sigc

The Quant's Compiler: From Alpha Idea to Production in Minutes

A type-safe DSL and high-performance runtime for quantitative trading strategies.

Write signals like sentences. Backtest in milliseconds. Deploy with confidence.

Documentation | Crates.io | GitHub

---

Why sigc?

Every quant team that scales past a handful of researchers eventually builds the same thing: a typed, reproducible signal language. We've seen it at every major desk. sigc is that system — open-sourced.

Problem	sigc Solution
Ad-hoc notebooks with mismatched calendars	Type-safe DSL with compile-time shape checking
Non-deterministic backtests	Content-addressed caching, identical inputs = identical outputs
Fragile pandas joins at 3am	120+ vectorized operators, SIMD-optimized
"It worked in research" → fails in prod	Same binary: sigc run → sigc daemon
Factor code copy-pasted across teams	Macros, functions, importable signal libraries

Quick Start

# Install from crates.io
cargo install sigc

# Or build from source
git clone https://github.com/skelf-Research/sigc.git
cd sigc && cargo build --release

Create momentum.sig:

data:
  prices: load parquet from "prices.parquet"

params:
  lookback = 20
  top_pct = 0.2

signal momentum:
  returns = ret(prices, lookback)
  score = zscore(returns)
  emit winsor(score, p=0.01)

portfolio main:
  weights = rank(momentum).long_short(top=top_pct, bottom=top_pct)
  backtest from 2024-01-01 to 2024-12-31

Run it:

$ sigc run momentum.sig

=== Backtest Results ===
Total Return:      15.23%
Sharpe Ratio:       1.45
Max Drawdown:       8.12%
Turnover:         312.00%

For Quants: Express Alpha, Not Boilerplate

Write signals that read like your research notes:

// Momentum with skip-month
signal momentum:
  total_ret = ret(prices, 252)
  skip_ret = ret(prices, 21)
  mom = total_ret - skip_ret
  emit zscore(mom)

// Mean reversion on residuals
signal stat_arb:
  beta = rolling_beta(stock, market, 60)
  residual = stock - beta * market
  z = (residual - rolling_mean(residual, 20)) / rolling_std(residual, 20)
  emit -z

// Combine factors with explicit weights
signal multi_factor:
  emit 0.4 * momentum + 0.3 * value + 0.3 * quality

120+ operators built-in: zscore, rank, rolling_mean, ema, rsi, macd, atr, vwap, neutralize, winsor, and more.

For Developers: Rust-Native, Zero Runtime Overhead

use sigc::{Strategy, Backtest};

// Embed in your systems
let strategy = Strategy::from_file("alpha.sig")?;
let results = strategy
    .with_param("lookback", 60)
    .with_param("threshold", 1.5)
    .run()?;

println!("Sharpe: {:.2}", results.sharpe_ratio());

// Access raw data
let weights: Vec<f64> = results.weights();
let returns: Vec<f64> = results.daily_returns();

Architecture:

• Single binary — compiler, runtime, daemon, CLI in one executable
• Polars/Arrow columnar execution with Rayon parallelism
• SIMD kernels for factor math (AVX2/AVX-512 where available)
• Content-addressed cache (sled + blake3) for reproducibility
• nng RPC for daemon mode, sub-millisecond latency

For Architects: Production-Grade Infrastructure

# sigc.yaml - Production configuration
mode: production

safety:
  circuit_breakers:
    max_drawdown: 0.15
    max_position: 0.10
    kill_switch: true

  rate_limits:
    orders_per_minute: 100

alerts:
  slack:
    webhook: ${SLACK_WEBHOOK}
    channels: ["#trading-alerts"]

schedule:
  jobs:
    - name: rebalance
      cron: "0 9 * * 1-5"
      strategy: momentum

Enterprise features:

• Circuit breakers with automatic position flattening
• Prometheus metrics (/metrics endpoint)
• Structured audit logging (JSON, compliance-ready)
• Slack/email alerting on anomalies
• Docker & Kubernetes ready (ghcr.io/skelf-Research/sigc)

Crate Ecosystem

Crate	Purpose
sigc	CLI and main entry point
sig_compiler	DSL parser and type checker
sig_runtime	Execution engine with 120+ operators
sig_types	Core type definitions
sig_cache	Deterministic caching layer
sig_lsp	Language server for IDE support

IDE Support

VS Code Extension with full language server:

• Real-time error diagnostics
• Hover documentation for all operators
• Code completion with signatures
• Go-to-definition for signals/functions
• 25+ snippets for common patterns

code --install-extension skelf-Research.sigc-vscode

Integrations

Integration	Status	Description
Alpaca	✅	Paper & live trading
Yahoo Finance	✅	Free market data
PostgreSQL	✅	Async connection pooling
S3/GCS	✅	Cloud data sources
Python	✅	pysigc bindings for notebooks

Performance

Benchmarked on 5 years of daily data, 500 securities:

Operation	Time
Parse + compile	2ms
Full backtest	45ms
Incremental update	3ms

Memory-mapped data loading. Warm cache hits in microseconds.

Documentation

• Full Documentation — Tutorials, API reference, deployment guides
• Quant Guide — 9-chapter deep dive for researchers
• Strategy Library — 23 ready-to-use strategies
• Operator Reference — All 120+ operators documented

Contributing

We welcome contributions. Please open an issue to discuss before submitting large PRs.

# Development setup
git clone https://github.com/skelf-Research/sigc.git
cd sigc
cargo build
cargo test

See CONTRIBUTING.md for guidelines.

License

MIT License — see LICENSE for details.

---

Built by quants, for quants.

Get Started | View Strategies | Join Discussions

Skelf Research — skelfresearch.com