43 Commits

Author SHA1 Message Date
Simon Moisy
ada6150413 Added multiple technical indicators for feature engineering, including ADX, TRIX, Vortex, KAMA, Force Index, EOM, MFI, ADI, TEMA, StochRSI, and Awesome Oscillator. Improved NaN handling and implemented leave-one-out feature evaluation with results saved to CSV. 2025-05-30 17:59:09 +08:00
Simon Moisy
ced64825bd reverted to sequential computing for features, added one distribution visualization graph 2025-05-30 15:54:48 +08:00
Simon Moisy
2f98463df8 more uv updates 2025-05-30 15:54:14 +08:00
Simon Moisy
2a52ffde9a cleanup and uv updates 2025-05-30 15:36:43 +08:00
Simon Moisy
a22914731f gitignore updated, model file 2025-05-30 12:31:20 +08:00
Simon Moisy
81e4b640a7 model updated 2025-05-30 12:29:37 +08:00
Simon Moisy
2dba88b620 Added mode indicators, still WIP 2025-05-29 12:45:45 +08:00
Simon Moisy
de67b27e37 XGBoost first iteration 2025-05-29 18:28:53 +08:00
Simon Moisy
1284549106 progress print 2025-05-29 11:04:03 +08:00
Simon Moisy
5f03524d6a never fallback to default values for fee_usd 2025-05-28 02:50:40 +08:00
Simon Moisy
74c8048ed5 shifted one day back on the metatrend to avoid lookahead bias, reverted metatrend calculus to use no cpu optimization for readability 2025-05-27 17:49:55 +08:00
Simon Moisy
2fd73085b8 Refactor backtest logic for improved index retrieval
- Updated the method for determining the start index of the current trade to directly use the DataFrame index, enhancing clarity and performance.
- Removed the deprecated get_current_min1_end_idx method to streamline the codebase.
2025-05-21 17:06:16 +08:00
Simon Moisy
806697116d Refactor backtesting logic and introduce new components
- Replaced TrendDetectorSimple with a new Backtest class for improved backtesting functionality.
- Integrated argparse for configuration file input, allowing dynamic parameter setting.
- Added MarketFees and Supertrends classes to handle fee calculations and trend detection, respectively.
- Removed deprecated main_debug.py and trend_detector_simple.py files to streamline the codebase.
- Enhanced process_timeframe_data to utilize the new Backtest class for executing trades and calculating results.
- Updated Storage class to support writing backtest results with metadata.
2025-05-21 17:03:34 +08:00
Simon Moisy
14905017c8 Add total fees calculation to storage results
- Included total_fees_usd in the results dictionary of the Storage class to enhance fee tracking in the output.
- removed plots from TrendDetectorSimple
2025-05-21 15:35:12 +08:00
Simon Moisy
ec1a86e098 Fixing last merge 2025-05-21 15:14:00 +08:00
Simon Moisy
0a919f825e Merge branch 'main' of ssh://dep.sokaris.link:2222/Simon/Cycles 2025-05-21 15:06:56 +08:00
Simon Moisy
c2886a2aab Enhance trading logic and fee calculations in main.py and trend_detector_simple.py
- Added total fees calculation to process_timeframe_data and aggregate_results functions in main.py.
- Updated trade logging in TrendDetectorSimple to include transaction fees in USD.
- Introduced calculate_okx_fee function for consistent fee calculations based on maker/taker status.
- Adjusted backtesting logic to account for fees when buying and selling, ensuring accurate profit calculations.
- Expanded stop loss percentages and timeframes for broader analysis in main.py.
2025-05-21 14:54:44 +08:00
Ajasra
10cc047975 Merge branch 'main' with resolved conflicts 2025-05-20 18:44:24 +08:00
Ajasra
955a340d02 docs 2025-05-20 18:40:16 +08:00
Ajasra
07b9824b69 docs 2025-05-20 18:36:59 +08:00
Ajasra
369b3c1daf added seaborn to plot ( a bit simpler and more features than matplotlib) 2025-05-20 18:29:12 +08:00
Ajasra
08c871e05a Boilinger Band and RSI implementation 2025-05-20 18:28:53 +08:00
Ajasra
837c505828 Refactor 2025-05-20 16:59:17 +08:00
Ajasra
1cdfe3973a Uv package manager 2025-05-20 16:22:11 +08:00
Simon Moisy
8ff86339d6 Add taxes functionality and refactor trading logic
- Introduced Taxes class in taxes.py to calculate and apply taxes on profits in backtest results.
- Updated main.py to include tax calculations in the results processing flow.
- Refactored trade logging in TrendDetectorSimple to account for transaction fees and ensure accurate profit calculations.
- Added a utility script (apply_taxes_to_file.py) for applying taxes to existing CSV files.
- Adjusted date range and timeframe settings in main.py for broader analysis.
2025-05-20 16:14:40 +08:00
Simon Moisy
7f788a4d4e Update .gitignore to include JSON files for version control, ensuring proper handling of data formats in the project. 2025-05-19 18:04:24 +08:00
Simon Moisy
0eb7fc77f9 Enhance data loading in main.py to support both CSV and JSON formats, improving flexibility for input data. Update date filtering and logging for better traceability. Refactor trade collection logic to ensure all trades are captured and written to separate CSV files by timeframe and stop loss percentage. Adjusted main execution parameters for broader timeframe analysis. 2025-05-19 17:35:17 +08:00
Simon Moisy
170751db0e Implement Google Sheets integration in main.py for batch updates of backtest results
- Added GSheetBatchPusher class to handle background updates to Google Sheets.
- Refactored write_results_per_combination function to write results directly to Google Sheets instead of CSV files.
- Updated process_timeframe function to handle single stop loss percentages.
- Introduced a global queue for batching results and trades for efficient updates.
- Enhanced error handling for Google Sheets API quota limits.
- Adjusted main execution flow to start the batch pusher and ensure all results are pushed after processing.
2025-05-19 02:02:03 +08:00
Simon Moisy
f7f0fc6dd5 Cleanup 2025-05-17 14:48:30 +08:00
Simon Moisy
e4ded694b1 Update .gitignore to exclude PNG files from version control 2025-05-17 14:48:13 +08:00
Simon Moisy
fa12bcb61a Update .gitignore to exclude CSV files from version control 2025-05-17 14:47:51 +08:00
Simon Moisy
125d4f7d52 Add BacktestCharts class for visualizing backtest results and update main.py for enhanced data processing
- Introduced BacktestCharts class in charts.py to plot profit ratio vs stop loss and average trade vs stop loss for different timeframes.
- Updated main.py to integrate new charting functionality and streamline data processing without monthly splits.
- Enhanced backtesting logic in TrendDetectorSimple to include transaction costs and improved stop loss handling using 1-minute data for accuracy.
- Added functionality to write results to individual CSV files for better organization and analysis.
2025-05-17 13:07:40 +08:00
Simon Moisy
ec8b1a7cf2 Enhance main.py with optimized data loading, logging setup, and concurrent processing for backtesting. Introduce new functions for data handling and results aggregation. Update TrendDetectorSimple to support meta supertrend backtesting and improve SuperTrend calculations with caching and parallel execution. Refactor TrendDetectorMACD for better performance in trend detection. 2025-05-16 02:44:22 +08:00
Simon Moisy
7c4db08b1b Refactor main.py to remove commented-out cycle detection code and streamline trend plotting. Update TrendDetectorSimple by eliminating unnecessary logging statements for improved clarity during trend detection. 2025-05-09 15:24:10 +08:00
Simon Moisy
f316571a3c Update date filtering in main.py and enhance TrendDetectorSimple with SuperTrend calculations and improved plotting functionality. Introduce color configurations for better visualization and streamline trend analysis methods. 2025-05-09 15:17:30 +08:00
Simon Moisy
c7732881c5 Refactor TrendDetectorSimple by removing redundant method definitions for trend detection and plotting, streamlining the code for better readability. 2025-05-09 12:27:40 +08:00
Simon Moisy
b0ffedc6af Update .gitignore to include additional files and modify README header 2025-05-09 12:26:42 +08:00
Simon Moisy
e9bfcd03eb Refactor cycle detection and trend analysis; enhance trend detection with linear regression and moving averages. Update main script for improved data handling and visualization. 2025-05-09 12:26:42 +08:00
cbc6a7493d More analysis WIP 2025-05-08 16:23:25 +08:00
1decd0d0ce Refactor trend detection logic in trend_detector_simple.py by removing the implementation for finding valleys between peaks 2025-05-06 16:22:20 +08:00
e1f1e7f561 Refactor trend detection logic and update date filtering in main.py. Removed width parameter from detect_trends method and enhanced peak detection to include valleys and peaks between consecutive points. 2025-05-06 16:20:43 +08:00
Simon Moisy
0bbe308321 first trend detection WIP 2025-05-06 15:24:36 +08:00
e2cd746dc6 Initial commit 2025-04-14 06:46:26 +00:00