R中igraph绘图：如何排除或编辑单节点模块的边界

Absolutely! You can easily tweak igraph's plotting behavior to hide module boundaries for communities that only contain a single node (like your nodes 9 and 13). Here are a couple of straightforward approaches depending on your needs:

Approach 1: Filter Out Single-Node Communities Entirely #

If you don't need to mark single-node communities at all, just filter them out before passing them to the plotting function:

from igraph import Graph, plot

# Example setup (replace with your actual graph and community detection result)
g = Graph.Famous("Zachary")
communities = g.community_infomap()

# Keep only communities with more than one node
filtered_groups = [group for group in communities if len(group) > 1]

# Plot with only the filtered groups marked
plot(g, mark_groups=filtered_groups, bbox=(600, 600))

Approach 2: Hide Boundaries for Specific Single-Node Communities #

If you want to keep all community-based styling (like node colors) but just hide boundaries for specific nodes (e.g., 9 and 13), you can customize the mark styles per community. Note: igraph uses 0-indexed node IDs by default—if your nodes are numbered starting at 1, adjust the target IDs accordingly (e.g., use 8 and 12 instead of 9 and 13).

# Define the nodes whose community boundaries you want to hide
target_nodes = [9, 13]  # Adjust IDs if using 1-based indexing

# Get the community IDs for these target nodes
target_comm_ids = [communities.membership[node] for node in target_nodes]

# Create custom styles: hide boundaries for target communities, show others
mark_styles = []
for comm_idx in range(len(communities)):
    if comm_idx in target_comm_ids:
        # Disable the boundary entirely
        mark_styles.append({"mark_shape": None})
    else:
        # Keep default or custom boundary styling
        mark_styles.append({"mark_width": 2, "mark_color": "lightblue", "mark_alpha": 0.3})

# Plot with the custom styles
plot(g, mark_groups=communities, mark_style=mark_styles, bbox=(600, 600))

Bonus: Hide All Single-Node Community Boundaries #

If you want to hide boundaries for all single-node communities (not just specific ones), you can automate the style creation:

# Get the size of each community
comm_sizes = [len(group) for group in communities]

mark_styles = []
for size in comm_sizes:
    if size == 1:
        mark_styles.append({"mark_shape": None})
    else:
        mark_styles.append({"mark_width": 2, "mark_color": "lightgray"})

plot(g, mark_groups=communities, mark_style=mark_styles, bbox=(600, 600))

All these methods let you control exactly which community boundaries are displayed, so you can clean up your plot by hiding those unnecessary single-node boxes.

内容的提问来源于stack exchange，提问作者cannin