restructure: proper project layout, README, kill training

- cuda/ — main LBM kernel (khra_gixx_1024_v5.cu)
- navigator/ — lattice_observer, golden_weave, bridges, mock daemon
- scripts/ — compile, start, launch, setup (paths updated)
- docs/ — system manual
- archive/ — everything else (old kernels, inquiries, experiments)
- README.md — full setup guide: requirements, quick start, use your own LLM
- removed training/ entirely (broken LoRA scripts + datasets)
- .gitignore: exclude build/ logs/ training/ *.jsonl

archive/misc/quick_analyze.py

@@ -0,0 +1,69 @@
+#!/usr/bin/env python3
+"""Quick sweep analysis - Windows native paths"""
+
+import json
+import glob
+import re
+from collections import defaultdict
+
+# Read all status files
+files = glob.glob('sweep_results/run_*_status.json')
+print(f"Found {len(files)} status files")
+
+# Collect data
+data_points = []
+for f in files:
+    try:
+        with open(f) as fp:
+            data = json.load(fp)
+            match = re.search(r'run_(\d+)', f)
+            if match:
+                run_id = int(match.group(1))
+                coherence = data.get('coherence')
+                asymmetry = data.get('asymmetry')
+                if coherence and asymmetry and 10 < asymmetry < 20:
+                    data_points.append({
+                        'run_id': run_id,
+                        'coherence': coherence,
+                        'asymmetry': asymmetry,
+                        'cycle': data.get('cycle', 0)
+                    })
+    except:
+        pass
+
+print(f"Valid data points: {len(data_points)}")
+
+if data_points:
+    # Basic stats
+    coherences = [d['coherence'] for d in data_points]
+    asymmetries = [d['asymmetry'] for d in data_points]
+    
+    print(f"\nCoherence: {min(coherences):.4f} to {max(coherences):.4f}, mean={sum(coherences)/len(coherences):.4f}")
+    print(f"Asymmetry: {min(asymmetries):.4f} to {max(asymmetries):.4f}, mean={sum(asymmetries)/len(asymmetries):.4f}")
+    
+    # Correlation
+    import math
+    n = len(data_points)
+    mean_c = sum(coherences)/n
+    mean_a = sum(asymmetries)/n
+    
+    cov = sum((c - mean_c) * (a - mean_a) for c, a in zip(coherences, asymmetries))
+    var_c = sum((c - mean_c)**2 for c in coherences)
+    var_a = sum((a - mean_a)**2 for a in asymmetries)
+    
+    if var_c > 0 and var_a > 0:
+        correlation = cov / math.sqrt(var_c * var_a)
+        print(f"\nCorrelation (coherence vs asymmetry): {correlation:.4f}")
+    
+    # Run progression
+    print("\n=== RUN PROGRESSION ===")
+    by_run = defaultdict(list)
+    for d in data_points:
+        by_run[d['run_id']].append(d['asymmetry'])
+    
+    # Show first 10 runs
+    for run_id in sorted(by_run.keys())[:10]:
+        asymms = by_run[run_id]
+        print(f"Run {run_id}: {len(asymms)} samples, asymmetry {min(asymms):.2f}-{max(asymms):.2f}")
+
+print("\nAnalysis complete.")