opt_cpp — C++17 pybind11 optimizers

This module provides a minimal‑memory C++ backend for portfolio optimization.

Delivered bindings

• NSGA‑II multi‑objective optimizer: nsga2_optimize(returns, objectives, params, lo, hi, popsize, ngen, seed, crossover_prob, mutation_prob, w_prev=None)
• Projection utility: prop_box_sum1(w, lo, hi, iters=6)

Design goals

• C++17 + pybind11 module (opt_cpp)
• Zero‑copy NumPy views; no STL copies of inputs
• Reuse preallocated buffers; O(T·N + popsize·N) working memory
• Optional OpenMP parallel evaluation
• Deterministic for fixed seed

Build

1. Ensure a C++17 compiler and Python dev headers are available.

2. From repo root:

   mkdir -p build && cd build cmake -S ../cpp -B . -DOPTCPP_USE_OPENMP=ON cmake --build . --config Release

3. Add the build dir to PYTHONPATH or install the built module

   export PYTHONPATH=$PWD:$PYTHONPATH python -c "import opt_cpp; print(opt_cpp)"

Usage

Python

import numpy as np, opt_cpp R = np.random.normal(0, 0.01, size=(2000, 30)).astype(np.float64) W, S = opt_cpp.nsga2_optimize(R, ["sharpe","min_vol","cvar"], {"alpha":0.05, "ann":252.0}, lo=0.0, hi=0.3, popsize=64, ngen=20, seed=42, crossover_prob=0.9, mutation_prob=0.1) print(W.shape, S.shape) # Pareto front sizes

Notes

• Objectives match Python implementations used in this repo (annualized Sharpe/Sortino/Calmar, CVaR, Kelly, diversification ratio, target‑min‑vol, etc.).
• For return_to_turnover, if w_prev is None the L1 norm of w is used in the denominator.
• Diversification ratio uses per‑asset standard deviations from the input returns (no full covariance allocation).
• Compile with -DOPTCPP_FLOAT32=ON to switch to float32.

Docker

Local build and run of the full app (API + Next.js UI):

1. Build and start services

   docker compose up --build

   • API: http://localhost:8000/docs
   • UI: http://localhost:3000

2. Data cache persistence

   Parquet data downloaded by the API is persisted on the host in data_management/data/ via a bind mount.

3. Environment

   • Frontend proxies /opt/* and /data/* to the API using API_URL (defaults to http://backend:8000 in compose).
   • You can toggle API debug logs by setting DM_DEBUG=true in docker-compose.yml.

Git Init & Push

This repo now includes a comprehensive .gitignore to exclude venv/, node_modules/, build artifacts, and cached parquet data.

Initialize and push to your remote:

1. Initialize the repository

   git init git add . git commit -m "Initial commit"

2. Add your remote and push

   git remote add origin <YOUR_REMOTE_URL> git branch -M main git push -u origin main

Quant_app_JS

Quant_app_JS

Quant_app_Note.js

Quant_app_JS

Quant_app_JS