rangebar

Non-lookahead range bar construction for cryptocurrency and forex trading.

This crate provides algorithms for constructing range bars from tick data with temporal integrity guarantees, ensuring no lookahead bias in financial backtesting.

Installation

Add to your Cargo.toml:

[dependencies]
rangebar = "6.1"

Meta-Crate

This is a meta-crate that re-exports all rangebar sub-crates for backward compatibility with v4.0.0. New code should depend on specific sub-crates directly:

• rangebar-core - Core algorithm and types
• rangebar-providers - Data providers (Binance, Exness)
• rangebar-config - Configuration management
• rangebar-io - I/O operations and Polars integration
• rangebar-streaming - Real-time streaming processor
• rangebar-batch - Batch analytics engine
• rangebar-cli - Command-line tools

Features

• core - Core algorithm (always enabled)
• providers - Data providers (Binance, Exness)
• config - Configuration management
• io - I/O operations and Polars integration
• streaming - Real-time streaming processor
• batch - Batch analytics engine
• full - Enable all features

Basic Usage

use rangebar::{RangeBarProcessor, AggTrade, FixedPoint};

// Create processor with 25 basis points threshold (0.25%)
let mut processor = RangeBarProcessor::new(250).unwrap();

// Create sample aggTrade
let trade = AggTrade {
    agg_trade_id: 1,
    price: FixedPoint::from_str("50000.0").unwrap(),
    volume: FixedPoint::from_str("1.0").unwrap(),
    first_trade_id: 1,
    last_trade_id: 1,
    timestamp: 1609459200000,
    is_buyer_maker: false,
    is_best_match: None,
};

// Process aggTrade records into range bars
let agg_trade_records = vec![trade];
let bars = processor.process_agg_trade_records(&agg_trade_records).unwrap();

Dual-Path Architecture

Streaming Mode (Real-time)

use rangebar::streaming::StreamingProcessor;

let threshold_decimal_bps = 250; // 25 bps = 0.25% range bars
let processor = StreamingProcessor::new(threshold_decimal_bps);
// Real-time processing with bounded memory

Batch Mode (Analytics)

use rangebar::batch::BatchAnalysisEngine;
use rangebar::core::types::RangeBar;

let range_bars: Vec<RangeBar> = vec![]; // Your range bar data
let engine = BatchAnalysisEngine::new();
// let result = engine.analyze_single_symbol(&range_bars, "BTCUSDT").unwrap();

Links

• GitHub Repository
• API Documentation
• Changelog
• Algorithm Specification

License

MIT license. See LICENSE for details.