Fix: remap hyperedges in community-aggregated meta-graph view

[aadi-novice]

Summary

When a graph exceeds GRAPHIFY_VIZ_NODE_LIMIT (default 1000 via _viz_node_limit(), falling back to MAX_NODES_FOR_VIZ = 5000), to_html() builds a community-aggregated meta-graph and recursively calls itself. The recursive call never carried hyperedges onto the meta-graph, so the rendered graph.html always emits const hyperedges = []; — even when graph.json contains plenty.

The visible result: the afterDrawing handler that renders hyperedge regions silently no-ops on every large project, even though hyperedges are extracted, persisted to graph.json, and merged correctly through the rest of the pipeline.

Root Cause

In graphify/export.py — to_html at the over-limit branch builds a fresh meta-graph and recursively calls itself, but never sets meta.graph["hyperedges"]. Even if forwarded as-is, the hyperedges reference semantic node IDs (e.g., "UserService") that don't exist in the meta-graph (which uses community IDs like "3" as node IDs).

Fix

Remap hyperedges from semantic node IDs to community IDs before the recursive call. Drop entries that collapse to <2 distinct communities — they wouldn't render as a polygon anyway.

• Handles both "nodes" and "members" keys for backward compatibility with different hyperedge formats
• Deduplicates community IDs per hyperedge
• Falls back from label to relation for display text when label is missing
• Attaches remapped hyperedges to meta.graph["hyperedges"] so the recursive call picks them up

Testing

• All 37 existing tests in tests/test_export.py and tests/test_hypergraph.py pass

Fixes #1005

[safishamsi]

The fix is correct and surgical — remapping logic handles both key variants, orphaned nodes, and collapsed-to-one-community edges properly. Two small requests before merge:

1. Unique hyperedge IDs: he.get('id', '') produces duplicate empty-string IDs when multiple hyperedges lack an id field. A simple f'he_{i}' fallback would be safer for any downstream JS that keys on id.

2. Regression test: the broken path (graph over node limit + non-empty hyperedges) had no coverage — that's why it shipped silently. A unit test that creates a graph over the limit with hyperedges, calls to_html(), and asserts the output HTML contains non-empty hyperedge data would lock this in.

Otherwise looks good to merge.