelemetry/llamatelemetry.git@v0.1.1

Cell 4 (Markdown)¶

2) Define a GPU infrastructure knowledge graph¶

Cell 5 (Code)¶

Summary: Imports core libraries: llamatelemetry. Sets up Graphistry for graph visualization or analytics.

from llamatelemetry.graphistry import GraphistryBuilders

entities = [
    {"id": "GPU0", "label": "GPU 0 (T4)"},
    {"id": "GPU1", "label": "GPU 1 (T4)"},
    {"id": "llama.cpp", "label": "llama.cpp server"},
    {"id": "nccl", "label": "NCCL"},
    {"id": "otel", "label": "OpenTelemetry"},
]

relationships = [
    {"source": "llama.cpp", "target": "GPU0", "type": "runs_on"},
    {"source": "llama.cpp", "target": "GPU1", "type": "splits_to"},
    {"source": "nccl", "target": "GPU0", "type": "communicates"},
    {"source": "nccl", "target": "GPU1", "type": "communicates"},
    {"source": "otel", "target": "llama.cpp", "type": "observes"},
]

nodes_df, edges_df = GraphistryBuilders.knowledge_graph(entities, relationships)

Cell 6 (Markdown)¶

3) Inspect the DataFrames¶

Cell 7 (Code)¶

Summary: Executes notebook-specific logic or data processing for this step.

print("=== Nodes ===")
print(nodes_df.to_string(index=False))
print(f"\n=== Edges ({len(edges_df)} relationships) ===")
print(edges_df.to_string(index=False))

Cell 8 (Markdown)¶

4) Visualize (optional)¶

Pass the DataFrames to Graphistry for interactive GPU-accelerated visualization.

Cell 9 (Code)¶

Summary: Sets up Graphistry for graph visualization or analytics.

# from llamatelemetry.graphistry import GraphistrySession
# session = GraphistrySession.from_kaggle_secrets()
# g = session.connector.create_graph(edges_df, nodes_df=nodes_df)
# g.plot()