大型社交網路的社群結構分析與視覺化探勘

import networkx as nx
import networkx.community as nxcom
import matplotlib.pyplot as plt
import collections
from pathlib import Path

# --- 數據載入與網路創建 
---
# 假設 'data' 目錄與腳本在同一層級，並且 'mcauley2012' 子目錄存在
# 為了範例運行，我們模擬一個大型網路，實際運行時請確保路徑正確
try:
    # 嘗試從實際路徑載入
    data_dir = Path('.') / 'data'
    file_path = data_dir / 'mcauley2012' / 'facebook_combined.txt'
    if file_path.exists():
        G_social_demo = nx.read_edgelist(file_path)
        print(f"成功載入線上社交網路數據: {file_path}")
    else:
        print(f"警告: 數據文件 '{file_path}' 不存在。創建一個模擬的大型網路。")
        # 創建一個模擬的大型網路，包含社群結構
        G_social_demo = nx.barabasi_albert_graph(n=1000, m=3, seed=42) # 使用 BA 模型生成一個具有無尺度特徵的網路
        # 為了增加社群感，可以手動添加一些密集連接的區域
        for i in range(5):
            subgraph_nodes = list(range(i*200, (i+1)*200))
            for u_idx in range(len(subgraph_nodes)):
                for v_idx in range(u_idx + 1, len(subgraph_nodes)):
                    if G_social_demo.number_of_edges() < 10000: # 限制邊的數量
                        G_social_demo.add_edge(subgraph_nodes[u_idx], subgraph_nodes[v_idx])
        print("已創建一個模擬的大型網路。")

except Exception as e:
    print(f"載入網路時發生錯誤: {e}。創建一個模擬的大型網路。")
    G_social_demo = nx.barabasi_albert_graph(n=1000, m=3, seed=42)
    for i in range(5):
        subgraph_nodes = list(range(i*200, (i+1)*200))
        for u_idx in range(len(subgraph_nodes)):
            for v_idx in range(u_idx + 1, len(subgraph_nodes)):
                if G_social_demo.number_of_edges() < 10000:
                    G_social_demo.add_edge(subgraph_nodes[u_idx], subgraph_nodes[v_idx])
    print("已創建一個模擬的大型網路。")

# --- 初步視覺化 (僅邊) 
---
print("\n--- 進行線上社交網路的初步視覺化 
---
")
if G_social_demo and G_social_demo.number_of_nodes() > 0:
    try:
        # 計算佈局
        pos_social = nx.spring_layout(G_social_demo, k=0.05, iterations=50, seed=42) # k 值較小以適應大型網路

        plt.figure(figsize=(15, 10))

        # 繪製邊
        nx.draw_networkx(
            G_social_demo, pos=pos_social, node_size=0, # 不繪製節點
            edge_color="#333333", alpha=0.05, with_labels=False # 稀疏、半透明的邊
        )

        plt.title("Online Social Network - Initial Edge Visualization", fontsize=18)
        plt.axis('off')
        plt.show()
        print("初步視覺化完成，展示了網路的整體結構和潛在的社群聚集。")
    except Exception as e:
        print(f"初步視覺化時發生錯誤: {e}")
else:
    print("網路物件為空或節點數不足，無法進行初步視覺化。")

# --- 社群偵測 
---
print("\n--- 對線上社交網路進行社群偵測 
---
")
if G_social_demo and G_social_demo.number_of_nodes() > 0:
    try:
        # 使用 greedy_modularity_communities 偵測社群
        # 對於大型網路，這可能需要一些時間
        communities_generator_social = nxcom.greedy_modularity_communities(G_social_demo, seed=42)

        # 將結果排序
        communities_social = sorted(communities_generator_social, key=len, reverse=True)

        # 輸出社群數量
        num_communities_social = len(communities_social)
        print(f"偵測到的社群數量: {num_communities_social}")

        # 輸出前幾個最大社群的大小
        print("前 5 個最大社群的大小:")
        for i, comm in enumerate(communities_social[:5]):
            print(f"  社群 {i+1}: {len(comm)} 個節點")

    except Exception as e:
        print(f"社群偵測失敗: {e}")
        communities_social = []
else:
    print("網路物件為空或節點數不足，無法進行社群偵測。")

# --- 視覺化輔助函數 (沿用前述定義) 
---

# 輔助函數：為節點設置社群屬性
def set_node_community(G, communities_list):
    '''為節點添加 'community' 屬性'''
    if not communities_list: return
    for c, nodes_in_community in enumerate(communities_list):
        for node in nodes_in_community:
            G.nodes[node]['community'] = c + 1

# 輔助函數：為邊設置社群屬性 (僅標記內部邊)
def set_edge_community(G):
    '''為社群內的邊添加 'community' 屬性'''
    for u, v in G.edges():
        if 'community' in G.nodes[u] and 'community' in G.nodes[v]:
            if G.nodes[u]['community'] == G.nodes[v]['community']:
                G.edges[u, v]['community'] = G.nodes[u]['community']
            else:
                G.edges[u, v]['community'] = 0 # 外部邊標記為 0
        else:
            pass # 節點無社群屬性時，邊也不標記

# 顏色映射函數 (使用 matplotlib 的 colormap)
def get_color(community_id, n_communities, low=0.1, high=0.9):
    if community_id == 0: # 外部邊的顏色
        return (0.5, 0.5, 0.5) # 灰色

    if n_communities <= 1: return (0.8, 0.2, 0.2) # 單一社群，預設顏色

    normalized_community_id = (community_id - 1) / (n_communities - 1)
    return plt.cm.viridis(normalized_community_id)

# --- 準備視覺化社群結構 
---
print("\n--- 準備視覺化線上社交網路的社群結構 
---
")
if communities_social and G_social_demo.number_of_nodes() > 0:
    try:
        # 1. 設置節點和邊的社群屬性
        set_node_community(G_social_demo, communities_social)
        set_edge_community(G_social_demo)

        # 2. 準備節點顏色
        n_communities_social = len(communities_social)
        node_colors_social = []
        for node in G_social_demo.nodes():
            community_id = G_social_demo.nodes[node].get('community', 0)
            node_colors_social.append(get_color(community_id, n_communities_social))

        # 3. 準備邊列表
        external_edges_social = []
        internal_edges_social = []
        internal_edge_colors_social = []

        for u, v, d in G_social_demo.edges(data=True):
            community_id = d.get('community', 0)
            if community_id == 0:
                external_edges_social.append((u, v))
            else:
                internal_edges_social.append((u, v))
                internal_edge_colors_social.append(get_color(community_id, n_communities_social))

        print("節點和邊的社群屬性已設置。")

    except Exception as e:
        print(f"準備視覺化數據時發生錯誤: {e}")
        node_colors_social, external_edges_social, internal_edges_social, internal_edge_colors_social = [], [], [], []
else:
    print("無法準備視覺化數據，因為社群偵測失敗或網路為空。")

# --- 繪製社群結構圖 
---
print("\n--- 繪製線上社交網路的社群結構圖 
---
")
if communities_social and G_social_demo.number_of_nodes() > 0 and len(node_colors_social) == G_social_demo.number_of_nodes():
    try:
        plt.figure(figsize=(18, 12)) # 增大圖形尺寸以容納更多節點

        # 繪製外部邊
        nx.draw_networkx(
            G_social_demo, pos=pos_social,
            edgelist=external_edges_social, edge_color="#333333", width=0.3, alpha=0.1, style='dashed',
            node_size=0
        )

        # 繪製節點和內部邊
        nx.draw_networkx(
            G_social_demo, pos=pos_social,
            node_color=node_colors_social, node_size=10, alpha=0.7, # 節點大小可以稍微調大一點
            edgelist=internal_edges_social, edge_color=internal_edge_colors_social, width=0.5, alpha=0.4
        )

        # 不繪製標籤以避免過於擁擠
        # nx.draw_networkx_labels(G_social_demo, pos_social, font_size=8)

        plt.title("Online Social Network - Visualized Communities", fontsize=20)
        plt.axis('off')
        plt.show()
        print("社群結構視覺化完成。")

    except Exception as e:
        print(f"繪製社群結構圖時發生錯誤: {e}")
else:
    print("無法繪製社群結構圖，因為數據準備不完整或網路為空。")