feat: Jupyter Notebook / IPython Magic Integration (#454)

[MDUYN]

Summary

Add %%backtest cell magic and %backtest line magic for running backtests directly in Jupyter notebooks.

Closes #454

Changes

New files

• investing_algorithm_framework/notebook/__init__.py — Module entry point
• investing_algorithm_framework/notebook/magic.py — IPython magic extension
• tests/notebook/__init__.py
• tests/notebook/test_magic.py — 15 unit tests

Modified files

• investing_algorithm_framework/__init__.py — Added load_ipython_extension for %load_ext support

Usage

# Load the extension
%load_ext investing_algorithm_framework

# Cell magic — define and run a backtest inline
%%backtest --start 2023-01-01 --end 2023-12-31 --initial-amount 10000 --market BITVAVO --trading-symbol EUR -o results
from investing_algorithm_framework import TradingStrategy, DataSource, TimeUnit

class MyStrategy(TradingStrategy):
    time_unit = TimeUnit.DAY
    interval = 1
    data_sources = [
        DataSource(identifier="btc", symbol="BTC/EUR", data_type="OHLCV", time_frame="1d")
    ]

    def run_strategy(self, context, data):
        ...

# Line magic — run from an existing file
%backtest strategies/my_strategy.py --start 2023-01-01 --end 2023-12-31 -o results

Supported parameters

Flag	Description
--start	Start date (YYYY-MM-DD or YYYY-MM-DD-HH)
--end	End date (defaults to now)
--initial-amount	Initial balance (default: 1000)
--market	Market (e.g. BITVAVO)
--trading-symbol	Quote currency (e.g. EUR)
-o / --output	Store result as notebook variable
--vectorized	Use vectorized backtesting
--show-report	Display inline HTML report
--show-progress	Show progress bars
--risk-free-rate	Risk-free rate for metrics
--snapshot-interval	DAILY or TRADE_CLOSE
--no-fill-missing-data	Skip filling missing data

Backtesting modes

• Default (event-driven): calls app.run_backtest() — strategy's run_strategy is called at each time step
• --vectorized: calls app.run_vector_backtest() — uses generate_buy_signals/generate_sell_signals

Tests

All 15 tests passing:

• Date parsing (YYYY-MM-DD, YYYY-MM-DD-HH, invalid, rounding)
• Argument parser (minimal, full, strategy path, risk-free-rate, no-fill)
• Strategy class discovery (subclass, multiple, base class, non-classes)
• Extension registration (magic registration, top-level load_ext)