Replay a fixture dataset through Nautilus Evaluator

[sybapp]

Parent

• PRD: ES TradingView-to-Nautilus strategy evolution system #1

What to build

Create the first Python-side Nautilus Evaluator skeleton in this repository and prove it can consume the Versioned Dataset Contract. The evaluator should read the fixture dataset, convert bars and minimal feature data into NautilusTrader-compatible inputs, and run a smoke backtest over fixture data.

This slice should establish the evaluator package boundary without requiring real TradingView collection or strategy optimization.

Acceptance criteria

• A Python evaluator package or directory exists in the repository with a clear entry point for reading a Versioned Dataset.
• The evaluator can load the fixture dataset produced for the Versioned Dataset Contract.
• Bars are converted into NautilusTrader-compatible bar inputs.
• Feature records are loaded and available to the evaluator without calling TradingView live.
• A minimal smoke backtest can replay fixture bars in NautilusTrader.
• Tests verify dataset reading, basic conversion, and the smoke backtest path.

Blocked by

• Define and validate the Versioned Dataset Contract #2