Skip to content

v7.10.0

Choose a tag to compare

View all tags
@enarjord enarjord released this 25 Apr 17:45
· 246 commits to master since this release
v7.10.0
28b8f0c
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Release Notes for v7.10.0

These notes describe the user-facing changes from v7.9.1 to v7.10.0.

Highlights

  • Added the OHLCV v2 foundation under caches/ohlcvs/, with monthly chunk storage, SQLite catalog metadata, legacy cache import, persistent gap tracking, and v2-aware backtest preparation.
  • Added passivbot tool inspect-ohlcvs for inspecting v2 OHLCV cache coverage, chunk validity, persistent gaps, and recent fetch attempts.
  • Updated the canonical schema and mirrored example profile to configs/examples/default_trailing_grid_long_npos7.json, with config_version = v7.10.0.
  • Removed inflated grid re-entry behavior. Grid re-entries are now always normal-or-cropped, while historical inflated order-type ids remain decodable for old fills and restart compatibility.
  • Renamed collateral-agnostic strategy-equity metrics to canonical *_strategy_eq names, while keeping old *_strategy_pnl_rebased and *_hsl names as input/result aliases.
  • Added day-denominated duration metrics alongside existing hour metrics for high exposure, peak recovery, position held, and position unchanged.
  • Added richer backtest artifacts: dataset.json, strategy_equity in balance_and_equity.csv.gz, artifact loading helpers, and single-coin fill plotting for notebooks.

Upgrade Notes

  • Reinstall after upgrading:
    python3 -m pip install -e .
    or
    python3 -m pip install -e ".[full]"
  • If the Rust extension looks stale, rebuild it in the active environment:
    maturin develop --release
  • New configs should use config_version = "v7.10.0".
  • configs/examples/default_trailing_grid_long_npos10.json has been replaced by configs/examples/default_trailing_grid_long_npos7.json.
  • bot.{long,short}.entry_grid_inflation_enabled is no longer a supported runtime behavior. Older configs that still contain it are normalized during config loading.
  • Canonical optimizer configs should use *_strategy_eq metric names and day-duration metrics. Deprecated metric names remain accepted as compatibility aliases.

What Changed

OHLCV v2 Foundation

  • Added v2 OHLCV chunk storage and catalog modules for backtest-oriented candle preparation.
  • Backtests can prepare HLCV payloads from the v2 local store when coverage is already available, while keeping caches/hlcvs_data/ as the fast reusable prepared bundle cache.
  • Legacy daily OHLCV shards, including compressed .npz shards, can be imported into the v2 store.
  • Persistent exchange-side gaps are tracked so repeated fetch attempts do not endlessly retry known unavailable ranges.
  • Combined-exchange backtests now have a v2-aware load path and clearer progress logging while preparing candles.

Backtest Artifacts and Analysis

  • Backtest runs now persist dataset.json with the HLCV cache files used by the run.
  • Added src/backtest_artifacts.py helpers for loading config, analysis, fills, balance/equity, HLCVs, timestamps, BTC/USD prices, and market settings from an artifact directory.
  • Notebook workflows can call load_backtest_artifact_workspace(...), candles_for_coin(...), and plot_fills_for_coin(...).
  • balance_and_equity.csv.gz now includes collateral-agnostic strategy_equity.
  • Backtest BTC collateral is initialized at the first active trading step instead of during EMA warmup.
  • drawdown_worst_mean_1pct metrics now derive drawdowns from the full-resolution equity curve before averaging the worst 1% of daily worst drawdowns.

Metrics and Optimization

  • Canonical collateral-agnostic strategy-equity metrics now use *_strategy_eq.
  • Deprecated *_strategy_pnl_rebased and *_hsl metric names are accepted as aliases for configs, limits, visibility filters, Pareto tools, and older stored result files.
  • peak_recovery_hours_pnl now uses net realized PnL (pnl + fee_paid) and includes the open tail from the last realized-PnL peak to the end of the backtest.
  • Day-duration metric variants were added for high exposure, peak recovery, position held, and position unchanged.
  • Suite-mode limit semantics are centralized so passivbot optimize and passivbot tool pareto resolve omitted stat= consistently from backtest.aggregate.

Config and Runtime Defaults

  • The hardcoded schema defaults and mirrored example config now use a trailing-grid n_positions = 7 profile.
  • Default approved coins, scenarios, optimizer bounds, scoring, and limits were refreshed.
  • Shorts are disabled by default through zero short total wallet exposure.
  • Example configs now prefer canonical day-duration metrics where appropriate.

Order Behavior

  • Inflated grid re-entry behavior was removed from current live, backtest, and runtime paths.
  • Grid entries remain normal-or-cropped so effective wallet exposure limits are observed without pulling future size forward.
  • Historical inflated order-type ids remain readable for old fills and restart compatibility.

Short Release Summary

Passivbot v7.10.0 focuses on backtest data infrastructure, clearer strategy-equity metrics, safer grid-entry behavior, and more useful research artifacts. The main operational changes are the new OHLCV v2 foundation, the new n_positions = 7 default profile, removal of inflated grid re-entries, canonical *_strategy_eq metrics, and notebook-friendly backtest artifact helpers.

Assets 2
Loading