arco 0.4.1

Automated Research into Computational Ontologies — a platform for discovering the conditions under which computation emerges
Documentation 
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
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285

#!/usr/bin/env python3
"""scripts/analyze.py — Analyze ARCO sweep data and generate plots.

Usage:
    ./scripts/sweep.sh                    # Run experiments and save JSON
    python3 scripts/analyze.py sweep_data # Analyze and plot

Requires: matplotlib, numpy
"""

import json
import os
import sys
from collections import defaultdict
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.ticker as mticker

DATA_DIR = sys.argv[1] if len(sys.argv) > 1 else "sweep_data"

# ===================================================================
# Load data
# ===================================================================

def load_records(substrate):
    """Load all JSON records for a substrate."""
    records = {}
    for filename in os.listdir(DATA_DIR):
        if not filename.endswith('.json'):
            continue
        parts = filename.replace('.json', '').split('_')
        if parts[0] != substrate:
            continue
        seed = parts[1]
        with open(os.path.join(DATA_DIR, filename)) as f:
            records[seed] = json.load(f)
    return records

graph = load_records("graph")
ca = load_records("ca")

# ===================================================================
# Extract data
# ===================================================================

def extract_hypotheses(records):
    """Extract hypothesis accuracies across seeds."""
    data = defaultdict(list)
    survival = defaultdict(int)
    for seed, record in records.items():
        for h in record['hypotheses']:
            data[h['name']].append(h['accuracy'] * 100)
            if h['survives']:
                survival[h['name']] += 1
    return data, survival

def extract_spectrum(records):
    """Extract storage rates by structured ratio bracket."""
    brackets = {
        'Noise': (0.00, 0.15),
        'Noise-dominated': (0.15, 0.40),
        'Balanced': (0.40, 0.60),
        'Structure-dominated': (0.60, 0.85),
        'Structured': (0.85, 1.01),
    }
    data = defaultdict(list)
    for seed, record in records.items():
        threshold = record['thresholds'].get('storage', 0.0)
        for bracket, (low, high) in brackets.items():
            group = [r for r in record['results']
                     if low <= r['structured_ratio'] < high]
            if group:
                rate = 100.0 * sum(1 for r in group if r['storage'] > threshold) / len(group)
                data[bracket].append(rate)
    return data

def extract_storage_distribution(records):
    """Extract raw storage values by bracket for distribution plots."""
    brackets = {
        'Noise': (0.00, 0.15),
        'Noise-dominated': (0.15, 0.40),
        'Balanced': (0.40, 0.60),
        'Structure-dominated': (0.60, 0.85),
        'Structured': (0.85, 1.01),
    }
    data = defaultdict(list)
    for seed, record in records.items():
        for bracket, (low, high) in brackets.items():
            for r in record['results']:
                if low <= r['structured_ratio'] < high:
                    data[bracket].append(r['storage'])
    return data

# ===================================================================
# Plot 1: Structure-Storage Gradient
# ===================================================================

def plot_spectrum(graph_spectrum, ca_spectrum, title, filename):
    fig, ax = plt.subplots(figsize=(10, 5))
    
    brackets = ['Noise', 'Noise-dominated', 'Balanced', 'Structure-dominated', 'Structured']
    x = np.arange(len(brackets))
    width = 0.35
    
    # Graph data
    means = [np.mean(graph_spectrum.get(b, [0])) for b in brackets]
    mins = [np.min(graph_spectrum.get(b, [0])) for b in brackets]
    maxs = [np.max(graph_spectrum.get(b, [0])) for b in brackets]
    errors_low = [m - lo for m, lo in zip(means, mins)]
    errors_high = [hi - m for m, hi in zip(means, maxs)]
    
    bars1 = ax.bar(x - width/2, means, width, 
                   yerr=[errors_low, errors_high],
                   label='Binary Graph', color='#58a6ff', capsize=3)
    
    # CA doesn't have the same brackets — skip for now or use simplified brackets
    # For CA, we have "Low structure" and "High structure"
    
    ax.set_ylabel('Storage Rate (%)')
    ax.set_title(title, fontweight='bold')
    ax.set_xticks(x)
    ax.set_xticklabels(brackets, rotation=15)
    ax.legend()
    ax.set_ylim(0, 110)
    ax.yaxis.set_major_formatter(mticker.FormatStrFormatter('%.0f%%'))
    
    plt.tight_layout()
    plt.savefig(filename, dpi=150, bbox_inches='tight')
    plt.close()
    print(f"Saved {filename}")

# ===================================================================
# Plot 2: Hypothesis Survival
# ===================================================================

def plot_hypotheses(graph_data, graph_survival, ca_data, ca_survival, filename):
    fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 6))
    
    for ax, (data, survival), title in [
        (ax1, (graph_data, graph_survival), 'Binary Graph'),
        (ax2, (ca_data, ca_survival), 'Cellular Automata')
    ]:
        names = sorted(data.keys())
        means = [np.mean(data[n]) for n in names]
        mins = [np.min(data[n]) for n in names]
        maxs = [np.max(data[n]) for n in names]
        errors_low = [m - lo for m, lo in zip(means, mins)]
        errors_high = [hi - m for m, hi in zip(means, maxs)]
        
        colors = []
        for n in names:
            s = survival.get(n, 0)
            if s >= 8:
                colors.append('#3fb950')  # green
            elif s >= 5:
                colors.append('#d29922')  # yellow
            else:
                colors.append('#f85149')  # red
        
        x = np.arange(len(names))
        bars = ax.bar(x, means, color=colors, capsize=3)
        ax.errorbar(x, means, yerr=[errors_low, errors_high], fmt='none', 
                    ecolor='#8b949e', capsize=3)
        
        # Survival count labels
        for i, n in enumerate(names):
            ax.text(i, means[i] + errors_high[i] + 1, 
                   f'{survival.get(n, 0)}/10',
                   ha='center', fontsize=7, color='#8b949e')
        
        ax.set_title(title, fontweight='bold')
        ax.set_ylabel('Accuracy (%)')
        ax.set_xticks(x)
        ax.set_xticklabels([n.replace('_', '\n') for n in names], 
                          rotation=0, fontsize=8)
        ax.set_ylim(0, 110)
        ax.axhline(y=50, color='#8b949e', linestyle='--', alpha=0.5)
        ax.yaxis.set_major_formatter(mticker.FormatStrFormatter('%.0f%%'))
    
    plt.suptitle('Hypothesis Survival Across 10 Seeds', fontweight='bold')
    plt.tight_layout()
    plt.savefig(filename, dpi=150, bbox_inches='tight')
    plt.close()
    print(f"Saved {filename}")

# ===================================================================
# Plot 3: Storage Distribution
# ===================================================================

def plot_distribution(graph_dist, filename):
    fig, ax = plt.subplots(figsize=(10, 5))
    
    brackets = ['Noise', 'Noise-dominated', 'Balanced', 'Structure-dominated', 'Structured']
    data = [graph_dist.get(b, []) for b in brackets]
    
    parts = ax.violinplot(data, positions=np.arange(len(brackets)), showmeans=True, showmedians=True)
    
    for pc in parts['bodies']:
        pc.set_facecolor('#58a6ff')
        pc.set_alpha(0.7)
    
    ax.set_ylabel('Storage Value')
    ax.set_title('Storage Distribution by Structured Ratio (Binary Graph)', fontweight='bold')
    ax.set_xticks(np.arange(len(brackets)))
    ax.set_xticklabels(brackets, rotation=15)
    ax.set_ylim(0, 1.1)
    
    plt.tight_layout()
    plt.savefig(filename, dpi=150, bbox_inches='tight')
    plt.close()
    print(f"Saved {filename}")


# ===================================================================
# Print spectrum
# ===================================================================

def print_spectrum_table(spectrum, substrate_name):
    """Print the full Structure-Storage Gradient table."""
    brackets = ['Noise', 'Noise-dominated', 'Balanced', 'Structure-dominated', 'Structured']
    
    print(f"\n{substrate_name} — Structure-Storage Gradient:")
    print(f"{'Bracket':<22} {'Storage Rate Range':<22} {'Mean':<8}")
    print("-" * 52)
    for b in brackets:
        if b in spectrum and spectrum[b]:
            vals = spectrum[b]
            print(f"{b:<22} {min(vals):.1f}--{max(vals):.1f}%{'':<8} {np.mean(vals):.1f}%")


def print_hypothesis(graph_hyp, graph_surv, ca_hyp, ca_surv):
    """Print hypothesis summary table"""

    print("\nHypothesis Summary:")
    print(f"{'Hypothesis':<35} {'Survival':<10} {'Acc Range':<18} {'Mean':<8}")
    print("-" * 71)

    all_hyp = {}
    for name, accs in graph_hyp.items():
        all_hyp[f"graph/{name}"] = (accs, graph_surv.get(name, 0))
    for name, accs in ca_hyp.items():
        all_hyp[f"ca/{name}"] = (accs, ca_surv.get(name, 0))

    for name in sorted(all_hyp.keys()):
        accs, surv = all_hyp[name]
        print(f"{name:<35} {surv}/10{'':<4} {min(accs):.1f}--{max(accs):.1f}%{'':<8} {np.mean(accs):.1f}%")


# ===================================================================
# Main
# ===================================================================

# Extract
graph_hyp, graph_surv = extract_hypotheses(graph)
ca_hyp, ca_surv = extract_hypotheses(ca)
graph_spectrum = extract_spectrum(graph)
graph_dist = extract_storage_distribution(graph)

# Plot
plot_spectrum(graph_spectrum, {}, 'Structure-Storage Gradient (Binary Graph)', f'{DATA_DIR}/plot_spectrum.png')
plot_hypotheses(graph_hyp, graph_surv, ca_hyp, ca_surv, f'{DATA_DIR}/plot_hypotheses.png')
plot_distribution(graph_dist, f'{DATA_DIR}/plot_distribution.png')

# Print spectrum
print_spectrum_table(graph_spectrum, "Binary Graph")

# Print hypothesis
print_hypothesis(graph_hyp, graph_surv, ca_hyp, ca_surv)

exit()

# Print summary table
print("\nHypothesis Summary:")
print(f"{'Hypothesis':<35} {'Survival':<10} {'Acc Range':<18} {'Mean':<8}")
print("-" * 71)

all_hyp = {}
for name, accs in graph_hyp.items():
    all_hyp[f"graph/{name}"] = (accs, graph_surv.get(name, 0))
for name, accs in ca_hyp.items():
    all_hyp[f"ca/{name}"] = (accs, ca_surv.get(name, 0))

for name in sorted(all_hyp.keys()):
    accs, surv = all_hyp[name]
    print(f"{name:<35} {surv}/10{'':<4} {min(accs):.1f}--{max(accs):.1f}%{'':<8} {np.mean(accs):.1f}%")