Commit Graph

294 Commits

Author SHA1 Message Date
Matthias
61c41fd919 Merge branch 'develop' into feat/freqai 2022-07-24 16:18:58 +02:00
Robert Caulk
88e10f7306 add exception for not passing timerange. Remove hard coded arguments for CatboostPredictionModels. Update docs 2022-07-24 09:01:23 +02:00
Robert Caulk
fff39eff9e fix multitarget bug 2022-07-24 08:42:50 +02:00
robcaulk
f3d46613ee move prediction denormalization into datakitchen. remove duplicate associated code. avoid normalization/denormalization for string dtypes. 2022-07-23 17:14:33 +02:00
Matthias
8fa6e8b4ba Remove freqAI model bridge in favor of self.freqai 2022-07-23 17:08:05 +02:00
robcaulk
c91e23dc50 let user avoid normalizing labels 2022-07-23 16:14:13 +02:00
robcaulk
50d630a155 remove unnecessary comments from data_drawer.py 2022-07-23 13:35:44 +02:00
robcaulk
a1cff377ec add record of contribution to data_kitchen.py 2022-07-23 13:32:04 +02:00
robcaulk
c2d6a0e891 add record of contribution to doc and source 2022-07-23 13:04:06 +02:00
robcaulk
5559e605b8 small PR conversation resolutions 2022-07-22 17:46:14 +02:00
robcaulk
40f00196eb use cloudpickle in place of pickle. define Paths once in data_drawer. 2022-07-22 17:37:51 +02:00
robcaulk
98c8a447b2 add LightGBMPredictionMultiModel 2022-07-22 12:40:51 +02:00
robcaulk
afcb0bec00 clean up obsolete comments, move remove_features_from_df to datakitchen 2022-07-22 12:29:20 +02:00
robcaulk
3205788bce extend doc to include descriptions of the return values from FreqAI to the strategy 2022-07-21 22:11:46 +02:00
robcaulk
8033e0bf23 add counter to backtesting log so users know how many more pairs and how many more models will need to be trained 2022-07-21 13:22:12 +02:00
robcaulk
183dec866a remove ability to backtest open ended timeranges (safer) 2022-07-21 13:02:52 +02:00
robcaulk
e694ea1cfd make sure backtesting gets the populated indicators with slimmed down user strat 2022-07-21 12:48:09 +02:00
robcaulk
ca4dd58642 remove superceded function from datakitchen 2022-07-21 12:40:54 +02:00
robcaulk
8f86b0deaa *breaking change* simplify user strat by consolidating feature loops into backend 2022-07-21 12:24:22 +02:00
robcaulk
e7337728bf add separator in folder name just incase an asset ends in an integer 2022-07-21 11:25:28 +02:00
robcaulk
921a7ef216 add requirements-freqai.txt to builds 2022-07-20 15:51:25 +02:00
robcaulk
d43c146676 add more tests for datakitchen functionalities, add regression tests for freqai_interface train/backtest 2022-07-20 12:56:46 +02:00
lolong
9c051958a6
Feat/freqai (#7105)
Vectorize weight setting, log training dates

Co-authored-by: robcaulk <rob.caulk@gmail.com>
2022-07-19 17:49:18 +02:00
robcaulk
714d9534b6 start adding tests 2022-07-19 16:16:44 +02:00
lolong
ed0f8b1189
Improve FreqAI documentation (#7072)
Improve doc + some other small fixes

Co-authored-by: robcaulk <rob.caulk@gmail.com>
2022-07-18 11:57:52 +02:00
Robert Caulk
921f3899f0 revert pickle reading for historic predictions 2022-07-17 16:06:36 +02:00
Robert Caulk
41eeb99177 load pickle file for writing 2022-07-17 10:05:21 +02:00
Robert Caulk
9d184586f1 fix bug in historic prediction saving 2022-07-16 21:16:59 +02:00
Wagner Costa Santos
ca2029a46b fix issue when bot restarts with same identifier, does not load prediction history 2022-07-14 18:55:24 -03:00
robcaulk
4141d165ff add BaseTensorFlowModel class 2022-07-12 19:10:09 +02:00
robcaulk
ef409dd345 Add ground work for TensorFlow models, add protections from common mistakes 2022-07-12 18:09:17 +02:00
Robert Caulk
fea63fba12 Fix saving/loading historic predictions 2022-07-12 10:12:50 +02:00
Robert Caulk
8ce6b18318 start collecting indefinite history of predictions. Allow user to generate statistics on these predictions. Direct FreqAI to save these to disk and reload them if available. 2022-07-11 22:01:48 +02:00
Matthias
3fc92b1b21 Create BaseRegression model - designed to reduce code duplication across currently available models. 2022-07-11 11:33:59 +02:00
Robert Caulk
607455919e Change config parameter names to improve clarity and consistency throughout the code (!!breaking change, please check discord support channel for migration instructions or review templates/FreqaiExampleStrategy.py config_examples/config_freqai_futures.example.json file changes!!) 2022-07-10 12:35:44 +02:00
Matthias
819cc9c0e4 Fully align LightGBM with Catboost 2022-07-10 11:06:18 +02:00
Matthias
58b18770e3 Fix LightGBM missing argument in predict method 2022-07-10 11:05:35 +02:00
Matthias
2e1061af64 Fix faulty LightGBM model 2022-07-09 08:21:42 +00:00
robcaulk
d9acdc9767 remove excess, increase no model warning clarity 2022-07-06 18:20:21 +02:00
robcaulk
4cac67fd66 Catch infrequent issue associated with grabbing first candle 2022-07-05 12:43:33 +02:00
robcaulk
bd3a6ba2fe update backtesting to handle new output framework 2022-07-03 17:34:44 +02:00
robcaulk
8ac8d53c32 All LGBMRegressor model parameters are now set in config 2022-07-03 16:30:01 +02:00
robcaulk
4ff0ef7359 fix bug returning multiple targets for training 2022-07-03 12:15:59 +02:00
robcaulk
ffb39a5029 black formatting on freqai files 2022-07-03 10:59:38 +02:00
robcaulk
106131ff0f Rehaul organization of return values 2022-07-02 18:09:38 +02:00
robcaulk
93e1410ed9 first step toward cleaning output and enabling multimodel training per pair 2022-07-01 14:00:30 +02:00
robcaulk
6c7d02cb18 expose nu in the SVM outlier detection via svm_nu in config 2022-06-28 15:12:25 +02:00
robcaulk
7dfbd432d1 fix config saving bug, assign priorities to pairs in scanning, sleep the scanning loop to reduce CPU usage 2022-06-27 11:35:33 +02:00
robcaulk
68bafa9517 archive config to the model folder, filter out features before returning dataframe to strategy (to alleviate frequi issues)` 2022-06-26 23:03:48 +02:00
robcaulk
051b99791d reduce unnecessary verbosity, fix error on first training sweep, add LightGBMPredictionModel 2022-06-26 19:04:23 +02:00
Robert Caulk
852706cd6b
Fix default behavior for expiration_hours 2022-06-21 08:12:51 +02:00
robcaulk
b35c64b6c0 fix bug in backtest, typo in example strat 2022-06-19 16:41:09 +02:00
robcaulk
3599d18ff6 fix bug in follow_mode, thanks @blood4rc 2022-06-18 12:05:28 +02:00
robcaulk
6da7a98857 add docstrings to new functions, remove superceded code 2022-06-17 16:16:23 +02:00
robcaulk
f631ae911b add model expiration feature, fix bug in DI return values 2022-06-17 14:55:40 +02:00
robcaulk
0b0688a91e ensure scanning purges models 2022-06-16 16:12:38 +02:00
robcaulk
c5de0c49e4 first functional scanning commit 2022-06-16 00:24:18 +02:00
robcaulk
4d472a0ea1 merging datarehaul into scanning branch 2022-06-16 00:22:49 +02:00
Matthias
c981ad4608 Fix missing space 2022-06-12 08:31:02 +02:00
robcaulk
eb47c74096 merge datarehaul into main freqai branch 2022-06-10 20:26:19 +02:00
robcaulk
d9b79d94e4 increase candle update flexibility to allow long sequential trainings that may last more than one candle 2022-06-07 20:57:10 +02:00
robcaulk
66800c7a45 ensure newest candles are always appended 2022-06-07 20:24:23 +02:00
robcaulk
15d049cffe detect if upper tf candles are new or not, append if so. Correct the epoch for candle update check 2022-06-07 19:49:20 +02:00
robcaulk
cab8f517b4 add lock to datadrawer 2022-06-07 01:07:30 +02:00
robcaulk
4b26b6aaec add lock to any historic data access 2022-06-07 00:54:18 +02:00
Robert Caulk
2451ed8c88
Quick bug fix 2022-06-06 15:11:54 -06:00
robcaulk
d6b8801f41 fix follower bug 2022-06-05 04:40:58 +02:00
robcaulk
e8c0dcf9f3 add debug message to timerange 2022-06-03 17:14:07 +02:00
robcaulk
f2762e3b4b fix bug in return_values() 2022-06-03 16:58:51 +02:00
robcaulk
16b4a5b71f rehaul of backend data management - increasing performance by holding history in memory, reducing load on the ratelimit by only pinging exchange once per candle. Improve code readability. 2022-06-03 15:19:46 +02:00
robcaulk
15a971346d catch infinity values when filtering 2022-06-02 17:13:20 +02:00
robcaulk
ace951bf7e another queue bug fix for fresh starts 2022-06-02 13:28:06 +02:00
robcaulk
833d25bda0 Save data after queue reorg 2022-06-01 08:05:38 +02:00
robcaulk
0b0dd8dd80 Force high priority pair to be trained before anything else. 2022-06-01 07:55:05 +02:00
robcaulk
7523ed825e automatically detect maximum required data based on user fed indicators (to avoid NaNs in dataset for rolling indicators), add new config parameter for backtesting to let users increase their startup_candles to accommodate high timeframe indicators, add docs to explain all. Add new feature for automatic indicator duplication according to user defined intervals (exhibited in example strat and configs now). 2022-05-31 18:42:27 +02:00
robcaulk
9b3b08a2bb let follower purge old model files 2022-05-31 15:37:38 +02:00
robcaulk
bac4ced382 Ensure follower predictions are persistent and uniquely stored 2022-05-31 14:35:04 +02:00
robcaulk
70adf55643 Automatically detect and change follower data_path to accommodate remote systems 2022-05-31 12:35:09 +02:00
robcaulk
0306f5ca13 Add autopurge feature so that FreqAI cleans up after itself when it no longer needs old models on disk 2022-05-31 11:58:21 +02:00
robcaulk
45f4f0f603 ensure follower sends back null arrays in case leader hasnt created a model file yet 2022-05-31 01:48:48 +02:00
robcaulk
29d2f59f12 fix PCA bug 2022-05-31 00:40:45 +02:00
robcaulk
606f18e5c1 Add follow_mode feature so that secondary bots can be launched with the same identifier and load models trained by the leader 2022-05-30 21:35:48 +02:00
robcaulk
5b4c649d43 detect variable sized dataframes coming from strat, adjust our stored/returned data accordingly 2022-05-30 13:55:46 +02:00
robcaulk
e229902381 fix bug in previous commit 2022-05-30 12:48:22 +02:00
robcaulk
a20651efd8 Increase performance by only predicting on most recent candle instead of full strat provided dataframe. Collect predictions and store them so that we can feed true predictions back to strategy (so that frequi isnt updating historic predictions based on newly trained models). 2022-05-30 11:37:05 +02:00
robcaulk
2f1a2c1cd7 allow users to store data in custom formats, update spot config to reflect better target horizon to training period ratio 2022-05-30 02:12:31 +02:00
robcaulk
d59eac3321 revert a79032b 2022-05-29 21:33:38 +02:00
robcaulk
a79032bf75 fixing bug in training queue 2022-05-29 20:19:32 +02:00
robcaulk
3f72263278 allow pairs deeper in the queue to get trained if the higher priority pairs dont need training 2022-05-29 20:02:43 +02:00
robcaulk
cc6cae47ec allow pairs deeper in the queue to get trained if the higher priority pairs dont need training 2022-05-29 19:49:43 +02:00
robcaulk
4eb4753e20 allow subdaily retraining for backtesting 2022-05-29 17:44:35 +02:00
robcaulk
ce365eb9e3 improve example strat so that it has dynamic buy and sell logic according to original prediction 2022-05-29 14:45:46 +02:00
robcaulk
4eb29c8810 Dont reset pair priority if it doesnt successfully train 2022-05-28 18:34:26 +02:00
robcaulk
83dd453723 catch errors occuring on background thread, and make sure to keep the ball rolling. Improve pair retraining queue. 2022-05-28 18:26:19 +02:00
robcaulk
e54614fa2f remove remnants of single threaded version, ensure pair queue priority is checked before retraining 2022-05-28 14:55:07 +02:00
robcaulk
2a4d1e2d64 fix bug in setting new timerange for retraining 2022-05-28 12:23:26 +02:00
robcaulk
7870a86e9a fix live retraining bug 2022-05-28 11:38:57 +02:00
robcaulk
0bf915054d handle key check correctly 2022-05-28 11:22:32 +02:00
robcaulk
c5a16e91fb throw user error if user tries to load models but feeds the wrong features (while using PCA) 2022-05-28 11:11:41 +02:00
robcaulk
b8f9c3557b dirty dirty, dont look here (hacking a flag to avoid reloading leverage_tiers in dry/live) 2022-05-27 13:56:34 +02:00
robcaulk
891fb87712 give load_cached_data_for_updating the right flags to avoid redownloading data in dry/live 2022-05-27 13:38:22 +02:00
robcaulk
65fdebab75 let load_pairs_histories load futures candles in live 2022-05-27 13:01:33 +02:00
robcaulk
c080571b7a help futures go dry/live with auto download feature 2022-05-27 12:23:32 +02:00
robcaulk
8a501831d6 fix the error logic on previous commit 2022-05-27 01:15:55 +02:00
robcaulk
23c30dbc10 add error for user trying to backtest with backtest_period<1 2022-05-27 00:43:52 +02:00
robcaulk
6193205012 fix bug for target_mean/std array merging in backtesting 2022-05-26 21:07:50 +02:00
robcaulk
b79d4e8876 Allow user to go live and start from pretrained models (after a completed backtest) by simply reusing the identifier config parameter while dry/live. 2022-05-25 14:40:32 +02:00
robcaulk
7486d9d9e2 proper validation of freqai config parameters 2022-05-25 12:37:25 +02:00
robcaulk
7ff3258607 remove assertions, log error if user has not assigned freqai in config, fix stratify bug 2022-05-25 11:43:45 +02:00
robcaulk
35bed842cb cleanup, add clarity to comments and docstrings 2022-05-25 11:31:03 +02:00
robcaulk
58b5abbaa6 improve multithreaded training queue system 2022-05-24 15:28:38 +02:00
robcaulk
31ae2b3060 alleviate FutureWarning in sklearn about ensuring svm model features are passed with identical order 2022-05-24 14:46:16 +02:00
robcaulk
255d35976e add priority metadata to pairs to avoid a sync of train time + train period 2022-05-24 12:58:53 +02:00
robcaulk
059c285425 paying closer attention to managing live retraining on separate thread without affecting prediction of other coins on master thread 2022-05-24 12:01:01 +02:00
robcaulk
b0d2d13eb1 improve data persistence/mapping for live/dry. This accommodates quick reloads after crash and handles multi-pair cleanly 2022-05-23 21:05:05 +02:00
robcaulk
e1c068ca66 add config asserts, use .get method with default values for optional functionality, move data_cleaning_* to freqai_interface (away from user custom pred model) since it is controlled by config params. 2022-05-23 12:07:09 +02:00
robcaulk
ee3cdd0ffe more cleanup 2022-05-23 09:55:58 +02:00
robcaulk
3587bd82e1 cleanup superceded code 2022-05-23 00:10:36 +02:00
robcaulk
af0cc21af9 Enable hourly/minute retraining in live/dry. Suppress catboost folder output. Update config + constants + docs to reflect updates. 2022-05-23 00:06:26 +02:00
robcaulk
42d95af829 Aggregated commit. Adding support vector machine for outlier detection, improve user interface to dry/live, better standardization, fix various other bugs 2022-05-22 17:51:49 +02:00
robcaulk
c5ecf94177 move live retraining to separate thread. 2022-05-19 21:15:58 +02:00
robcaulk
1fae6c9ef7 keep model accessible in memory to avoid loading objects from disk during live/dry 2022-05-19 19:27:38 +02:00
robcaulk
89eacf2f47 Retrain model if FreqAI found a pretrained model but user strategy is not passing the expected features (user has changed the features in the strategy but has passed a the same config[freqai][identifier]). Logger warning output to user. 2022-05-19 17:15:50 +02:00
Matthias
c81b960791 Fix some typos 2022-05-17 19:58:36 +02:00
robcaulk
db66b82f6f accept open-ended timeranges from user 2022-05-17 19:50:06 +02:00
robcaulk
d1d451c27e auto populate features based on a prepended % in the strategy (remove feature assignment from config). Update doc/constants/example strategy to reflect change 2022-05-17 18:15:03 +02:00
robcaulk
8664e8f9a3 create a prediction_models folder where basic prediction models can live (similar to optimize/hyperopt-loss. Update resolver/docs/and gitignore to accommodate change 2022-05-17 17:13:38 +02:00
robcaulk
80dcd88abf allow user to run config from anywhere on their system 2022-05-15 17:42:15 +02:00
robcaulk
a8022c104a give beta testers more information in the doc 2022-05-15 17:42:15 +02:00
robcaulk
9b3e5faebe create more flexible whitelist, avoid duplicating whitelist features into corr_pairlist, update docs 2022-05-15 17:42:15 +02:00
robcaulk
22bd5556ed add self-retraining functionality for live/dry 2022-05-15 17:42:15 +02:00
robcaulk
178c2014b0 appease mypy 2022-05-15 17:42:15 +02:00
robcaulk
f653ace24b another attempt at fixing datalength bug 2022-05-15 17:42:15 +02:00
robcaulk
b08c0888bb add USERPATH_FREQAIMODELS, remove return values from @abstract methods 2022-05-15 17:42:15 +02:00
robcaulk
b03c7b514d optional style for interfacing freqai with backtesting 2022-05-15 17:42:15 +02:00
robcaulk
3020218096 fix bug on backtest timerange 2022-05-15 17:41:34 +02:00
robcaulk
00ff0c9b91 ensure user defined timerange truncates final backtest so that we arent mismatching data lengths upon return to strategy. Rename DataHandler class to FreqaiDataKitchen 2022-05-15 17:41:34 +02:00
robcaulk
def71a0afe auto build full_timerange and self manage training_timerange 2022-05-15 17:41:34 +02:00
robcaulk
764f9449b4 fix logger, debug some flake8 appeasements 2022-05-15 17:41:34 +02:00
robcaulk
29c2d1d189 use logger in favor of print 2022-05-15 17:38:58 +02:00
robcaulk
99f7e44c30 flake8 passing, use pathlib in lieu of os.path to accommodate windows/mac OS 2022-05-15 17:38:58 +02:00
robcaulk
630d201546 remove trained_stake 2022-05-15 17:38:58 +02:00
robcaulk
b40f8f88ac cleaning and bug fixing 2022-05-15 17:38:58 +02:00
robcaulk
fc837c4daa add freqao backend machinery, user interface, documentation 2022-05-15 17:38:58 +02:00