Implements a TradingView-like backtester with next-bar-open entries, intrabar price path simulation (high-before-low or low-before-high based on open proximity), and SL/TP exit ordering. Computes comprehensive metrics including Sharpe ratio with timeframe-appropriate annualization, Calmar ratio, and drawdown from a combined equity curve. The engine is self-contained except for importing dataclasses from hyperview/models.py, which are generic and reusable. The intrabar path logic and metric accumulation are non-obvious and demonstrate careful engineering.

A robust configuration loader that merges defaults, primary JSON config, additional config files, and environment variable overrides (HYPERVIEW__ prefix with __ as level separator). Uses deep merge to combine nested dictionaries, coerces env values via JSON parsing, and validates the final config against a schema of allowed values and types. This pattern is reusable in any CLI tool requiring layered configuration and demonstrates clean separation of concerns.

Provides 20+ wrapped TA-Lib indicators (EMA, RSI, MACD, Stochastic, ADX, Bollinger Bands, etc.) that accept and return pandas Series, handling numpy conversion internally. Also includes pure-pandas signal helpers (crossed_above, crossed_below, barssince) and a to_unix_timestamp utility. The wrappers are clean, well-documented, and can be extracted as a standalone library for any Python trading project. The barssince implementation uses vectorized numpy operations instead of loops.

Parses Pine Script input() declarations using regex, extracts type, default, min/max, step, and options. Maps Pine input names to strategy keys via a predefined mapping (e.g., htfTf -> htf_timeframe). Resolves search dimensions for optimization, handling categorical (string, bool, timeframe, session) and numeric (int, float) types with sensible default ranges. This module is self-contained and could be reused for any Pine Script optimization tool.