Spaces:
Runtime error
Runtime error
File size: 20,923 Bytes
e92be04 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 | import { db } from '../config/gun.js';
import { gunSafe } from '../utils/gunUtils.js';
import { sandbox } from './IsolateSandbox.js';
/**
* Gene Definitions — structured genome for P2PCLAW protocol optimization
* Each gene is a continuous [0,1] parameter governing network behavior.
*/
export const GENE_DEFS = [
{ key: 'research_depth', min: 0, max: 1, optimum: 0.65, label: 'Research Depth', desc: 'Depth vs. breadth of topic investigation per agent' },
{ key: 'validation_strictness', min: 0, max: 1, optimum: 0.70, label: 'Validation Strictness', desc: 'Rigor of peer review applied to submitted papers' },
{ key: 'publication_rate', min: 0, max: 1, optimum: 0.40, label: 'Publication Rate', desc: 'Frequency of publishing findings (higher = more spam risk)' },
{ key: 'consensus_threshold', min: 0, max: 1, optimum: 0.68, label: 'Consensus Threshold', desc: 'Minimum agreement ratio required to promote a paper' },
{ key: 'collaboration_weight', min: 0, max: 1, optimum: 0.55, label: 'Collaboration Weight', desc: 'Tendency to collaborate vs. isolated solo research' },
{ key: 'exploration_rate', min: 0, max: 1, optimum: 0.38, label: 'Exploration Rate', desc: 'Explore new topics vs. exploit established research areas' },
{ key: 'fault_tolerance', min: 0, max: 1, optimum: 0.80, label: 'Fault Tolerance', desc: 'Network resilience to agent failures and adversarial nodes' },
{ key: 'convergence_speed', min: 0, max: 1, optimum: 0.45, label: 'Convergence Speed', desc: 'Speed of consensus convergence (too fast = premature, too slow = stagnation)' },
];
/**
* GeneticService — Full Evolutionary Engine
*
* Implements:
* - Real genetic algorithm (selection, crossover, mutation, elitism)
* - Multi-objective fitness function based on network optimization theory
* - Lineage tracking (parent IDs per offspring)
* - Population diversity metric (avg pairwise gene distance)
* - Persistence via Gun.js
* - Code mutation sandbox (legacy IsolateSandbox integration)
*/
export class GeneticService {
constructor() {
this.population = []; // current live population
this.generation = 0;
this.populationSize = 12;
this.mutationRate = 0.12;
this.eliteCount = 2; // elitism: always carry top N
this._historyBuf = []; // [{generation, best, avg, diversity}]
}
// ─────────────────────────────────────────────────────────────────
// Gene helpers
// ─────────────────────────────────────────────────────────────────
_randGene(def) {
return +(Math.random() * (def.max - def.min) + def.min).toFixed(4);
}
_randomGenome(overrides = {}) {
const genes = {};
for (const def of GENE_DEFS) {
genes[def.key] = overrides[def.key] !== undefined ? overrides[def.key] : this._randGene(def);
}
return genes;
}
// ─────────────────────────────────────────────────────────────────
// Fitness function — multi-objective, range [0, 1]
// ─────────────────────────────────────────────────────────────────
evaluateFitness(genes) {
// 1. Network efficiency: research_depth ≈ 0.65, exploration_rate ≈ 0.38
const netEff = Math.max(0,
1 - Math.abs(genes.research_depth - 0.65) * 1.4
- 0.6 * Math.abs(genes.exploration_rate - 0.38)
);
// 2. Quality gate: validation_strictness × (1 - publication_rate × 0.4)
// High strictness + moderate rate = good. Too loose + too fast = spam.
const qualityGate = Math.min(1,
genes.validation_strictness * (1 - genes.publication_rate * 0.45) * 1.25
);
// 3. Consensus health: threshold sweet-spot around 0.68
const consensusScore = genes.consensus_threshold >= 0.5
? Math.max(0, 1 - Math.abs(genes.consensus_threshold - 0.68) * 2.5)
: genes.consensus_threshold * 0.6;
// 4. Collaboration balance: neither isolated (→0) nor echo chamber (→1)
const collabScore = Math.max(0, 1 - Math.abs(genes.collaboration_weight - 0.55) * 2.2);
// 5. Fault tolerance: monotone reward, strongly penalise < 0.5
const resilienceScore = genes.fault_tolerance >= 0.5
? genes.fault_tolerance
: genes.fault_tolerance * 0.4;
// 6. Convergence speed: sweet-spot at 0.45
const convergenceScore = Math.max(0, 1 - Math.abs(genes.convergence_speed - 0.45) * 2.4);
// Weighted aggregate
const raw =
netEff * 0.22 +
qualityGate * 0.22 +
consensusScore * 0.16 +
collabScore * 0.14 +
resilienceScore * 0.14 +
convergenceScore * 0.12;
return Math.max(0, Math.min(1, raw));
}
_fitnessComponents(genes) {
const fc = {
network_efficiency: Math.max(0, 1 - Math.abs(genes.research_depth - 0.65)*1.4 - 0.6*Math.abs(genes.exploration_rate - 0.38)),
quality_gate: Math.min(1, genes.validation_strictness * (1 - genes.publication_rate * 0.45) * 1.25),
consensus_health: genes.consensus_threshold >= 0.5 ? Math.max(0, 1 - Math.abs(genes.consensus_threshold - 0.68)*2.5) : genes.consensus_threshold*0.6,
collaboration_balance: Math.max(0, 1 - Math.abs(genes.collaboration_weight - 0.55)*2.2),
resilience: genes.fault_tolerance >= 0.5 ? genes.fault_tolerance : genes.fault_tolerance*0.4,
convergence_score: Math.max(0, 1 - Math.abs(genes.convergence_speed - 0.45)*2.4),
};
return Object.fromEntries(Object.entries(fc).map(([k, v]) => [k, +v.toFixed(4)]));
}
// ─────────────────────────────────────────────────────────────────
// Genetic operators
// ─────────────────────────────────────────────────────────────────
/** Tournament selection — picks best of k random candidates */
_tournamentSelect(pop, k = 3) {
const candidates = [];
for (let i = 0; i < k; i++) candidates.push(pop[Math.floor(Math.random() * pop.length)]);
return candidates.reduce((best, c) => (c.fitness > best.fitness ? c : best));
}
/** Uniform crossover — each gene inherited independently with 50% probability */
_crossover(parentA, parentB) {
const childGenes = {};
for (const def of GENE_DEFS) {
childGenes[def.key] = Math.random() < 0.5 ? parentA.genes[def.key] : parentB.genes[def.key];
}
return childGenes;
}
/** Gaussian mutation — perturbs each gene with probability `rate` */
_mutate(genes, rate = this.mutationRate) {
const mutated = { ...genes };
for (const def of GENE_DEFS) {
if (Math.random() < rate) {
const sigma = (def.max - def.min) * 0.10;
// Box-Muller approximation
const delta = (Math.random() + Math.random() - 1) * sigma;
mutated[def.key] = +(Math.max(def.min, Math.min(def.max, genes[def.key] + delta))).toFixed(4);
}
}
return mutated;
}
// ─────────────────────────────────────────────────────────────────
// Population management
// ─────────────────────────────────────────────────────────────────
/** Seed a fresh random population (resets generation counter) */
seedPopulation(size = this.populationSize) {
this.population = [];
this.generation = 0;
this._historyBuf = [];
this.populationSize = size;
for (let i = 0; i < size; i++) {
const genes = this._randomGenome();
const fitness = this.evaluateFitness(genes);
const genome = this._buildGenome(`genome-g0-${i}`, 0, [], genes, fitness, 'EVALUATED');
this.population.push(genome);
db.get('genetic_population').get(genome.id).put(gunSafe(genome));
}
const stats = this.getStats();
this._historyBuf.push(stats);
db.get('genetic_stats').put(gunSafe({ ...stats, timestamp: Date.now() }));
db.get('genetic_history').get(`g0`).put(gunSafe(stats));
return this.population;
}
/** Evolve one full generation (selection → crossover → mutation → elitism) */
evolveGeneration() {
if (this.population.length < 2) throw new Error('Population too small — seed first (minimum 2)');
const sorted = [...this.population].sort((a, b) => b.fitness - a.fitness);
const nextGen = [];
// Elitism: carry over top N unchanged
for (let i = 0; i < this.eliteCount && i < sorted.length; i++) {
nextGen.push({ ...sorted[i], status: 'ELITE' });
}
// Generate offspring via tournament → crossover → mutation
while (nextGen.length < this.populationSize) {
const pa = this._tournamentSelect(sorted, 3);
const pb = this._tournamentSelect(sorted, 3);
let childGenes = this._crossover(pa, pb);
childGenes = this._mutate(childGenes);
const fitness = this.evaluateFitness(childGenes);
const idx = nextGen.length;
const child = this._buildGenome(
`genome-g${this.generation + 1}-${idx}`,
this.generation + 1,
[pa.id, pb.id],
childGenes,
fitness,
'EVALUATED'
);
nextGen.push(child);
}
this.generation++;
this.population = nextGen;
// Persist to Gun
for (const g of nextGen) {
db.get('genetic_population').get(g.id).put(gunSafe(g));
}
const stats = this.getStats();
this._historyBuf.push(stats);
db.get('genetic_stats').put(gunSafe({ ...stats, timestamp: Date.now() }));
db.get('genetic_history').get(`g${this.generation}`).put(gunSafe(stats));
return { generation: this.generation, population: nextGen, stats, history: this._historyBuf };
}
/** Manual crossover of two specific genomes by ID */
crossoverById(idA, idB) {
const pa = this.population.find(g => g.id === idA);
const pb = this.population.find(g => g.id === idB);
if (!pa) throw new Error(`Genome ${idA} not found`);
if (!pb) throw new Error(`Genome ${idB} not found`);
let childGenes = this._crossover(pa, pb);
childGenes = this._mutate(childGenes, 0.05); // light mutation for manual cross
const fitness = this.evaluateFitness(childGenes);
const child = this._buildGenome(
`genome-cross-${Date.now().toString(36)}`,
Math.max(pa.generation, pb.generation) + 1,
[pa.id, pb.id],
childGenes,
fitness,
'MANUAL_CROSS'
);
this.population.push(child);
db.get('genetic_population').get(child.id).put(gunSafe(child));
return child;
}
_buildGenome(id, generation, parent_ids, genes, fitness, status) {
return {
id,
generation,
parent_ids,
genes,
fitness: +fitness.toFixed(4),
fitness_components: this._fitnessComponents(genes),
status,
born_at: Date.now(),
};
}
// ─────────────────────────────────────────────────────────────────
// Stats & population access
// ─────────────────────────────────────────────────────────────────
getStats() {
if (this.population.length === 0) {
return { generation: this.generation, size: 0, best: 0, avg: 0, diversity: 0, elites: [] };
}
const fits = this.population.map(g => g.fitness);
const best = Math.max(...fits);
const avg = fits.reduce((a, b) => a + b, 0) / fits.length;
// Average pairwise Euclidean gene distance (normalised)
let divSum = 0, pairs = 0;
for (let i = 0; i < this.population.length; i++) {
for (let j = i + 1; j < this.population.length; j++) {
let dist = 0;
const ga = this.population[i].genes;
const gb = this.population[j].genes;
for (const def of GENE_DEFS) dist += Math.abs(ga[def.key] - gb[def.key]);
divSum += dist / GENE_DEFS.length;
pairs++;
}
}
const diversity = pairs > 0 ? divSum / pairs : 0;
return {
generation: this.generation,
size: this.population.length,
best: +best.toFixed(4),
avg: +avg.toFixed(4),
diversity: +diversity.toFixed(4),
elites: this.population.filter(g => g.status === 'ELITE').map(g => ({ id: g.id, fitness: g.fitness })),
};
}
getHistory() { return this._historyBuf; }
async getPopulation() {
if (this.population.length > 0) return this.population;
// Fallback: load from Gun (e.g. after server restart)
return new Promise((resolve) => {
const pop = [];
db.get('genetic_population').map().once((data) => {
if (data && data.id && data.genes) pop.push(data);
});
setTimeout(() => {
this.population = pop.sort((a, b) => (b.fitness || 0) - (a.fitness || 0));
resolve(this.population);
}, 1500);
});
}
// ─────────────────────────────────────────────────────────────────
// Legacy: code mutation sandbox (unchanged interface)
// ─────────────────────────────────────────────────────────────────
async submitProposal(agentId, { title, description, code, logicType = 'protocol' }) {
const proposalId = `mutation-${Date.now()}-${Math.random().toString(36).substring(2, 9)}`;
const proposal = {
id: proposalId, author: agentId, title, description, code, logicType,
status: 'PENDING_SANDBOX', consensusWeight: 0, timestamp: Date.now(), results: null,
};
db.get('genetic_tree').get(proposalId).put(gunSafe(proposal));
this.runSandboxCheck(proposalId, code);
return proposalId;
}
async runSandboxCheck(proposalId, code) {
const result = await sandbox.execute(code, { memory: '64m', cpus: '0.2', timeout: 5000 });
const status = result.success ? 'SANDBOX_PASSED' : 'SANDBOX_FAILED';
db.get('genetic_tree').get(proposalId).put(gunSafe({
status,
results: {
success: result.success,
exitCode: result.exitCode,
stdout: (result.stdout || '').slice(0, 500),
stderr: (result.stderr || '').slice(0, 300),
},
}));
console.log(`[GENETIC] Proposal ${proposalId} → ${status}`);
}
async getGeneticTree() {
return new Promise((resolve) => {
const tree = [];
db.get('genetic_tree').map().once((data) => { if (data && data.title) tree.push(data); });
setTimeout(() => resolve(tree), 1500);
});
}
}
export const geneticService = new GeneticService();
|