Precise Patterns

🚧 Work in Progress — Early Development Stage 🚧

Precise Patterns is a Python library for real-time detection of stock market chart patterns. It provides flexible candle aggregation, pivot detection, and pattern recognition tools designed for live trading systems and historical backtesting workflows.

Supports: Python 3.10 or higher

---

✨ Features

precise_patterns is the next major iteration of my earlier project, Stock-Pattern.

• Multi-timeframe candle aggregation

  • Intraday (minute-based) and end-of-day aggregation.
  • Supports arbitrary custom minute frequencies (e.g., 3, 5, 15, 75, 240…)
  • Accurate handling of sessions, weekends, and holidays
  • Uses event-driven architecture for streaming updates

• Pivot point identification

  • Detect local highs and lows with configurable rules
  • Automatically triggered from candle close events

• Pattern detection engine

  • A fresh, reimagined approach to pattern detection algorithms (in progress)
  • Pluggable architecture for future expansion

• Typed data models

  • Candle, OHLC, and Pivot
  • Built using TypedDict for structural validation and IDE code completion for fields.

---

🔌 Event-Driven Architecture

precise_patterns is built around a central event bus:

• Aggregators emit "candle.close" events with a Candle payload
• Pivot and pattern modules subscribe and react to those events
• Enables real-time streaming and modular processing pipelines

Users may register their own handlers to stream or store detected structures.

---

     ╭──────────────╮
     │  DataReader  │
     │   csv, db,   │
     │  websocket   │
     ╰──────┬───────╯
          Candle
            ▼
     ╭──────────────╮
     │  Aggregator  │
     ╰──────┬───────╯
      candle.closed
    ╭───────┴─────────╮
    ▼                 ▼
╭─────────────╮  ╭──────────╮
│PivotDetector│  │  Storage │
╰─────┬───────╯  ╰────┬─────╯
      │               ▲
  pivot.formed   pattern.formed
      ▼               │
╭───────────────╮     │   Output
│PatternDetector│──────► patterns
╰───────────────╯        to Chart

🔧 Components Overview

Component	Purpose
CandleBuilder	Accumulates OHLC updates within a timeframe
MinuteAggregator	Groups tick or 1m candles into higher intraday timeframes
EODAggregator	Resamples daily candles into D/W/M/Q intervals
PivotDetector	Emits local pivot highs/lows as they form

All components are designed to be composable and stateful.

Notes

• The CSVReader is an experimental custom CSV reader. It is not fully tested.
• The CSVStorage class is temporary class written for debug (not feature complete)
• Both of the above classes may not exist in the final implementation.

Example

See example.py for a partial implementation.

Documentation

To generate the docs using sphinx:

1. Install dependencies: pip install sphinx furo

2. Build the docs: sphinx-build docs/source docs/build

3. Open docs/build/index.html in your browser.

---

📦 Installation

The package has not been published yet.

For those wishing to tinker: clone or download the repo and run pip install <path to repo>

To install doc dependencies, pip install <path to repo>[docs]

Task and Future roadmap

See tasks.md