EMLN: Node type column naming is ambiguous across built-in datasets

[shainova]

This issue stems from #30

Problem

The type column in emln node tables is used inconsistently:

• Bipartite networks (e.g. 1, 2, 6, 14, 17, 21, 23–26, 28–30, 33, 35–40, 43–51, 54, 57–58, 60, 64–65, 69, 71–73, 78): type has exactly 2 values (e.g. plant|pollinator) representing bipartite set membership.
• Unipartite networks (e.g. 3–5, 7–12, 15, 18–20, 22, 27, 31–32, 34, 42, 47, 52, 55–56, 59, 61–63, 66–68, 70, 74–77): type has 1 value (taxon or organism) representing a taxonomic label — unrelated to bipartite structure.
• Ambiguous networks (13, 41, 53): type has 3–12 values representing trophic roles.

The MultilayerViz integration functions (multilayer_to_json, multilayer_to_csv) use type as a fallback for bipartite set membership, which is incorrect for unipartite networks.

Proposed fix

1. Rename type → node_type in all bipartite datasets (2-value type).
2. Rename type → taxon_group (or similar) in all non-bipartite datasets.
3. Regenerate all affected .rda files.
4. Remove the type fallback from multilayer_to_json and multilayer_to_csv — only node_type (and legacy node_group) would be accepted.

Affected datasets

• Bipartite (rename type → node_type): 1, 2, 6, 14, 17, 21, 23–26, 28–30, 33, 35–40, 43–51, 54, 57–58, 60, 64–65, 69, 71–73, 78
• Non-bipartite taxonomy (rename type → taxon_group): 3–5, 7–12, 15, 18–20, 22, 27, 31–32, 34, 42, 47, 52, 55–56, 59, 61–63, 66–68, 70, 74–77
• Trophic roles (decide separately): 13, 41, 53
• No type column (no action): 16