graph.neighbors() (scikit-image RAG / networkx) yields np.int64 values, which were stored directly into id_neighbours. When serialized to CSV, this produced unreadable output like [np.int64(31), np.int64(552), ...] instead of [31, 552, ...].

Changes

• src/epitools/analysis/cell_statistics.py — In _calculate_graph_statistics(), convert each neighbour ID to a plain Python int at collection time; also simplifies the conditional append with a ternary:

  # Before
  id_neighbours.append(list(graph.neighbors(index)))
  
  # After
  new_id_neighbours = [int(n) for n in graph.neighbors(index)]
  id_neighbours.append(new_id_neighbours if new_id_neighbours else None)

• tests/test_cell_statistics.py — Adds test_id_neighbours_contains_plain_ints to assert type(n) is int for every value in every id_neighbours list, guarding against regressions from future scikit-image/networkx version changes.

• Fixes id neighbours showing integer tpe #154

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.