Metadata-Version: 2.4
Name: XXSim
Version: 0.13.0
Summary: Exchange Execution Simulator - stock exchange single bar execution simulator
Author-email: Yanir Taflev <yanirta+xxsim@gmail.com>
License: MIT
Keywords: trade,Algotrading,simulation,trading,exchange,stocks,backtesting,stock market,execution simulator,OHLCV,order simulation,finance,trading bot,quantitative,algorithmic trading,market data
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.10
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: numpy
Requires-Dist: pydantic>=2.0
Requires-Dist: xtrading-models>=0.9.0
Provides-Extra: dev
Requires-Dist: plotly; extra == "dev"
Requires-Dist: pandas; extra == "dev"
Dynamic: license-file

# XXSim
Exchange Execution Simulator, development package.
This package simulates order(s) execution based on provided market OHLCV data.
This is not an independent package, but its the core logic required to implement backtesting when it comes on running on candlebar data.

**This is the first drop and wasn't tested out in the wild yet, more functionality and capabilities to come**

## disclaimer
**Own risk warning** - Execution prices are best estimations base on worst case scenarios and statistics, there will be price differences between simulations and real-world execution, using this package the user acknowledges his consent and takes full responsibility on the implications caused due to any error or misinterpetation of this package and it's results.

Trailing commands warning - Trailing commands are currently roughly estimated and carry high deviation from the real world

## The challenge
Core Problem: Reconstructing intra-bar price movement to determine order execution.

The golden standard of market data comes in chunks of Candle-bars providing Open, High, Low, Close and Volume of predefined time range
ie. 1-minute, 5-minutes, an hour, a day, a week, a month, etc...
Within each such data-unit there is a gap of the inner price motions, unless you work with tick-by-tick data which is expensive, noisy and resource intense.
On top of that, fill prices are results of a consiquent rules and formations that are hard to simulate.

## The solution
Output: Realistic execution fills within statistical uncertainty. or consiquent order either original or modified.

XSim relies on OHLC Data to simulate the inner motion of prices within single data-unit, and attempt to perform a set of decision to execute orders in the most authentic way.

## Supported order types
- MarketOrder
- LimitOrder
- StopOrder
- StopLimitOrder
- TrailingStopMarket

## Not supported order types (at the moment)
- Trailing Stop Limit Orders
- Market-on-Close (MOC) / Limit-on-Close (LOC)
- Market-if-Touched (MIT)
- Others...

## Current Execution algorithm assumptions
- **Optional fill drift model** — volatility-based normal distribution drift, configurable via `ExecutionConfig`
  - `fill_drift_model="none"` (default): deterministic fills at exact price
  - `fill_drift_model="normal"`: drift drawn from `N(0, bar_range / std_divider)`
  - Direction follows intra-bar price movement (not order side) — fill drifts where the market moves
  - Trail orders use next price fragment for direction; other orders use close vs fill price
  - Result always clamped to `[bar.low, bar.high]`
  - Seeded RNG via `random_seed` for reproducible backtests
- No partial fills
- Aggressive approach - Order will be filled if there's a possible path between order's formation and the candlebar.
- Trail orders assume the following order of the candles:
On Bullish bar: prev_extremePrice [optional] -> open -> low -> high -> close
On Bearish bar: prev_extremePrice [optional] -> open -> high -> low -> close

## Installation

```bash
pip install XXSim
```

## Usage

### Single-Bar Execution
For direct single-bar execution using the low-level engine:
```python
from XXSim import ExecutionEngine, MarketOrder, BarData
from datetime import datetime

engine = ExecutionEngine()
bar = BarData(
    date=datetime(2025, 1, 1, 9, 30),
    open=100.00,
    high=105.00,
    low=95.00,
    close=102.00,
    volume=1000000,
)
order = MarketOrder(action='BUY', totalQuantity=100)
fills = engine.execute(order, bar)
print(fills)
```

### Multi-Bar Simulation
For backtesting across multiple bars with order lifecycle management:
```python
from XXSim import Simulator, MarketOrder, LimitOrder, BarData

sim = Simulator()

# Register callbacks
sim.on_fill(lambda trade, fill: print(f"Filled: {fill.execution.price}"))
sim.on_cancel(lambda trade: print(f"Cancelled: {trade.log[-1].message}"))

# Submit orders
sim.submit_order(MarketOrder(action='BUY', totalQuantity=100))
sim.submit_order(LimitOrder(action='SELL', totalQuantity=100, price=110.0))

# Process bars
for bar in historical_bars:
    fills = sim.process_bar(bar)

# Or use ib_insync-style event loop
def strategy(bar, fills):
    if should_buy(bar):
        sim.submit_order(MarketOrder(action='BUY', totalQuantity=100))

sim.on_bar(strategy)
sim.run(historical_bars)
```

The Simulator supports:
- **Order management**: submit, cancel, update, query orders
- **TIF (Time-In-Force)**: GTC, DAY (expires on date change), GTD (expires after date)
- **GAT (Good After Time)**: Orders with `goodAfterTime` are not active until that time
- **OCO (One-Cancels-Other)**: Link orders via `ocaGroup` - when one fills, siblings are cancelled
- **Callbacks**: on_fill, on_cancel, on_update, on_bar

## Development

## Running Tests

```bash
pytest tests/ -v
```

## Visualizations
Stop-limit and Trailing test cases can be visualized
For stop-limit visualization run:

```bash
python docs/stop-limit-chart-generator.py test-data/stop-limit/<filename.csv>
```

For trailing visualization run:

```bash
python docs/trailing-stop-chart-generator.py test-data/trailing-stop/<filename.csv>
```

TODO: Trailing cases are not organized systemactically enough

### Contributing
Contributions are welcome! Please see the documentation in the `docs` folder for more details and test specifications.
