\u5468\u4e94\u4e0b\u5348，\u6280\u672f\u90e8\u4f1a\u8bae\u5ba4\u3002<\/p>\n
"\u8fd9\u6279\u65b0\u5939\u5177\u4ece\u8bbe\u8ba1\u5230\u9996\u4ef6\u5408\u683c，\u8d70\u4e86 23 \u5929，"\u9879\u76ee\u7ecf\u7406\u8001\u5218\u628a\u7518\u7279\u56fe\u6295\u5f71\u5728\u5899\u4e0a，"\u8bbe\u8ba1 5 \u5929，\u5de5\u827a 4 \u5929，\u52a0\u5de5 8 \u5929，\u8d28\u68c0 6 \u5929\u2014\u2014\u770b\u8d77\u6765\u5404\u5e72\u5404\u7684，\u4e32\u884c\u6392\u961f\u3002\u53ef\u5b9e\u9645\u4e0a\u8bbe\u8ba1\u6539\u4e86\u7b2c\u4e09\u7248\u7684\u65f6\u5019\u5de5\u827a\u5c31\u5df2\u7ecf\u5728\u5ba1\u7b2c\u4e8c\u7248\u4e86，\u52a0\u5de5\u90a3\u8fb9\u63d0\u524d\u770b\u4e86\u4e09\u7ef4\u6a21\u578b\u5728\u51c6\u5907\u6bdb\u576f，\u8d28\u68c0\u4e5f\u63d0\u524d\u4ecb\u5165\u4e86\u516c\u5dee\u8bc4\u5ba1\u3002\u95ee\u9898\u662f\u8fd9\u4e9b'\u5e76\u884c'\u5168\u9760\u5fae\u4fe1\u7fa4\u91cc\u543c，\u6ca1\u4eba\u8bf4\u5f97\u6e05\u5230\u5e95\u8c01\u548c\u8c01\u5728\u534f\u540c\u3001\u534f\u540c\u5f3a\u5ea6\u591a\u5927\u3001\u54ea\u6761\u94fe\u8def\u662f\u74f6\u9888\u3002"<\/p>\n
\u6211\u7ffb\u4e86\u7ffb\u90ae\u4ef6\u548c\u5fae\u4fe1\u7fa4\u8bb0\u5f55\u3002<\/p>\n
"\u8fd9\u6709\u4ec0\u4e48\u95ee\u9898？"\u8001\u5218\u95ee\u3002<\/p>\n
"\u95ee\u9898\u662f\u534f\u540c\u662f\u9690\u6027\u7684，"\u6211\u8bf4，"\u8bbe\u8ba1\u6539\u4e86\u5c3a\u5bf8，\u5de5\u827a\u8981\u8ddf\u7740\u6539\u5939\u5177\u65b9\u6848，\u52a0\u5de5\u8981\u91cd\u65b0\u9009\u5200\u8def，\u8d28\u68c0\u8981\u91cd\u65b0\u5b9a\u68c0\u5177\u3002\u8fd9\u6761'\u53d8\u66f4\u4f20\u64ad\u94fe'\u4f60\u770b\u5f97\u89c1\u5417？\u54ea\u4e2a\u5de5\u7a0b\u5e08\u8ddf\u54ea\u4e2a\u5de5\u7a0b\u5e08\u4ea4\u4e92\u6700\u591a？\u54ea\u6b21\u53d8\u66f4\u89e6\u53d1\u4e86\u6700\u591a\u7684\u4e0b\u6e38\u8fd4\u5de5？"<\/p>\n
"\u6211\u5c31\u60f3\u5e72\u4e00\u4ef6\u4e8b，"\u8001\u5218\u8bf4，"\u628a\u8bbe\u8ba1\u3001\u5de5\u827a\u3001\u52a0\u5de5\u3001\u8d28\u68c0\u56db\u4e2a\u89d2\u8272\u7684\u534f\u540c\u5173\u7cfb\u753b\u51fa\u6765\u2014\u2014\u8c01\u8ddf\u8c01\u6c9f\u901a\u6700\u591a\u3001\u6bcf\u6b21\u6c9f\u901a\u6d89\u53ca\u54ea\u4e9b\u53d8\u66f4\u3001\u53d8\u66f4\u4f20\u64ad\u4e86\u51e0\u8df3\u624d\u7a33\u5b9a\u3002\u6211\u60f3\u77e5\u9053\u5e76\u884c\u5de5\u7a0b\u91cc\u5230\u5e95\u54ea\u6761\u534f\u540c\u94fe\u8def\u6700\u6d3b\u8dc3\u3001\u54ea\u4e2a\u4eba\u662f\u67a2\u7ebd\u8282\u70b9\u3002"<\/p>\n
"\u6bd4\u5982\u8bbe\u8ba1\u6539\u4e86\u5b9a\u4f4d\u9500\u5b54\u5f84，\u5de5\u827a\u8981\u6539\u5b9a\u4f4d\u65b9\u6848，\u52a0\u5de5\u8981\u6362\u94bb\u5934，\u8d28\u68c0\u8981\u6362\u585e\u89c4，"\u6211\u63a5\u8bdd，"\u8fd9\u5c31\u662f\u4e00\u6761\u4f20\u64ad\u94fe：\u8bbe\u8ba1\u2192\u5de5\u827a\u2192\u52a0\u5de5\u2192\u8d28\u68c0，\u56db\u8df3\u3002\u5982\u679c\u8bbe\u8ba1\u6539\u4e86\u4e09\u6b21\u624d\u5b9a\u7a3f，\u6bcf\u6b21\u90fd\u5168\u94fe\u8def\u4f20\u64ad，\u90a3\u5c31\u662f 12 \u6b21\u4ea4\u4e92\u3002\u628a\u8fd9\u4e9b\u4ea4\u4e92\u9891\u6b21\u5efa\u6210\u7f51\u7edc\u56fe，\u8c01\u7684\u5ea6\u6570\u6700\u9ad8\u8c01\u5c31\u662f\u67a2\u7ebd\u3002"<\/p>\n
"\u5bf9，"\u8001\u5218\u70b9\u5934，"\u8fd8\u6709，\u6211\u60f3\u770b'\u53d8\u66f4\u4f20\u64ad\u6df1\u5ea6'\u2014\u2014\u6709\u7684\u53d8\u66f4\u53ea\u5f71\u54cd\u5de5\u827a，\u6709\u7684\u53d8\u66f4\u4e00\u8def\u4f20\u5230\u8d28\u68c0\u3002\u4f20\u64ad\u8d8a\u6df1\u7684\u53d8\u66f4，\u8bf4\u660e\u524d\u671f\u8bbe\u8ba1\u8bc4\u5ba1\u6ca1\u505a\u5230\u4f4d\u3002\u6700\u597d\u8fd8\u80fd\u805a\u7c7b，\u770b\u54ea\u4e9b\u5de5\u7a0b\u5e08\u662f\u4e00\u4e2a'\u534f\u540c\u793e\u533a'\u3002"<\/p>\n
"\u7528 networkx \u5efa\u56db\u89d2\u8272\u534f\u540c\u7f51\u7edc，\u8282\u70b9\u662f\u5de5\u7a0b\u5e08，\u8fb9\u662f\u534f\u540c\u4ea4\u4e92\u9891\u6b21，\u8fb9\u6743=\u4ea4\u4e92\u6b21\u6570，\u8282\u70b9\u5927\u5c0f=\u5ea6\u6570\u4e2d\u5fc3\u6027，"\u6211\u5f00\u5de5\u7a0b，"pandas \u7ba1\u534f\u540c\u65e5\u5fd7，numpy \u7b97\u4e2d\u5fc3\u6027\u6307\u6807，scipy \u505a\u793e\u533a\u53d1\u73b0\u540e\u7684\u6a21\u5757\u5ea6\u68c0\u9a8c，scikit-learn \u5bf9\u5de5\u7a0b\u5e08\u505a\u805a\u7c7b\u5206\u7fa4，matplotlib \u753b\u7f51\u7edc\u56fe+\u4f20\u64ad\u6df1\u5ea6\u76f4\u65b9\u56fe\u3002\u6570\u636e\u81ea\u5305\u542b，\u5408\u6210\u4e00\u6279\u5e76\u884c\u5de5\u7a0b\u534f\u540c\u8bb0\u5f55，\u4e0b\u8f7d\u5c31\u80fd\u8dd1\u3002"<\/p>\n
\u6572\u4e86\u884c\u539f\u578b：<\/p>\n
G.add_edge("\u8bbe\u8ba1-\u5f20\u4e09", "\u5de5\u827a-\u674e\u56db", weight=12, changes=["\u5b54\u5f84\u53d8\u66f4","\u516c\u5dee\u8c03\u6574"])<\/p>\n
centrality = nx.degree_centrality(G)<\/p>\n
communities = nx.community.greedy_modularity_communities(G)<\/p>\n
\u00a0<\/p>\n
"\u5b8c\u6574\u7248\u7528 OOP \u5c01\u597d，"\u6211\u8bf4，"\u4e00\u4e2a\u7c7b\u7ba1\u534f\u540c\u65e5\u5fd7\u52a0\u8f7d，\u4e00\u4e2a\u7c7b\u5efa\u7f51\u7edc，\u4e00\u4e2a\u7c7b\u7b97\u4e2d\u5fc3\u6027，\u4e00\u4e2a\u7c7b\u505a\u793e\u533a\u53d1\u73b0，\u4e00\u4e2a\u7c7b\u5206\u6790\u53d8\u66f4\u4f20\u64ad，\u4e00\u4e2a\u7c7b\u51fa\u56fe\u3002\u8f93\u51fa\u534f\u540c\u7f51\u7edc\u56fe\u3001\u67a2\u7ebd\u8282\u70b9\u6392\u540d\u3001\u793e\u533a\u5212\u5206\u3001\u53d8\u66f4\u4f20\u64ad\u6df1\u5ea6\u5206\u5e03，\u5b58 results\/\u3002"<\/p>\n
\u8001\u5218\u51d1\u8fd1\u770b："\u90a3\u4ee5\u540e\u770b\u62a5\u544a：\u5de5\u827a-\u674e\u56db\u662f\u67a2\u7ebd（\u5ea6\u6570 8，\u4ecb\u6570\u4e2d\u5fc3\u6027\u6700\u9ad8），\u8bbe\u8ba1-\u5f20\u4e09\u548c\u4f20\u64ad\u6df1\u5ea6\u6700\u6df1\u7684\u53d8\u66f4\u94fe（\u5e73\u5747 3.2 \u8df3），\u56db\u4e2a\u5de5\u7a0b\u5e08\u5f62\u6210\u4e24\u4e2a\u793e\u533a（\u8bbe\u8ba1\u548c\u5de5\u827a\u4e00\u7ec4\u3001\u52a0\u5de5\u548c\u8d28\u68c0\u4e00\u7ec4），\u6a21\u5757\u5ea6 0.42\u3002\u8bf4\u660e\u8bbe\u8ba1\u548c\u5de5\u827a\u9ecf\u6027\u591f，\u4f46\u52a0\u5de5\u548c\u8d28\u68c0\u4e4b\u95f4\u534f\u540c\u504f\u5f31，\u53d8\u66f4\u5230\u52a0\u5de5\u5c31\u65ad\u4e86，\u8d28\u68c0\u53ea\u80fd\u88ab\u52a8\u7b49\u7ed3\u679c\u3002"<\/p>\n
"\u5bf9，"\u6211\u63a5\u8bdd，"\u5e76\u884c\u5de5\u7a0b\u4e0d\u662f'\u540c\u65f6\u5f00\u5de5'\u5c31\u5b8c\u4e86，\u662f\u4fe1\u606f\u6d41\u8981\u901a\u7545\u3002\u7f51\u7edc\u56fe\u5c31\u662f\u4fe1\u606f\u6d41\u7684 X \u5149\u7247\u3002\u6570\u5b57\u5b6a\u751f\u91cc\u534f\u540c\u8bbe\u8ba1\u5e73\u53f0\u8981\u505a\u6d41\u7a0b\u4f18\u5316，\u8fd9\u5f20\u7f51\u7edc\u5c31\u662f\u73b0\u72b6\u57fa\u7ebf\u3002"<\/p>\n
\u00a0<\/p>\n
\u4e00\u3001\u5b9e\u9645\u5e94\u7528\u573a\u666f（\u771f\u5b9e\u75db\u70b9）<\/p>\n
\u00a0<\/p>\n
\u573a\u666f\u8bbe\u5b9a：\u591a\u54c1\u79cd\u5de5\u88c5\u5939\u5177\u5f00\u53d1\u9879\u76ee，\u91c7\u7528\u5e76\u884c\u5de5\u7a0b\u6a21\u5f0f，\u8bbe\u8ba1\/\u5de5\u827a\/\u52a0\u5de5\/\u8d28\u68c0\u56db\u89d2\u8272\u540c\u6b65\u63a8\u8fdb\u3002\u534f\u540c\u4f9d\u8d56\u5fae\u4fe1\u7fa4\u3001\u90ae\u4ef6\u3001\u53e3\u5934\u6c9f\u901a，\u65e0\u7ed3\u6784\u5316\u8bb0\u5f55\u3002\u9879\u76ee\u7ecf\u7406\u65e0\u6cd5\u91cf\u5316\u534f\u540c\u5f3a\u5ea6\u3001\u8bc6\u522b\u4fe1\u606f\u5b64\u5c9b\u3001\u5b9a\u4f4d\u53d8\u66f4\u4f20\u64ad\u74f6\u9888\u3002<\/p>\n
\u00a0<\/p>\n
\u73b0\u573a\u539f\u8bdd（\u53d9\u4e8b\u5316）：<\/p>\n
\u00a0<\/p>\n
"\u4e0d\u662f\u6211\u4eec\u4e0d\u5e76\u884c，"\u8001\u5218\u8bf4，"\u662f\u5e76\u884c\u4e86\u4f46\u770b\u4e0d\u89c1\u3002\u8bbe\u8ba1\u6539\u4e86\u4e2a\u5c3a\u5bf8，\u5de5\u827a\u90a3\u8fb9\u4e0d\u77e5\u9053，\u7167\u6837\u6309\u65e7\u56fe\u505a\u65b9\u6848，\u7b49\u53d1\u73b0\u4e86\u5df2\u7ecf\u767d\u5e72\u4e86\u4e24\u5929\u3002\u540e\u6765\u6211\u4eec\u89c4\u5b9a'\u8bbe\u8ba1\u53d8\u66f4\u5fc5\u987b\u901a\u77e5\u5de5\u827a\u548c\u52a0\u5de5'，\u53ef\u901a\u77e5\u4e86\u6ca1\u8ddf\u8e2a，\u8d28\u68c0\u90a3\u8fb9\u8fd8\u662f\u6700\u540e\u624d\u77e5\u9053\u3002\u6211\u60f3\u77e5\u9053\u5230\u5e95\u8c01\u8ddf\u8c01\u534f\u540c\u6700\u5bc6\u3001\u54ea\u6b21\u53d8\u66f4\u4f20\u64ad\u6700\u8fdc\u3001\u54ea\u4e2a\u4eba\u662f\u4fe1\u606f\u9ed1\u6d1e\u3002"<\/p>\n
"\u8fd8\u6709\u793e\u533a\u95ee\u9898，"\u8001\u5218\u8865\u5145，"\u8bbe\u8ba1\u4e09\u4e2a\u4eba\u3001\u5de5\u827a\u4e24\u4e2a\u3001\u52a0\u5de5\u4e09\u4e2a\u3001\u8d28\u68c0\u4e24\u4e2a，\u4e00\u5171\u5341\u4e2a\u4eba\u3002\u6211\u611f\u89c9\u8bbe\u8ba1\u548c\u5de5\u827a\u662f\u4e00\u4f19\u7684，\u52a0\u5de5\u81ea\u5df1\u73a9，\u8d28\u68c0\u6700\u540e\u624d\u8fdb\u6765\u3002\u53ef\u8fd9\u662f\u611f\u89c9，\u6211\u8981\u6570\u636e\u8bc1\u660e，\u624d\u80fd\u8c03\u6574\u4eba\u5458\u642d\u914d\u3002"<\/p>\n
\u6838\u5fc3\u77db\u76fe："\u9690\u5f0f\u534f\u540c\u4ea4\u4e92\u6d41\u6c34" \u4e0e "\u534f\u540c\u7f51\u7edc\u62d3\u6251 + \u4e2d\u5fc3\u6027\u5206\u6790 + \u793e\u533a\u53d1\u73b0 + \u53d8\u66f4\u4f20\u64ad\u6df1\u5ea6 + \u4fe1\u606f\u5b64\u5c9b\u5b9a\u4f4d" \u4e4b\u95f4\u7684\u65ad\u5c42\u3002<\/p>\n
\u00a0<\/p>\n
\u4e8c\u3001\u75db\u70b9\u5206\u6790（\u6620\u5c04\u5230\u6ee8\u5dde\u804c\u4e1a\u5b66\u9662\u300a\u5148\u8fdb\u5236\u9020\u6280\u672f\u300b\u8bfe\u7a0b\u6a21\u578b）<\/p>\n
\u00a0<\/p>\n
\u300a\u5148\u8fdb\u5236\u9020\u6280\u672f\u300b\u6a21\u5757 \u672c\u7bc7\u75db\u70b9\u5bf9\u5e94<\/p>\n
\u5148\u8fdb\u5236\u9020\u65b0\u6a21\u5f0f：\u5e76\u884c\u5de5\u7a0b(CE)\u3001\u534f\u540c\u8bbe\u8ba1\u3001DFX \u591a\u89d2\u8272\u534f\u540c\u7f51\u7edc\u5efa\u6a21\u4e0e\u4f18\u5316<\/p>\n
\u5148\u8fdb\u5236\u9020\u6280\u672f\u57fa\u7840：\u7cfb\u7edf\u5de5\u7a0b\u3001\u4fe1\u606f\u6d41\u5206\u6790 \u534f\u540c\u4ea4\u4e92\u7684\u7ed3\u6784\u5316\u91cf\u5316<\/p>\n
CAD\/CAM\u6280\u672f：\u8bbe\u8ba1-\u5de5\u827a-\u5236\u9020\u4fe1\u606f\u96c6\u6210 \u53d8\u66f4\u4f20\u64ad\u94fe\u5206\u6790<\/p>\n
\u667a\u80fd\u5236\u9020\u4e0e\u6570\u5b57\u5b6a\u751f：\u534f\u540c\u5e73\u53f0\u4fe1\u606f\u6d41\u4eff\u771f \u7f51\u7edc\u62d3\u6251\u4f5c\u4e3a\u6d41\u7a0b\u4f18\u5316\u57fa\u7ebf<\/p>\n
FMS\u4e0e\u5148\u8fdb\u751f\u4ea7\u7ba1\u7406：\u9879\u76ee\u7ba1\u7406\u3001\u5e76\u884c\u4efb\u52a1\u8c03\u5ea6 \u67a2\u7ebd\u8282\u70b9\u8bc6\u522b\u4e0e\u8d44\u6e90\u8c03\u914d<\/p>\n
\u00a0<\/p>\n
\u4e00\u53e5\u8bdd\u603b\u7ed3：\u6211\u4eec\u9700\u8981\u4e00\u4e2a"\u5e76\u884c\u5de5\u7a0b\u591a\u89d2\u8272\u534f\u540c\u7f51\u7edc\u5206\u6790\u7a0b\u5e8f"，\u7528\u00a0<\/p>\n
"networkx" \u5efa\u56fe+\u4e2d\u5fc3\u6027+\u793e\u533a\u53d1\u73b0，<\/p>\n
"pandas" \u7ba1\u534f\u540c\u65e5\u5fd7，<\/p>\n
"numpy" \u7b97\u4f20\u64ad\u6df1\u5ea6，<\/p>\n
"scipy" \u505a\u6a21\u5757\u5ea6\u8bc4\u4f30，<\/p>\n
"scikit-learn" \u805a\u7c7b\u5de5\u7a0b\u5e08\u5206\u7fa4，\u5b9e\u73b0\u4ece"\u5fae\u4fe1\u7fa4\u543c"\u5230"\u534f\u540c\u7f51\u7edc\u62d3\u6251 + \u67a2\u7ebd\u5b9a\u4f4d + \u4fe1\u606f\u5b64\u5c9b\u8bc6\u522b"\u3002<\/p>\n
\u00a0<\/p>\n
\u4e09\u3001\u6838\u5fc3\u903b\u8f91\u8bb2\u89e3（\u5927\u767d\u8bdd）<\/p>\n
\u00a0<\/p>\n
3.1 \u95ee\u9898\u672c\u8d28：\u628a\u534f\u540c\u60f3\u6210"\u670b\u53cb\u5708"<\/p>\n
\u00a0<\/p>\n
\u628a\u5e76\u884c\u5de5\u7a0b\u56e2\u961f\u60f3\u6210\u4e00\u4e2a\u5fae\u4fe1\u7fa4：<\/p>\n
\u00a0<\/p>\n
* \u8282\u70b9 = \u7fa4\u6210\u5458（\u8bbe\u8ba1\u5f20\u4e09\u3001\u5de5\u827a\u674e\u56db\u3001\u52a0\u5de5\u738b\u4e94\u3001\u8d28\u68c0\u8d75\u516d\u2026）<\/p>\n
* \u8fb9 = \u804a\u5929\u9891\u6b21（\u5f20\u4e09@\u674e\u56db 12 \u6b21 = \u8fb9\u6743 12）<\/p>\n
* \u5ea6\u6570 = \u8c01\u6700\u6d3b\u8dc3（\u88ab @ \u6700\u591a\u7684\u4eba = \u67a2\u7ebd）<\/p>\n
* \u4ecb\u6570\u4e2d\u5fc3\u6027 = \u8c01\u662f\u4e0d\u53ef\u66ff\u4ee3\u7684\u4e2d\u95f4\u4eba（\u7ed5\u8fc7\u4ed6\u5c31\u65ad\u8054）<\/p>\n
* \u793e\u533a = \u8c01\u8ddf\u8c01\u662f\u4e00\u4f19\u7684（\u8bbe\u8ba1+\u5de5\u827a\u9ecf\u4e00\u8d77，\u52a0\u5de5+\u8d28\u68c0\u5404\u73a9\u5404）<\/p>\n
* \u53d8\u66f4\u4f20\u64ad = \u4e00\u6761\u6d88\u606f\u4ece\u8bbe\u8ba1\u53d1\u51fa，\u7ecf\u8fc7\u51e0\u8df3\u624d\u5230\u8d28\u68c0（\u8df3\u6570=\u4f20\u64ad\u6df1\u5ea6）<\/p>\n
* \u4fe1\u606f\u5b64\u5c9b = \u5ea6\u6570\u4f4e+\u4ecb\u6570\u4f4e+\u4e0d\u5728\u4e3b\u793e\u533a\u7684\u8282\u70b9（\u88ab\u9057\u5fd8\u7684\u4eba）<\/p>\n
\u00a0<\/p>\n
3.2 \u4e1a\u52a1\u903b\u8f91 \u2192 \u4ee3\u7801\u6620\u5c04<\/p>\n
\u00a0<\/p>\n
\u5bfc\u5165\u534f\u540c\u4ea4\u4e92\u65e5\u5fd7<\/p>\n
\u00a0 \u00a0\u2502<\/p>\n
\u00a0 \u00a0\u25bc CollabLoader (pandas)<\/p>\n
\u8bfb\u53d6 CSV：<\/p>\n
\u00a0 source, target, role_source, role_target,<\/p>\n
\u00a0 interaction_count, change_type, change_depth<\/p>\n
\u00a0 \u6821\u9a8c\u89d2\u8272\u5408\u6cd5(\u8bbe\u8ba1\/\u5de5\u827a\/\u52a0\u5de5\/\u8d28\u68c0)<\/p>\n
\u00a0 \u00a0\u2502<\/p>\n
\u00a0 \u00a0\u25bc CollabNetworkBuilder (networkx)<\/p>\n
\u5efa\u56fe：<\/p>\n
\u00a0 \u8282\u70b9=\u5de5\u7a0b\u5e08，\u8fb9=\u4ea4\u4e92\u9891\u6b21<\/p>\n
\u00a0 \u8282\u70b9\u5c5e\u6027=\u89d2\u8272，\u8fb9\u5c5e\u6027=\u53d8\u66f4\u7c7b\u578b+\u4f20\u64ad\u6df1\u5ea6<\/p>\n
\u00a0 \u00a0\u2502<\/p>\n
\u00a0 \u00a0\u25bc CentralityAnalyzer (networkx\/numpy)<\/p>\n
\u4e2d\u5fc3\u6027\u8ba1\u7b97：<\/p>\n
\u00a0 \u5ea6\u6570\u4e2d\u5fc3\u6027(\u8c01\u6700\u6d3b\u8dc3)<\/p>\n
\u00a0 \u4ecb\u6570\u4e2d\u5fc3\u6027(\u8c01\u662f\u4e0d\u53ef\u66ff\u4ee3\u7684\u6865)<\/p>\n
\u00a0 \u63a5\u8fd1\u4e2d\u5fc3\u6027(\u8c01\u79bb\u6240\u6709\u4eba\u6700\u8fd1)<\/p>\n
\u00a0 \u7279\u5f81\u5411\u91cf\u4e2d\u5fc3\u6027(\u8c01\u8ddf\u91cd\u8981\u7684\u4eba\u8fde\u63a5)<\/p>\n
\u00a0 \u00a0\u2502<\/p>\n
\u00a0 \u00a0\u25bc CommunityDetector (networkx)<\/p>\n
\u793e\u533a\u53d1\u73b0：<\/p>\n
\u00a0 greedy_modularity_communities<\/p>\n
\u00a0 \u6a21\u5757\u5ea6\u8bc4\u4f30(scipy)<\/p>\n
\u00a0 \u00a0\u2502<\/p>\n
\u00a0 \u00a0\u25bc PropagationAnalyzer<\/p>\n
\u53d8\u66f4\u4f20\u64ad\u5206\u6790：<\/p>\n
\u00a0 \u6309\u53d8\u66f4\u7c7b\u578b\u7edf\u8ba1\u4f20\u64ad\u6df1\u5ea6\u5206\u5e03<\/p>\n
\u00a0 \u627e\u6700\u6df1\u4f20\u64ad\u94fe<\/p>\n
\u00a0 \u00a0\u2502<\/p>\n
\u00a0 \u00a0\u25bc CollabVisualizer (matplotlib)<\/p>\n
\u53ef\u89c6\u5316：<\/p>\n
\u00a0 1. \u534f\u540c\u7f51\u7edc\u62d3\u6251\u56fe(\u8282\u70b9\u5927\u5c0f=\u5ea6\u6570, \u989c\u8272=\u89d2\u8272, \u8fb9\u5bbd=\u9891\u6b21)<\/p>\n
\u00a0 2. \u4e2d\u5fc3\u6027\u96f7\u8fbe\u56fe(\u67a2\u7ebd\u8282\u70b9\u5bf9\u6bd4)<\/p>\n
\u00a0 3. \u793e\u533a\u5206\u5e03\u6761\u5f62\u56fe<\/p>\n
\u00a0 4. \u53d8\u66f4\u4f20\u64ad\u6df1\u5ea6\u76f4\u65b9\u56fe<\/p>\n
\u00a0 5. \u89d2\u8272\u95f4\u4ea4\u4e92\u70ed\u529b\u56fe<\/p>\n
\u00a0 \u00a0\u2502<\/p>\n
\u00a0 \u00a0\u25bc SyntheticCollabGenerator (numpy)<\/p>\n
\u5408\u6210\u6570\u636e：<\/p>\n
\u00a0 10\u5de5\u7a0b\u5e08\u00d74\u89d2\u8272，\u4ea4\u4e92\u9891\u6b21\u5e42\u5f8b\u5206\u5e03<\/p>\n
\u00a0 \u542b\u67a2\u7ebd\u8282\u70b9+\u4fe1\u606f\u5b64\u5c9b<\/p>\n
\u00a0 \u53d8\u66f4\u4f20\u64ad\u6df1\u5ea6 1~4 \u8df3<\/p>\n
\u00a0<\/p>\n
3.3 \u4e3a\u4ec0\u4e48\u9700\u8981\u7f51\u7edc\u5206\u6790<\/p>\n
\u00a0<\/p>\n
\u65b9\u6cd5 \u95ee\u9898<\/p>\n
\u6570\u90ae\u4ef6\/\u5fae\u4fe1 \u53ea\u80fd\u770b\u81ea\u5df1\u53c2\u4e0e\u7684，\u770b\u4e0d\u5230\u5168\u5c40<\/p>\n
\u753b\u7ec4\u7ec7\u67b6\u6784\u56fe \u662f\u6c47\u62a5\u5173\u7cfb，\u4e0d\u662f\u534f\u540c\u5173\u7cfb<\/p>\n
\u7f51\u7edc\u62d3\u6251 \u771f\u5b9e\u4fe1\u606f\u6d41，\u8c01\u8ddf\u8c01\u771f\u5728\u534f\u4f5c<\/p>\n
\u00a0<\/p>\n
3.4 \u5206\u6790\u524d\u540e\u5bf9\u6bd4<\/p>\n
\u00a0<\/p>\n
\u7ef4\u5ea6 \u73b0\u72b6 \u672c\u7a0b\u5e8f<\/p>\n
\u534f\u540c\u5f3a\u5ea6 \u611f\u89c9 \u8fb9\u6743\u5b9a\u91cf<\/p>\n
\u67a2\u7ebd\u4eba\u7269 \u9886\u5bfc\u6307\u5b9a \u4e2d\u5fc3\u6027\u6392\u540d<\/p>\n
\u4fe1\u606f\u5b64\u5c9b \u6ca1\u4eba\u63d0 \u4f4e\u5ea6\u6570+\u4f4e\u4ecb\u6570\u81ea\u52a8\u8bc6\u522b<\/p>\n
\u793e\u533a\u5212\u5206 \u51ed\u7ecf\u9a8c \u6a21\u5757\u5ea6\u4f18\u5316\u81ea\u52a8\u805a\u7c7b<\/p>\n
\u53d8\u66f4\u4f20\u64ad \u65e0\u8bb0\u5f55 \u4f20\u64ad\u6df1\u5ea6\u5206\u5e03+\u6700\u957f\u94fe<\/p>\n
\u00a0<\/p>\n
\u56db\u3001OOP \u4ee3\u7801\u5b9e\u73b0<\/p>\n
\u00a0<\/p>\n
4.1 \u9879\u76ee\u7ed3\u6784<\/p>\n
\u00a0<\/p>\n
parallel_collab_network\/<\/p>\n
\u251c\u2500\u2500 parallel_collab_network\/<\/p>\n
\u2502 \u251c\u2500\u2500 __init__.py<\/p>\n
\u2502 \u251c\u2500\u2500 collab_loader.py # \u534f\u540c\u65e5\u5fd7\u52a0\u8f7d<\/p>\n
\u2502 \u251c\u2500\u2500 network_builder.py # \u7f51\u7edc\u6784\u5efa<\/p>\n
\u2502 \u251c\u2500\u2500 centrality_analyzer.py # \u4e2d\u5fc3\u6027\u5206\u6790<\/p>\n
\u2502 \u251c\u2500\u2500 community_detector.py # \u793e\u533a\u53d1\u73b0<\/p>\n
\u2502 \u251c\u2500\u2500 propagation_analyzer.py # \u53d8\u66f4\u4f20\u64ad\u5206\u6790<\/p>\n
\u2502 \u251c\u2500\u2500 visualizer.py # \u53ef\u89c6\u5316<\/p>\n
\u2502 \u2514\u2500\u2500 synthetic_data.py # \u5408\u6210\u6570\u636e<\/p>\n
\u251c\u2500\u2500 tests\/<\/p>\n
\u2502 \u251c\u2500\u2500 __init__.py<\/p>\n
\u2502 \u2514\u2500\u2500 test_collab_network.py<\/p>\n
\u251c\u2500\u2500 results\/<\/p>\n
\u2502 \u251c\u2500\u2500 collab_network.png<\/p>\n
\u2502 \u251c\u2500\u2500 centrality_radar.png<\/p>\n
\u2502 \u251c\u2500\u2500 community_bar.png<\/p>\n
\u2502 \u251c\u2500\u2500 propagation_hist.png<\/p>\n
\u2502 \u251c\u2500\u2500 role_heatmap.png<\/p>\n
\u2502 \u251c\u2500\u2500 centrality_ranking.csv<\/p>\n
\u2502 \u251c\u2500\u2500 communities.csv<\/p>\n
\u2502 \u251c\u2500\u2500 propagation_depth.csv<\/p>\n
\u2502 \u2514\u2500\u2500 collab_report.txt<\/p>\n
\u2514\u2500\u2500 run_collab_network.py<\/p>\n
\u00a0<\/p>\n
4.2 \u6838\u5fc3\u6e90\u7801<\/p>\n
\u00a0<\/p>\n
<details><\/p>\n
\u00a0<\/p>\n
<summary><\/summary><\/p>\n
\u00a0<\/p>\n
"""\u5e76\u884c\u5de5\u7a0b\u534f\u540c\u65e5\u5fd7\u52a0\u8f7d\u5668"""<\/p>\n
\u00a0<\/p>\n
import pandas as pd<\/p>\n
from pathlib import Path<\/p>\n
from typing import Optional<\/p>\n
\u00a0<\/p>\n
VALID_ROLES = {"\u8bbe\u8ba1", "\u5de5\u827a", "\u52a0\u5de5", "\u8d28\u68c0"}<\/p>\n
\u00a0<\/p>\n
\u00a0<\/p>\n
class CollabLoader:<\/p>\n
\u00a0 \u00a0 """\u52a0\u8f7d\u534f\u540c\u4ea4\u4e92\u65e5\u5fd7"""<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def __init__(self, filepath: str = "collab_log.csv",<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0encoding: str = "utf-8"):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 self.filepath = Path(filepath)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 self.encoding = encoding<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 self._raw: Optional[pd.DataFrame] = None<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def load(self) -> pd.DataFrame:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 if not self.filepath.exists():<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 raise FileNotFoundError(f"\u6587\u4ef6\u4e0d\u5b58\u5728: {self.filepath}")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 self._raw = pd.read_csv(self.filepath, encoding=self.encoding)<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 rename = {}<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 for tgt, al in {<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "source": ["source", "\u53d1\u8d77\u65b9", "from"],<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "target": ["target", "\u63a5\u6536\u65b9", "to"],<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "role_source": ["role_source", "\u53d1\u8d77\u89d2\u8272", "role_from"],<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "role_target": ["role_target", "\u63a5\u6536\u89d2\u8272", "role_to"],<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "interaction_count": ["interaction_count", "\u4ea4\u4e92\u6b21\u6570", "count"],<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "change_type": ["change_type", "\u53d8\u66f4\u7c7b\u578b", "change"],<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "change_depth": ["change_depth", "\u4f20\u64ad\u6df1\u5ea6", "depth"],<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 }.items():<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 if tgt not in self._raw.columns:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 for a in al:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 if a in self._raw.columns:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 rename[a] = tgt<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 break<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 self._raw = self._raw.rename(columns=rename)<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 req = ["source", "target", "role_source", "role_target",<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0"interaction_count"]<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 miss = [c for c in req if c not in self._raw.columns]<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 if miss:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 raise ValueError(f"\u7f3a\u5c11\u5fc5\u8981\u5217: {miss}")<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 self._raw["interaction_count"] = pd.to_numeric(<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 self._raw["interaction_count"], errors="coerce").fillna(1).astype(int)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 self._raw["change_depth"] = pd.to_numeric(<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 self._raw.get("change_depth"), errors="coerce").fillna(0).astype(int)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 self._raw["change_type"] = self._raw.get("change_type", "\u65e0\u53d8\u66f4").fillna("\u65e0\u53d8\u66f4")<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 # \u89d2\u8272\u6821\u9a8c<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 for col in ["role_source", "role_target"]:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 if col not in self._raw.columns:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 self._raw[col] = "\u672a\u77e5"<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 self._raw[col] = self._raw[col].str.strip()<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 return self._raw.reset_index(drop=True)<\/p>\n
\u00a0<\/p>\n
<\/details><\/p>\n
\u00a0<\/p>\n
<details><\/p>\n
\u00a0<\/p>\n
<summary><\/summary><\/p>\n
\u00a0<\/p>\n
"""\u534f\u540c\u7f51\u7edc\u6784\u5efa (networkx)"""<\/p>\n
\u00a0<\/p>\n
import networkx as nx<\/p>\n
import pandas as pd<\/p>\n
from typing import Optional<\/p>\n
\u00a0<\/p>\n
\u00a0<\/p>\n
class CollabNetworkBuilder:<\/p>\n
\u00a0 \u00a0 """\u4ece\u534f\u540c\u65e5\u5fd7\u6784\u5efa\u6709\u5411\/\u65e0\u5411\u56fe"""<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def __init__(self, directed: bool = False):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 self.directed = directed<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 self.G = nx.DiGraph() if directed else nx.Graph()<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def build(self, df: pd.DataFrame) -> nx.Graph:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 self.G.clear()<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 for _, r in df.iterrows():<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 src = str(r["source"])<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 tgt = str(r["target"])<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 # \u8282\u70b9\u5c5e\u6027<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 self.G.add_node(src, role=str(r.get("role_source", "\u672a\u77e5")))<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 self.G.add_node(tgt, role=str(r.get("role_target", "\u672a\u77e5")))<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 # \u8fb9\u5c5e\u6027<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 w = int(r["interaction_count"])<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 if self.G.has_edge(src, tgt):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 self.G[src][tgt]["weight"] += w<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 self.G[src][tgt]["changes"].append(str(r.get("change_type", "")))<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 else:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 self.G.add_edge(src, tgt, weight=w,<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 changes=[str(r.get("change_type", ""))],<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 depth=int(r.get("change_depth", 0)))<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 return self.G<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def basic_stats(self) -> dict:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 return {<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "nodes": self.G.number_of_nodes(),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "edges": self.G.number_of_edges(),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "density": round(nx.density(self.G), 4),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "avg_degree": round(<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 sum(dict(self.G.degree()).values()) \/ max(1, self.G.number_of_nodes()), 2<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 ),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 }<\/p>\n
\u00a0<\/p>\n
<\/details><\/p>\n
\u00a0<\/p>\n
<details><\/p>\n
\u00a0<\/p>\n
<summary><\/summary><\/p>\n
\u00a0<\/p>\n
"""\u4e2d\u5fc3\u6027\u5206\u6790 (networkx\/numpy)"""<\/p>\n
\u00a0<\/p>\n
import numpy as np<\/p>\n
import pandas as pd<\/p>\n
import networkx as nx<\/p>\n
from typing import Optional, Dict, List<\/p>\n
\u00a0<\/p>\n
\u00a0<\/p>\n
class CentralityAnalyzer:<\/p>\n
\u00a0 \u00a0 """\u8ba1\u7b97\u591a\u79cd\u4e2d\u5fc3\u6027\u6307\u6807"""<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def __init__(self):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 pass<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def analyze(self, G: nx.Graph) -> pd.DataFrame:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 deg = nx.degree_centrality(G)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 betw = nx.betweenness_centrality(G, weight="weight", normalized=True)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 close = nx.closeness_centrality(G)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 try:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 eig = nx.eigenvector_centrality_numpy(G, weight="weight")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 except Exception:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 eig = {n: 0.0 for n in G.nodes()}<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 rows = []<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 for n in G.nodes():<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 rows.append({<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "node": n,<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "role": G.nodes[n].get("role", "\u672a\u77e5"),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "degree_centrality": round(deg.get(n, 0), 4),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "betweenness_centrality": round(betw.get(n, 0), 4),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "closeness_centrality": round(close.get(n, 0), 4),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "eigenvector_centrality": round(eig.get(n, 0), 4),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "degree": G.degree(n),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 })<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 return pd.DataFrame(rows).sort_values("degree_centrality",<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 ascending=False).reset_index(drop=True)<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def hub_nodes(self, df: pd.DataFrame, top_n: int = 3) -> pd.DataFrame:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 """\u67a2\u7ebd\u8282\u70b9 = \u4ecb\u6570\u4e2d\u5fc3\u6027\u6700\u9ad8"""<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 return df.nlargest(top_n, "betweenness_centrality").reset_index(drop=True)<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def info_islands(self, df: pd.DataFrame,<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0deg_thresh: float = 0.15,<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0betw_thresh: float = 0.1) -> pd.DataFrame:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 """\u4fe1\u606f\u5b64\u5c9b = \u5ea6\u6570\u548c\u4ecb\u6570\u90fd\u4f4e\u7684\u8282\u70b9"""<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 return df[<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 (df["degree_centrality"] < deg_thresh) &<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 (df["betweenness_centrality"] < betw_thresh)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ].reset_index(drop=True)<\/p>\n
\u00a0<\/p>\n
<\/details><\/p>\n
\u00a0<\/p>\n
<details><\/p>\n
\u00a0<\/p>\n
<summary><\/summary><\/p>\n
\u00a0<\/p>\n
"""\u793e\u533a\u53d1\u73b0 (networkx)"""<\/p>\n
\u00a0<\/p>\n
import networkx as nx<\/p>\n
import pandas as pd<\/p>\n
from typing import Optional, List, Dict<\/p>\n
\u00a0<\/p>\n
\u00a0<\/p>\n
class CommunityDetector:<\/p>\n
\u00a0 \u00a0 """\u57fa\u4e8e\u6a21\u5757\u5ea6\u4f18\u5316\u7684\u793e\u533a\u53d1\u73b0"""<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def __init__(self):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 pass<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def detect(self, G: nx.Graph) -> List[set]:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 """\u8d2a\u5a6a\u6a21\u5757\u5ea6\u793e\u533a\u53d1\u73b0"""<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 if G.number_of_edges() == 0:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 return [set(G.nodes())]<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 communities = nx.community.greedy_modularity_communities(G)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 return [set(c) for c in communities]<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def modularity(self, G: nx.Graph, communities: List[set]) -> float:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 """\u8ba1\u7b97\u6a21\u5757\u5ea6"""<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 if len(communities) <= 1:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 return 0.0<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 return round(nx.community.modularity(G, communities), 4)<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def community_summary(self, G: nx.Graph,<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 communities: List[set]) -> pd.DataFrame:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 """\u793e\u533a\u6c47\u603b"""<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 rows = []<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 for i, comm in enumerate(communities):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 roles = {}<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 for n in comm:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 r = G.nodes[n].get("role", "\u672a\u77e5")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 roles[r] = roles.get(r, 0) + 1<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 rows.append({<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "community_id": i + 1,<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "size": len(comm),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "members": ", ".join(sorted(comm)),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "role_distribution": ", ".join(f"{k}:{v}" for k, v in roles.items()),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "internal_edges": sum(1 for u in comm<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 for v in comm if u != v and G.has_edge(u, v)) \/\/ 2,<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 })<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 return pd.DataFrame(rows)<\/p>\n
\u00a0<\/p>\n
<\/details><\/p>\n
\u00a0<\/p>\n
<details><\/p>\n
\u00a0<\/p>\n
<summary><\/summary><\/p>\n
\u00a0<\/p>\n
"""\u53d8\u66f4\u4f20\u64ad\u5206\u6790"""<\/p>\n
\u00a0<\/p>\n
import numpy as np<\/p>\n
import pandas as pd<\/p>\n
from collections import defaultdict<\/p>\n
from typing import Optional<\/p>\n
\u00a0<\/p>\n
\u00a0<\/p>\n
class PropagationAnalyzer:<\/p>\n
\u00a0 \u00a0 """\u5206\u6790\u53d8\u66f4\u4f20\u64ad\u6df1\u5ea6\u548c\u8def\u5f84"""<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def __init__(self):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 pass<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def depth_distribution(self, df: pd.DataFrame) -> pd.DataFrame:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 """\u4f20\u64ad\u6df1\u5ea6\u5206\u5e03"""<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 depths = df[df["change_depth"] > 0]["change_depth"]<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 if depths.empty:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 return pd.DataFrame()<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 dist = depths.value_counts().sort_index().reset_index()<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 dist.columns = ["depth", "count"]<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 dist["ratio"] = (dist["count"] \/ dist["count"].sum()).round(4)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 return dist<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def deepest_chains(self, df: pd.DataFrame,<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0top_n: int = 5) -> pd.DataFrame:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 """\u4f20\u64ad\u6df1\u5ea6\u6700\u5927\u7684\u4ea4\u4e92"""<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 sub = df[df["change_depth"] > 0].copy()<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 if sub.empty:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 return pd.DataFrame()<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 sub = sub.sort_values("change_depth", ascending=False).head(top_n)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 return sub[["source", "target", "change_type",<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0"change_depth", "interaction_count"]].reset_index(drop=True)<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def change_type_stats(self, df: pd.DataFrame) -> pd.DataFrame:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 """\u6309\u53d8\u66f4\u7c7b\u578b\u7edf\u8ba1"""<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 sub = df[df["change_type"] != "\u65e0\u53d8\u66f4"].copy()<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 if sub.empty:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 return pd.DataFrame()<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 rows = []<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 for ct, g in sub.groupby("change_type"):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 rows.append({<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "change_type": ct,<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "count": len(g),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "avg_depth": round(g["change_depth"].mean(), 2),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "max_depth": int(g["change_depth"].max()),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "total_interactions": int(g["interaction_count"].sum()),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 })<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 return pd.DataFrame(rows).sort_values("avg_depth",<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 ascending=False).reset_index(drop=True)<\/p>\n
\u00a0<\/p>\n
<\/details><\/p>\n
\u00a0<\/p>\n
<details><\/p>\n
\u00a0<\/p>\n
<summary><\/summary><\/p>\n
\u00a0<\/p>\n
"""\u53ef\u89c6\u5316 (matplotlib)"""<\/p>\n
\u00a0<\/p>\n
import numpy as np<\/p>\n
import pandas as pd<\/p>\n
import matplotlib.pyplot as plt<\/p>\n
from pathlib import Path<\/p>\n
\u00a0<\/p>\n
plt.rcParams["font.sans-serif"] = ["SimHei", "DejaVu Sans"]<\/p>\n
plt.rcParams["axes.unicode_minus"] = False<\/p>\n
\u00a0<\/p>\n
\u00a0<\/p>\n
class CollabVisualizer:<\/p>\n
\u00a0 \u00a0 def __init__(self, results_dir: str = "results"):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 self.results_dir = Path(results_dir)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 self.results_dir.mkdir(exist_ok=True)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 self.role_colors = {<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "\u8bbe\u8ba1": "#E74C3C", "\u5de5\u827a": "#3498DB",<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "\u52a0\u5de5": "#27AE60", "\u8d28\u68c0": "#F39C12", "\u672a\u77e5": "#95A5A6"<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 }<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def network_plot(self, G, centrality_df):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 fig, ax = plt.subplots(figsize=(16, 12))<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 pos = nx.spring_layout(G, seed=42, k=0.8)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 # \u8282\u70b9\u989c\u8272=\u89d2\u8272<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 nc = [self.role_colors.get(G.nodes[n].get("role", "\u672a\u77e5"), "#999")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 for n in G.nodes()]<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 # \u8282\u70b9\u5927\u5c0f=\u5ea6\u6570\u4e2d\u5fc3\u6027<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 deg_c = dict(zip(centrality_df["node"], centrality_df["degree_centrality"]))<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ns = [max(300, deg_c.get(n, 0) * 5000) for n in G.nodes()]<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 # \u8fb9\u5bbd=\u6743\u91cd<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ew = [max(0.5, G[u][v].get("weight", 1) \/ 3) for u, v in G.edges()]<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 nx.draw_networkx_nodes(G, pos, node_color=nc, node_size=ns,<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0ax=ax, alpha=0.9, edgecolors="black", linewidths=1.0)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 nx.draw_networkx_edges(G, pos, width=ew, alpha=0.35, ax=ax,<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 edge_color="#7F8C8D")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 nx.draw_networkx_labels(G, pos, font_size=8, ax=ax,<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0font_weight="bold")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 # \u56fe\u4f8b<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 for role, color in self.role_colors.items():<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 ax.plot([], [], "o", color=color, label=role, markersize=10)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.legend(loc="upper left", framealpha=0.9)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.set_title("\u5e76\u884c\u5de5\u7a0b\u534f\u540c\u7f51\u7edc\\\\n(\u989c\u8272=\u89d2\u8272 \u5927\u5c0f=\u5ea6\u6570\u4e2d\u5fc3\u6027 \u7ebf\u5bbd=\u4ea4\u4e92\u9891\u6b21)",<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0fontsize=14, fontweight="bold")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.axis("off")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 plt.tight_layout()<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 plt.savefig(self.results_dir \/ "collab_network.png",<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0dpi=150, bbox_inches="tight")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 plt.close()<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def centrality_radar(self, centrality_df, top_n=5):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 sub = centrality_df.head(top_n)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 metrics = ["degree_centrality", "betweenness_centrality",<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "closeness_centrality", "eigenvector_centrality"]<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 labels = ["\u5ea6\u6570", "\u4ecb\u6570", "\u63a5\u8fd1", "\u7279\u5f81\u5411\u91cf"]<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 angles = np.linspace(0, 2 * np.pi, len(metrics), endpoint=False).tolist()<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 angles += angles[:1]<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 fig, ax = plt.subplots(figsize=(10, 10), subplot_kw=dict(polar=True))<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 for _, r in sub.iterrows():<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 vals = [r[m] for m in metrics]<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 max_val = centrality_df[m].max() if m in centrality_df.columns else 1<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 vals = [v \/ max(centrality_df[m].max(), 1e-6) for m in metrics]<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 vals += vals[:1]<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 ax.plot(angles, vals, "o-", label=r["node"], linewidth=2)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 ax.fill(angles, vals, alpha=0.1)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.set_xticks(angles[:-1])<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.set_xticklabels(labels, fontsize=12)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.set_title(f"Top{top_n} \u67a2\u7ebd\u8282\u70b9\u4e2d\u5fc3\u6027\u96f7\u8fbe\u56fe", fontsize=14, fontweight="bold")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.legend(loc="upper right", bbox_to_anchor=(1.3, 1.1))<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 plt.tight_layout()<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 plt.savefig(self.results_dir \/ "centrality_radar.png",<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0dpi=150, bbox_inches="tight")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 plt.close()<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def community_bar(self, comm_df):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 if comm_df.empty:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 return<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 fig, ax = plt.subplots(figsize=(10, 6))<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 x = range(len(comm_df))<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.bar(x, comm_df["size"], color="#3498DB", alpha=0.8)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.set_xticks(list(x))<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.set_xticklabels([f"C{i+1}" for i in range(len(comm_df))], fontsize=11)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.set_ylabel("\u793e\u533a\u4eba\u6570")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.set_title("\u793e\u533a\u89c4\u6a21\u5206\u5e03", fontsize=14, fontweight="bold")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 for i, r in comm_df.iterrows():<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 ax.text(i, r["size"] + 0.1, r["role_distribution"],<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 ha="center", fontsize=8)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.grid(axis="y", alpha=0.3)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 plt.tight_layout()<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 plt.savefig(self.results_dir \/ "community_bar.png",<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0dpi=150, bbox_inches="tight")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 plt.close()<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def propagation_hist(self, depth_df):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 if depth_df.empty:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 return<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 fig, ax = plt.subplots(figsize=(9, 6))<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.bar(depth_df["depth"].astype(str), depth_df["count"],<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0color="#E67E22", alpha=0.8)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.set_xlabel("\u4f20\u64ad\u6df1\u5ea6(\u8df3\u6570)")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.set_ylabel("\u53d8\u66f4\u6b21\u6570")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.set_title("\u53d8\u66f4\u4f20\u64ad\u6df1\u5ea6\u5206\u5e03", fontsize=14, fontweight="bold")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.grid(axis="y", alpha=0.3)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 total = depth_df["count"].sum()<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 for _, r in depth_df.iterrows():<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 pct = r["count"] \/ total * 100<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 ax.text(str(r["depth"]), r["count"] + 0.2, f"{pct:.1f}%",<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 ha="center", fontsize=10, fontweight="bold")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 plt.tight_layout()<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 plt.savefig(self.results_dir \/ "propagation_hist.png",<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0dpi=150, bbox_inches="tight")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 plt.close()<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def role_heatmap(self, df):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 """\u89d2\u8272\u95f4\u4ea4\u4e92\u70ed\u529b\u56fe"""<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 df2 = df.copy()<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 crosstab = pd.crosstab(<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 df2["role_source"], df2["role_target"],<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 values=df2["interaction_count"], aggfunc="sum"<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ).fillna(0)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 fig, ax = plt.subplots(figsize=(8, 6))<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 im = ax.imshow(crosstab.values, cmap="YlOrRd")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.set_xticks(range(len(crosstab.columns)))<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.set_xticklabels(crosstab.columns, fontsize=11)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.set_yticks(range(len(crosstab.index)))<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.set_yticklabels(crosstab.index, fontsize=11)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 ax.set_title("\u89d2\u8272\u95f4\u4ea4\u4e92\u70ed\u529b\u56fe(\u9891\u6b21)", fontsize=14, fontweight="bold")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 for i in range(len(crosstab.index)):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 for j in range(len(crosstab.columns)):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 ax.text(j, i, int(crosstab.values[i, j]),<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 ha="center", va="center", fontsize=10,<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 color="black" if crosstab.values[i, j] < crosstab.values.max()\/2 else "white")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 plt.colorbar(im, ax=ax)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 plt.tight_layout()<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 plt.savefig(self.results_dir \/ "role_heatmap.png",<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0dpi=150, bbox_inches="tight")<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 plt.close()<\/p>\n
\u00a0<\/p>\n
<\/details><\/p>\n
\u00a0<\/p>\n
<details><\/p>\n
\u00a0<\/p>\n
<summary><\/summary><\/p>\n
\u00a0<\/p>\n
"""\u5408\u6210\u5e76\u884c\u5de5\u7a0b\u534f\u540c\u6570\u636e\u751f\u6210\u5668"""<\/p>\n
\u00a0<\/p>\n
import numpy as np<\/p>\n
import pandas as pd<\/p>\n
from pathlib import Path<\/p>\n
from typing import Optional<\/p>\n
\u00a0<\/p>\n
\u00a0<\/p>\n
class SyntheticCollabGenerator:<\/p>\n
\u00a0 \u00a0 """<\/p>\n
\u00a0 \u00a0 \u751f\u6210 10 \u5de5\u7a0b\u5e08\u00d74\u89d2\u8272 \u534f\u540c\u65e5\u5fd7<\/p>\n
\u00a0 \u00a0 \u89d2\u8272: \u8bbe\u8ba1(3\u4eba) \u5de5\u827a(2\u4eba) \u52a0\u5de5(3\u4eba) \u8d28\u68c0(2\u4eba)<\/p>\n
\u00a0 \u00a0 \u4ea4\u4e92\u9891\u6b21: \u5e42\u5f8b\u5206\u5e03(\u67a2\u7ebd\u591a, \u8fb9\u7f18\u5c11)<\/p>\n
\u00a0 \u00a0 \u53d8\u66f4\u4f20\u64ad: \u6df1\u5ea61~4\u8df3<\/p>\n
\u00a0 \u00a0 """<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def __init__(self, rng: Optional[np.random.RandomState] = None):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 self.rng = rng or np.random.RandomState(42)<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 def generate(self, output_path: str = "collab_log.csv",<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0n_engineers: int = 10) -> pd.DataFrame:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 roles = {<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "\u8bbe\u8ba1": ["\u5f20\u5de5", "\u738b\u5de5", "\u674e\u5de5"],<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "\u5de5\u827a": ["\u8d75\u5de5", "\u5b59\u5de5"],<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "\u52a0\u5de5": ["\u5468\u5de5", "\u5434\u5de5", "\u90d1\u5de5"],<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 "\u8d28\u68c0": ["\u94b1\u5de5", "\u51af\u5de5"],<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 }<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 engineers = []<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 for role, names in roles.items():<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 for name in names:<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 engineers.append((f"{role}-{name}", role))<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 change_types = ["\u5c3a\u5bf8\u53d8\u66f4", "\u516c\u5dee\u8c03\u6574", "\u6750\u6599\u53d8\u66f4",<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0"\u8868\u9762\u5904\u7406\u53d8\u66f4", "\u88c5\u914d\u987a\u5e8f\u53d8\u66f4", "\u65e0\u53d8\u66f4"]<\/p>\n
\u00a0<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 records = []<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 # \u751f\u6210\u4ea4\u4e92(\u5e42\u5f8b: \u90e8\u5206\u4eba\u4ea4\u4e92\u591a)<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 for i, (s, sr) in enumerate(engineers):<\/p>\n
\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 f<\/p>\n
\u5229\u7528AI\u89e3\u51b3\u5b9e\u9645\u95ee\u9898，\u5982\u679c\u4f60\u89c9\u5f97\u8fd9\u4e2a\u5de5\u5177\u597d\u7528，\u6b22\u8fce\u5173\u6ce8\u957f\u5b89\u7267\u7b1b！<\/p>\n","protected":false},"excerpt":{"rendered":"
\u5468\u4e94\u4e0b\u5348，\u6280\u672f\u90e8\u4f1a\u8bae\u5ba4\u3002\\”\u8fd9\u6279\u65b0\u5939\u5177\u4ece\u8bbe\u8ba1\u5230\u9996\u4ef6\u5408\u683c，\u8d70\u4e86 23 \u5929，\\”\u9879\u76ee\u7ecf\u7406\u8001\u5218\u628a\u7518\u7279\u56fe\u6295\u5f71\u5728\u5899\u4e0a，\\”\u8bbe\u8ba1 5 \u5929，\u5de5\u827a 4 \u5929，\u52a0\u5de5 8 \u5929，\u8d28\u68c0 6 \u5929\u2014\u2014\u770b\u8d77\u6765\u5404\u5e72\u5404\u7684，\u4e32\u884c\u6392\u961f\u3002\u53ef\u5b9e\u9645\u4e0a\u8bbe\u8ba1\u6539\u4e86\u7b2c\u4e09\u7248\u7684\u65f6\u5019\u5de5\u827a\u5c31\u5df2\u7ecf\u5728\u5ba1\u7b2c\u4e8c\u7248\u4e86，\u52a0\u5de5\u90a3\u8fb9\u63d0\u524d\u770b\u4e86\u4e09\u7ef4\u6a21\u578b\u5728\u51c6\u5907\u6bdb\u576f，\u8d28\u68c0\u4e5f\u63d0\u524d\u4ecb\u5165\u4e86\u516c\u5dee\u8bc4\u5ba1\u3002\u95ee\u9898\u662f\u8fd9\u4e9b\u5e76\u884c\u5168\u9760\u5fae\u4fe1\u7fa4\u91cc\u543c<\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[1818,4278,81,737],"topic":[],"class_list":["post-112154","post","type-post","status-publish","format-standard","hentry","category-server","tag-numpy","tag-pandas","tag-python","tag-scikit-learn"],"yoast_head":"\n