Spaces:
Paused
Paused
File size: 8,418 Bytes
18a3a92 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 | # utils.py
import networkx as nx
import config
def calculate_metric(G, metric_name):
"""
Calculates metrics. Includes:
Originals: Density, Clustering, Cycles, Interdependence, etc
"""
try:
n = G.number_of_nodes()
if metric_name == "Nodes": return str(n)
if metric_name == "Edges": return str(G.number_of_edges())
if n == 0: return "0"
if metric_name == "Density":
return f"{nx.density(G):.3f}"
if metric_name == "Avg Degree":
avg_deg = sum([d for _, d in G.degree()]) / n
return f"{avg_deg:.2f}"
if metric_name == "Avg Cycle Length":
try:
cycles = list(nx.simple_cycles(G))
if not cycles: return "0.0 (None)"
lengths = [len(c) for c in cycles]
avg = sum(lengths) / len(lengths)
return f"{avg:.2f} {lengths}"
except: return "Err"
if metric_name == "Interdependence":
try:
m = G.number_of_edges()
if m == 0: return "0.000"
cross_boundary_edges = 0
for u, v in G.edges():
agent_u = G.nodes[u].get('agent', 'Unassigned')
agent_v = G.nodes[v].get('agent', 'Unassigned')
if agent_u != agent_v: cross_boundary_edges += 1
return f"{(cross_boundary_edges / m):.3f}"
except: return "Err"
if metric_name == "Cyclomatic Number":
try:
e = G.number_of_edges()
p = nx.number_weakly_connected_components(G)
return str(e - n + p)
except: return "Err"
if metric_name == "Critical Loop Nodes":
try:
fvs = nx.approximation.min_weighted_feedback_vertex_set(G)
return str(len(fvs))
except: return "0"
if metric_name == "Total Cycles":
try:
count = 0
for _ in nx.simple_cycles(G):
count += 1
if count > 100: return "100+"
return str(count)
except: return "Err"
if metric_name == "Global Efficiency":
# Measures how integrated the system is (0.0 to 1.0)
try:
# Treated as undirected to measure potential for information flow
eff = nx.global_efficiency(G.to_undirected())
return f"{eff:.3f}"
except: return "Err"
if metric_name == "Modularity":
# Detects if system splits into distinct groups (Q-Score)
try:
communities = nx.community.greedy_modularity_communities(G.to_undirected())
num_communities = len(communities)
q_score = nx.community.modularity(G.to_undirected(), communities)
return f"Q={q_score:.2f} ({num_communities} Grps)"
except: return "Err"
if metric_name == "Brittleness Ratio": # This isn't a great name, will rename the button
# A. Ratio of Soft to Hard edges
soft = sum(1 for u, v, d in G.edges(data=True) if d.get('type') == config.EDGE_TYPE_SOFT)
hard = sum(1 for u, v, d in G.edges(data=True) if d.get('type') == config.EDGE_TYPE_HARD)
if hard == 0: return "Infinite (No Hard Edges)"
ratio = soft / hard
return f"{ratio:.2f} (S:{soft}/H:{hard})"
if metric_name == "Supportive Gain":
# B. Supportive Gain: Efficiency(Total) - Efficiency(HardOnly)
# Measures how much integration is lost if soft edges fail.
# 1. Total Efficiency
eff_total = nx.global_efficiency(G.to_undirected())
# 2. Hard-edges Only Efficiency
hard_edges = [(u, v) for u, v, d in G.edges(data=True) if d.get('type') == config.EDGE_TYPE_HARD]
G_hard = nx.Graph()
G_hard.add_nodes_from(G.nodes())
G_hard.add_edges_from(hard_edges)
eff_hard = nx.global_efficiency(G_hard)
gain = eff_total - eff_hard
return f"{gain:.3f} (Tot: {eff_total:.2f})"
if metric_name == "Critical Vulnerability":
# C. Critical Path Vulnerability:
# Does the graph fragment if soft edges are removed?
# Measures Connected Components of the Hard-Only graph.
hard_edges = [(u, v) for u, v, d in G.edges(data=True) if d.get('type') == config.EDGE_TYPE_HARD]
G_hard = nx.DiGraph() # di-graph to catch strict flow breaks
G_hard.add_nodes_from(G.nodes())
G_hard.add_edges_from(hard_edges)
# Weakly connected components (islands of connectivity)
num_components = nx.number_weakly_connected_components(G_hard)
if num_components == 1:
return "Robust (1 Comp)"
else:
return f"Fractured ({num_components} Comps)"
### Metric Ideas for analyzing SHared Authority
if metric_name == "Functional Redundancy":
# "Average number of agents" capable of performing a function.
# Idea: 1.0 = Brittle (No backup), >1.2 = Resilient (more backups)
total_agents = 0
func_count = 0
for n, d in G.nodes(data=True):
if d.get('type') == 'Function':
# Get list of agents, handling single strings if necessary
ag = d.get('agent', [])
if not isinstance(ag, list): ag = [ag]
# Count real agents (exclude "Unassigned")
real_agents = [x for x in ag if x != "Unassigned"]
total_agents += len(real_agents)
func_count += 1
if func_count == 0: return "0.0"
avg = total_agents / func_count
return f"{avg:.2f} (Avg Agents)"
if metric_name == "Agent Criticality":
# try to find the agent with the most "authority" in the system
# for the most functions? (If they disconnect, these functions fail)
agent_sole_counts = {}
for n, d in G.nodes(data=True):
if d.get('type') == 'Function':
ag = d.get('agent', [])
if not isinstance(ag, list): ag = [ag]
real_agents = [x for x in ag if x != "Unassigned"]
# If exactly one agent owns this, they are critical to it
if len(real_agents) == 1:
sole_agent = real_agents[0]
agent_sole_counts[sole_agent] = agent_sole_counts.get(sole_agent, 0) + 1
if not agent_sole_counts:
return "None (Robust)"
# Find the agent with the highest count
worst_agent = max(agent_sole_counts, key=agent_sole_counts.get)
count = agent_sole_counts[worst_agent]
return f"{worst_agent} ({count} Sole Tasks)"
if metric_name == "Collaboration Ratio":
# Percent of functions that involve multiple agents
shared_funcs = 0
total_funcs = 0
for n, d in G.nodes(data=True):
if d.get('type') == 'Function':
ag = d.get('agent', [])
if not isinstance(ag, list): ag = [ag]
# Filter out 'Unassigned' (only count other agents)
real_agents = [x for x in ag if x != "Unassigned"]
if len(real_agents) > 0:
total_funcs += 1
# if more than 1 agent is assigned then it's a collaborative task
if len(real_agents) > 1:
shared_funcs += 1
if total_funcs == 0: return "0.0%"
ratio = (shared_funcs / total_funcs) * 100
return f"{ratio:.1f}%"
except Exception as e:
print(f"Error calculating {metric_name}: {e}")
return "Err"
return "" |