ingest/graph.py

import plotly.graph_objects as go

def build_cluster_graph(results, clusters):
    xs, ys, labels = [], [], []

    i = 0
    for agency, docs in clusters.items():
        for d in docs:
            xs.append(i)
            ys.append(len(docs))
            labels.append(f"{agency}: {d['title']}")
            i += 1

    fig = go.Figure(
        go.Scatter(
            x=xs,
            y=ys,
            mode="markers",
            text=labels,
            marker=dict(size=10),
        )
    )

    fig.update_layout(
        title="Semantic Cluster Graph (Click nodes to inspect documents)",
        xaxis_title="Document Index",
        yaxis_title="Cluster Density",
    )
    return fig