@vibedcoder/invespro-core

Core engine for Invespro, a rules-based investment profiling and portfolio allocation system.

This package evaluates applicant answers against the default profiling model or a versioned custom definition, then returns a normalized score, risk profile, allocation, and definition metadata.

Installation

pnpm add @vibedcoder/invespro-core

npm install @vibedcoder/invespro-core

Quick Start

import { RiskProfilerEngine } from '@vibedcoder/invespro-core';

const engine = new RiskProfilerEngine();

const result = await engine.evaluate({
  applicantId: 'APP-001',
  answers: {
    investmentHorizonYears: 10,
    riskAttitude: 'hold',
    investmentObjective: 'balanced_growth',
    annualIncome: 75000,
    dtiRatio: 20,
    liquidityMonths: 4,
    investmentExperience: 'intermediate',
  },
});

console.log(result.profile.label);
console.log(result.allocation);

Custom Definitions

You can compile and run a custom, versioned profiler definition while keeping the standard Invespro result contract.

import { RiskProfilerEngine } from '@vibedcoder/invespro-core';

const engine = new RiskProfilerEngine({ definition: customDefinition });

Custom definitions can change question labels, options, scores, weights, bands, profile IDs, and allocations. Expert custom JDM graphs are supported when they follow the Invespro graph contract.

CSV Batch Input

parseCsvBatch converts CSV rows into the same batch item shape accepted by RiskProfilerEngine.evaluateMany.

import {
  DEFAULT_RISK_PROFILE_DEFINITION,
  parseCsvBatch,
  RiskProfilerEngine,
} from '@vibedcoder/invespro-core';

const csv = [
  'applicantId,investmentHorizonYears,riskAttitude,investmentObjective,annualIncome,dtiRatio,liquidityMonths,investmentExperience',
  'APP-001,10,hold,balanced_growth,75000,20,4,intermediate',
].join('\n');

const items = parseCsvBatch(csv, DEFAULT_RISK_PROFILE_DEFINITION);
const engine = new RiskProfilerEngine();
const result = await engine.evaluateMany({ items });

CSV columns should use the active definition's question IDs. The parser handles numbers, booleans, select option values, optional applicantId, and omitted empty cells.

Result Shape

{
  "applicantId": "APP-001",
  "rawScore": 38,
  "normalizedScore": 67.86,
  "profile": {
    "id": "moderatelyAggressive",
    "label": "Moderately Aggressive"
  },
  "overrideApplied": false,
  "allocation": {
    "equities": 70,
    "fixedIncome": 20,
    "cash": 5,
    "alternatives": 5
  },
  "definition": {
    "id": "invesproDefaultRiskProfiler",
    "version": "0.1.0",
    "schemaVersion": "1.0",
    "graphChecksum": "sha256:..."
  }
}

Related Packages

• @vibedcoder/invespro-types provides schemas and types.
• @vibedcoder/invespro-hono exposes this engine through REST.
• @vibedcoder/invespro-cli evaluates inputs from the command line.

See the main project documentation at github.com/vibedcoder/invespro.