diff --git a/config_examples/config_freqai.example.json b/config_examples/config_freqai.example.json index 330bcd9f5..13c7a94ea 100644 --- a/config_examples/config_freqai.example.json +++ b/config_examples/config_freqai.example.json @@ -53,7 +53,6 @@ ], "freqai": { "enabled": true, - "startup_candles": 10000, "purge_old_models": true, "train_period_days": 15, "backtest_period_days": 7, @@ -76,7 +75,10 @@ "principal_component_analysis": false, "use_SVM_to_remove_outliers": true, "indicator_max_period_candles": 20, - "indicator_periods_candles": [10, 20] + "indicator_periods_candles": [ + 10, + 20 + ] }, "data_split_parameters": { "test_size": 0.33, @@ -92,4 +94,4 @@ "internals": { "process_throttle_secs": 5 } -} +} \ No newline at end of file diff --git a/docs/freqai.md b/docs/freqai.md index e4a451324..482a56d2b 100644 --- a/docs/freqai.md +++ b/docs/freqai.md @@ -38,14 +38,14 @@ The example strategy, example prediction model, and example config can be found The user provides FreqAI with a set of custom *base* indicators (the same way as in a typical Freqtrade strategy) as well as target values (*labels*). FreqAI trains a model to predict the target values based on the input of custom indicators, for each pair in the whitelist. These models are consistently retrained to adapt to market conditions. FreqAI offers the ability to both backtest strategies (emulating reality with periodic retraining) and deploy dry/live runs. In dry/live conditions, FreqAI can be set to constant retraining in a background thread in an effort to keep models as up to date as possible. -An overview of the algorithm is shown below, explaining the data processing pipeline and the model usage. +An overview of the algorithm is shown below, explaining the data processing pipeline and the model usage. ![freqai-algo](assets/freqai_algo.jpg) ### Important machine learning vocabulary **Features** - the quantities with which a model is trained. All features for a single candle is stored as a vector. In FreqAI, the user -builds the feature sets from anything they can construct in the strategy. +builds the feature sets from anything they can construct in the strategy. **Labels** - the target values that a model is trained toward. Each set of features is associated with a single label that is @@ -53,12 +53,12 @@ defined by the user within the strategy. These labels intentionally look into th future, and are not available to the model during dry/live/backtesting. **Training** - the process of feeding individual feature sets, composed of historic data, with associated labels into the -model with the goal of matching input feature sets to associated labels. +model with the goal of matching input feature sets to associated labels. **Train data** - a subset of the historic data that is fed to the model during training. This data directly influences weight connections in the model. -**Test data** - a subset of the historic data that is used to evaluate the performance of the model after training. This data does not influence nodal weights within the model. +**Test data** - a subset of the historic data that is used to evaluate the performance of the model after training. This data does not influence nodal weights within the model. ## Install prerequisites @@ -104,21 +104,21 @@ Mandatory parameters are marked as **Required**, which means that they are requi | `include_corr_pairlist` | A list of correlated coins that FreqAI will add as additional features to all `pair_whitelist` coins. All indicators set in `populate_any_indicators` during feature engineering (see details [here](#feature-engineering)) will be created for each coin in this list, and that set of features is added to the base asset feature set.
**Datatype:** List of assets (strings). | `label_period_candles` | Number of candles into the future that the labels are created for. This is used in `populate_any_indicators` (see `templates/FreqaiExampleStrategy.py` for detailed usage). The user can create custom labels, making use of this parameter or not.
**Datatype:** Positive integer. | `include_shifted_candles` | Add features from previous candles to subsequent candles to add historical information. FreqAI takes all features from the `include_shifted_candles` previous candles, duplicates and shifts them so that the information is available for the subsequent candle.
**Datatype:** Positive integer. -| `weight_factor` | Used to set weights for training data points according to their recency. See details about how it works [here](#controlling-the-model-learning-process).
**Datatype:** Positive float (typically < 1). -| `indicator_max_period_candles` | The maximum period used in `populate_any_indicators()` for indicator creation. FreqAI uses this information in combination with the maximum timeframe to calculate how many data points that should be downloaded so that the first data point does not have a NaN.
**Datatype:** Positive integer. +| `weight_factor` | Used to set weights for training data points according to their recency. See details about how it works [here](#controlling-the-model-learning-process).
**Datatype:** Positive float (typically < 1). +| `indicator_max_period_candles` | **No longer used**. User must use the strategy set `startup_candle_count` which defines the maximum *period* used in `populate_any_indicators()` for indicator creation (timeframe independent). FreqAI uses this information in combination with the maximum timeframe to calculate how many data points it should download so that the first data point does not have a NaN
**Datatype:** positive integer. | `indicator_periods_candles` | Calculate indicators for `indicator_periods_candles` time periods and add them to the feature set.
**Datatype:** List of positive integers. | `stratify_training_data` | This value is used to indicate the grouping of the data. For example, 2 would set every 2nd data point into a separate dataset to be pulled from during training/testing. See details about how it works [here](#stratifying-the-data-for-training-and-testing-the-model)
**Datatype:** Positive integer. -| `principal_component_analysis` | Automatically reduce the dimensionality of the data set using Principal Component Analysis. See details about how it works [here](#reducing-data-dimensionality-with-principal-component-analysis)
**Datatype:** Boolean. +| `principal_component_analysis` | Automatically reduce the dimensionality of the data set using Principal Component Analysis. See details about how it works [here](#reducing-data-dimensionality-with-principal-component-analysis)
**Datatype:** Boolean. | `DI_threshold` | Activates the Dissimilarity Index for outlier detection when > 0. See details about how it works [here](#removing-outliers-with-the-dissimilarity-index).
**Datatype:** Positive float (typically < 1). | `use_SVM_to_remove_outliers` | Train a support vector machine to detect and remove outliers from the training data set, as well as from incoming data points. See details about how it works [here](#removing-outliers-using-a-support-vector-machine-svm).
**Datatype:** Boolean. | `svm_params` | All parameters available in Sklearn's `SGDOneClassSVM()`. See details about some select parameters [here](#removing-outliers-using-a-support-vector-machine-svm).
**Datatype:** Dictionary. -| `use_DBSCAN_to_remove_outliers` | Cluster data using DBSCAN to identify and remove outliers from training and prediction data. See details about how it works [here](#removing-outliers-with-dbscan).
**Datatype:** Boolean. +| `use_DBSCAN_to_remove_outliers` | Cluster data using DBSCAN to identify and remove outliers from training and prediction data. See details about how it works [here](#removing-outliers-with-dbscan).
**Datatype:** Boolean. | `outlier_protection_percentage` | If more than `outlier_protection_percentage` fraction of points are removed as outliers, FreqAI will log a warning message and ignore outlier detection while keeping the original dataset intact.
**Datatype:** float. Default: `30` | `reverse_train_test_order` | If true, FreqAI will train on the latest data split and test on historical split of the data. This allows the model to be trained up to the most recent data point, while avoiding overfitting. However, users should be careful to understand unorthodox nature of this parameter before employing it.
**Datatype:** bool. Default: False | | **Data split parameters** | `data_split_parameters` | Include any additional parameters available from Scikit-learn `test_train_split()`, which are shown [here](https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.train_test_split.html) (external website).
**Datatype:** Dictionary. | `test_size` | Fraction of data that should be used for testing instead of training.
**Datatype:** Positive float < 1. -| `shuffle` | Shuffle the training data points during training. Typically, for time-series forecasting, this is set to `False`.
+| `shuffle` | Shuffle the training data points during training. Typically, for time-series forecasting, this is set to `False`.
| | **Model training parameters** | `model_training_parameters` | A flexible dictionary that includes all parameters available by the user selected model library. For example, if the user uses `LightGBMRegressor`, this dictionary can contain any parameter available by the `LightGBMRegressor` [here](https://lightgbm.readthedocs.io/en/latest/pythonapi/lightgbm.LGBMRegressor.html) (external website). If the user selects a different model, this dictionary can contain any parameter from that model.
**Datatype:** Dictionary.**Datatype:** Boolean. | `n_estimators` | The number of boosted trees to fit in regression.
**Datatype:** Integer. @@ -167,7 +167,6 @@ The user interface is isolated to the typical Freqtrade config file. A FreqAI co ], "label_period_candles": 24, "include_shifted_candles": 2, - "indicator_max_period_candles": 20, "indicator_periods_candles": [10, 20] }, "data_split_parameters" : { @@ -184,6 +183,9 @@ The user interface is isolated to the typical Freqtrade config file. A FreqAI co The FreqAI strategy requires the user to include the following lines of code in the standard Freqtrade strategy: ```python + # user should define the maximum startup candle count (the largest number of candles + # passed to any single indicator) + startup_candle_count: int = 20 def informative_pairs(self): whitelist_pairs = self.dp.current_whitelist() @@ -200,9 +202,9 @@ The FreqAI strategy requires the user to include the following lines of code in def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame: - # the model will return all labels created by user in `populate_any_indicators` - # (& appended targets), an indication of whether or not the prediction should be accepted, - # the target mean/std values for each of the labels created by user in + # the model will return all labels created by user in `populate_any_indicators` + # (& appended targets), an indication of whether or not the prediction should be accepted, + # the target mean/std values for each of the labels created by user in # `populate_any_indicators()` for each training period. dataframe = self.freqai.start(dataframe, metadata, self) @@ -310,7 +312,7 @@ The user is encouraged to inherit `train()` and `predict()` to let them customiz ## Feature engineering Features are added by the user inside the `populate_any_indicators()` method of the strategy -by prepending indicators with `%`, and labels with `&`. +by prepending indicators with `%`, and labels with `&`. There are some important components/structures that the user *must* include when building their feature set; the use of these is shown below: @@ -419,13 +421,13 @@ In total, the number of features the user of the presented example strat has cre length of `include_timeframes` * no. features in `populate_any_indicators()` * length of `include_corr_pairlist` * no. `include_shifted_candles` * length of `indicator_periods_candles` $= 3 * 3 * 3 * 2 * 2 = 108$. -Another structure to consider is the location of the labels at the bottom of the example function (below `if set_generalized_indicators:`). +Another structure to consider is the location of the labels at the bottom of the example function (below `if set_generalized_indicators:`). This is where the user will add single features and labels to their feature set to avoid duplication of them from various configuration parameters that multiply the feature set, such as `include_timeframes`. !!! Note Features **must** be defined in `populate_any_indicators()`. Definining FreqAI features in `populate_indicators()` - will cause the algorithm to fail in live/dry mode. If the user wishes to add generalized features that are not associated with + will cause the algorithm to fail in live/dry mode. If the user wishes to add generalized features that are not associated with a specific pair or timeframe, they should use the following structure inside `populate_any_indicators()` (as exemplified in `freqtrade/templates/FreqaiExampleStrategy.py`): @@ -434,7 +436,7 @@ various configuration parameters that multiply the feature set, such as `include ... - # Add generalized indicators here (because in live, it will call only this function to populate + # Add generalized indicators here (because in live, it will call only this function to populate # indicators for retraining). Notice how we ensure not to add them multiple times by associating # these generalized indicators to the basepair/timeframe if set_generalized_indicators: @@ -545,7 +547,7 @@ and adding this to the `train_period_days`. The units need to be in the base can ### Defining model expirations -During dry/live mode, FreqAI trains each coin pair sequentially (on separate threads/GPU from the main Freqtrade bot). This means that there is always an age discrepancy between models. If a user is training on 50 pairs, and each pair requires 5 minutes to train, the oldest model will be over 4 hours old. This may be undesirable if the characteristic time scale (the trade duration target) for a strategy is less than 4 hours. The user can decide to only make trade entries if the model is less than +During dry/live mode, FreqAI trains each coin pair sequentially (on separate threads/GPU from the main Freqtrade bot). This means that there is always an age discrepancy between models. If a user is training on 50 pairs, and each pair requires 5 minutes to train, the oldest model will be over 4 hours old. This may be undesirable if the characteristic time scale (the trade duration target) for a strategy is less than 4 hours. The user can decide to only make trade entries if the model is less than a certain number of hours old by setting the `expiration_hours` in the config file: ```json @@ -632,15 +634,15 @@ The user can stratify (group) the training/testing data using: This will split the data chronologically so that every Xth data point is used to test the model after training. In the example above, the user is asking for every third data point in the dataframe to be used for -testing; the other points are used for training. +testing; the other points are used for training. -The test data is used to evaluate the performance of the model after training. If the test score is high, the model is able to capture the behavior of the data well. If the test score is low, either the model either does not capture the complexity of the data, the test data is significantly different from the train data, or a different model should be used. +The test data is used to evaluate the performance of the model after training. If the test score is high, the model is able to capture the behavior of the data well. If the test score is low, either the model either does not capture the complexity of the data, the test data is significantly different from the train data, or a different model should be used. ### Controlling the model learning process Model training parameters are unique to the machine learning library selected by the user. FreqAI allows the user to set any parameter for any library using the `model_training_parameters` dictionary in the user configuration file. The example configuration file (found in `config_examples/config_freqai.example.json`) show some of the example parameters associated with `Catboost` and `LightGBM`, but the user can add any parameters available in those libraries. -Data split parameters are defined in `data_split_parameters` which can be any parameters associated with `Sklearn`'s `train_test_split()` function. +Data split parameters are defined in `data_split_parameters` which can be any parameters associated with `Sklearn`'s `train_test_split()` function. FreqAI includes some additional parameters such as `weight_factor`, which allows the user to weight more recent data more strongly than past data via an exponential function: @@ -670,7 +672,7 @@ The user can tell FreqAI to remove outlier data points from the training/test da } ``` -Equity and crypto markets suffer from a high level of non-patterned noise in the form of outlier data points. The Dissimilarity Index (DI) aims to quantify the uncertainty associated with each prediction made by the model. The DI allows predictions which are outliers (not existent in the model feature space) to be thrown out due to low levels of certainty. +Equity and crypto markets suffer from a high level of non-patterned noise in the form of outlier data points. The Dissimilarity Index (DI) aims to quantify the uncertainty associated with each prediction made by the model. The DI allows predictions which are outliers (not existent in the model feature space) to be thrown out due to low levels of certainty. To do so, FreqAI measures the distance between each training data point (feature vector), $X_{a}$, and all other training data points: @@ -688,7 +690,7 @@ which enables the estimation of the Dissimilarity Index as: $$ DI_k = d_k/\overline{d} $$ -The user can tweak the DI through the `DI_threshold` to increase or decrease the extrapolation of the trained model. +The user can tweak the DI through the `DI_threshold` to increase or decrease the extrapolation of the trained model. Below is a figure that describes the DI for a 3D data set. @@ -707,11 +709,11 @@ The user can tell FreqAI to remove outlier data points from the training/test da ``` FreqAI will train an SVM on the training data (or components of it if the user activated -`principal_component_analysis`) and remove any data point that the SVM deems to be beyond the feature space. +`principal_component_analysis`) and remove any data point that the SVM deems to be beyond the feature space. -The parameter `shuffle` is by default set to `False` to ensure consistent results. If it is set to `True`, running the SVM multiple times on the same data set might result in different outcomes due to `max_iter` being to low for the algorithm to reach the demanded `tol`. Increasing `max_iter` solves this issue but causes the procedure to take longer time. +The parameter `shuffle` is by default set to `False` to ensure consistent results. If it is set to `True`, running the SVM multiple times on the same data set might result in different outcomes due to `max_iter` being to low for the algorithm to reach the demanded `tol`. Increasing `max_iter` solves this issue but causes the procedure to take longer time. -The parameter `nu`, *very* broadly, is the amount of data points that should be considered outliers. +The parameter `nu`, *very* broadly, is the amount of data points that should be considered outliers. #### Removing outliers with DBSCAN diff --git a/docs/strategy-customization.md b/docs/strategy-customization.md index 260e253c4..a452b8f05 100644 --- a/docs/strategy-customization.md +++ b/docs/strategy-customization.md @@ -166,7 +166,7 @@ Additional technical libraries can be installed as necessary, or custom indicato Most indicators have an instable startup period, in which they are either not available (NaN), or the calculation is incorrect. This can lead to inconsistencies, since Freqtrade does not know how long this instable period should be. To account for this, the strategy can be assigned the `startup_candle_count` attribute. -This should be set to the maximum number of candles that the strategy requires to calculate stable indicators. +This should be set to the maximum number of candles that the strategy requires to calculate stable indicators. In the case where a user includes higher timeframes with informative pairs, the `startup_candle_count` does not necessarily change. The value is the maximum period (in candles) that any of the informatives timeframes need to compute stable indicators. In this example strategy, this should be set to 100 (`startup_candle_count = 100`), since the longest needed history is 100 candles. diff --git a/freqtrade/data/dataprovider.py b/freqtrade/data/dataprovider.py index 21cead77f..4151b7419 100644 --- a/freqtrade/data/dataprovider.py +++ b/freqtrade/data/dataprovider.py @@ -91,9 +91,9 @@ class DataProvider: timerange = TimeRange.parse_timerange(None if self._config.get( 'timerange') is None else str(self._config.get('timerange'))) # Move informative start time respecting startup_candle_count - timerange.subtract_start( - timeframe_to_seconds(str(timeframe)) * self._config.get('startup_candle_count', 0) - ) + startup_candles = self.get_required_startup(str(timeframe)) + tf_seconds = timeframe_to_seconds(str(timeframe)) + timerange.subtract_start(tf_seconds * startup_candles) self.__cached_pairs_backtesting[saved_pair] = load_pair_history( pair=pair, timeframe=timeframe or self._config['timeframe'], @@ -105,6 +105,21 @@ class DataProvider: ) return self.__cached_pairs_backtesting[saved_pair].copy() + def get_required_startup(self, timeframe: str) -> int: + freqai_config = self._config.get('freqai', {}) + if not freqai_config.get('enabled', False): + return self._config.get('startup_candle_count', 0) + else: + startup_candles = self._config.get('startup_candle_count', 0) + indicator_periods = freqai_config['feature_parameters']['indicator_periods_candles'] + # make sure the startupcandles is at least the set maximum indicator periods + self._config['startup_candle_count'] = max(startup_candles, max(indicator_periods)) + tf_seconds = timeframe_to_seconds(timeframe) + train_candles = freqai_config['train_period_days'] * 86400 / tf_seconds + total_candles = int(self._config['startup_candle_count'] + train_candles) + logger.info(f'Increasing startup_candle_count for freqai to {total_candles}') + return total_candles + def get_pair_dataframe( self, pair: str, diff --git a/freqtrade/freqai/data_kitchen.py b/freqtrade/freqai/data_kitchen.py index 8dc94e9ec..361d9872d 100644 --- a/freqtrade/freqai/data_kitchen.py +++ b/freqtrade/freqai/data_kitchen.py @@ -16,8 +16,6 @@ from sklearn.model_selection import train_test_split from sklearn.neighbors import NearestNeighbors from freqtrade.configuration import TimeRange -from freqtrade.data.dataprovider import DataProvider -from freqtrade.data.history.history_utils import refresh_backtest_ohlcv_data from freqtrade.exceptions import OperationalException from freqtrade.exchange import timeframe_to_seconds from freqtrade.strategy.interface import IStrategy @@ -911,9 +909,7 @@ class FreqaiDataKitchen: # We notice that users like to use exotic indicators where # they do not know the required timeperiod. Here we include a factor # of safety by multiplying the user considered "max" by 2. - max_period = self.freqai_config["feature_parameters"].get( - "indicator_max_period_candles", 20 - ) * 2 + max_period = self.config.get('startup_candle_count', 20) * 2 additional_seconds = max_period * max_tf_seconds if trained_timestamp != 0: @@ -959,31 +955,6 @@ class FreqaiDataKitchen: self.model_filename = f"cb_{coin.lower()}_{int(trained_timerange.stopts)}" - def download_all_data_for_training(self, timerange: TimeRange, dp: DataProvider) -> None: - """ - Called only once upon start of bot to download the necessary data for - populating indicators and training the model. - :param timerange: TimeRange = The full data timerange for populating the indicators - and training the model. - :param dp: DataProvider instance attached to the strategy - """ - new_pairs_days = int((timerange.stopts - timerange.startts) / SECONDS_IN_DAY) - if not dp._exchange: - # Not realistic - this is only called in live mode. - raise OperationalException("Dataprovider did not have an exchange attached.") - refresh_backtest_ohlcv_data( - dp._exchange, - pairs=self.all_pairs, - timeframes=self.freqai_config["feature_parameters"].get("include_timeframes"), - datadir=self.config["datadir"], - timerange=timerange, - new_pairs_days=new_pairs_days, - erase=False, - data_format=self.config.get("dataformat_ohlcv", "json"), - trading_mode=self.config.get("trading_mode", "spot"), - prepend=self.config.get("prepend_data", False), - ) - def set_all_pairs(self) -> None: self.all_pairs = copy.deepcopy( diff --git a/freqtrade/freqai/freqai_interface.py b/freqtrade/freqai/freqai_interface.py index ad64588a7..2297811b4 100644 --- a/freqtrade/freqai/freqai_interface.py +++ b/freqtrade/freqai/freqai_interface.py @@ -312,14 +312,8 @@ class IFreqaiModel(ABC): ) dk.set_paths(metadata["pair"], new_trained_timerange.stopts) - # download candle history if it is not already in memory + # load candle history into memory if it is not yet. if not self.dd.historic_data: - logger.info( - "Downloading all training data for all pairs in whitelist and " - "corr_pairlist, this may take a while if you do not have the " - "data saved" - ) - dk.download_all_data_for_training(data_load_timerange, strategy.dp) self.dd.load_all_pair_histories(data_load_timerange, dk) if not self.scanning: diff --git a/freqtrade/freqai/utils.py b/freqtrade/freqai/utils.py new file mode 100644 index 000000000..6a70f050f --- /dev/null +++ b/freqtrade/freqai/utils.py @@ -0,0 +1,134 @@ +import logging +from datetime import datetime, timezone + +from freqtrade.configuration import TimeRange +from freqtrade.data.dataprovider import DataProvider +from freqtrade.data.history.history_utils import refresh_backtest_ohlcv_data +from freqtrade.exceptions import OperationalException +from freqtrade.exchange import timeframe_to_seconds +from freqtrade.exchange.exchange import market_is_active +from freqtrade.plugins.pairlist.pairlist_helpers import dynamic_expand_pairlist + + +logger = logging.getLogger(__name__) + + +def download_all_data_for_training(dp: DataProvider, config: dict) -> None: + """ + Called only once upon start of bot to download the necessary data for + populating indicators and training the model. + :param timerange: TimeRange = The full data timerange for populating the indicators + and training the model. + :param dp: DataProvider instance attached to the strategy + """ + + if dp._exchange is None: + raise OperationalException('No exchange object found.') + markets = [p for p, m in dp._exchange.markets.items() if market_is_active(m) + or config.get('include_inactive')] + + all_pairs = dynamic_expand_pairlist(config, markets) + + timerange = get_required_data_timerange(config) + + new_pairs_days = int((timerange.stopts - timerange.startts) / 86400) + + refresh_backtest_ohlcv_data( + dp._exchange, + pairs=all_pairs, + timeframes=config["freqai"]["feature_parameters"].get("include_timeframes"), + datadir=config["datadir"], + timerange=timerange, + new_pairs_days=new_pairs_days, + erase=False, + data_format=config.get("dataformat_ohlcv", "json"), + trading_mode=config.get("trading_mode", "spot"), + prepend=config.get("prepend_data", False), + ) + + +def get_required_data_timerange( + config: dict +) -> TimeRange: + """ + Used to compute the required data download time range + for auto data-download in FreqAI + """ + time = datetime.now(tz=timezone.utc).timestamp() + + timeframes = config["freqai"]["feature_parameters"].get("include_timeframes") + + max_tf_seconds = 0 + for tf in timeframes: + secs = timeframe_to_seconds(tf) + if secs > max_tf_seconds: + max_tf_seconds = secs + + startup_candles = config.get('startup_candle_count', 0) + indicator_periods = config["freqai"]["feature_parameters"]["indicator_periods_candles"] + + # factor the max_period as a factor of safety. + max_period = int(max(startup_candles, max(indicator_periods)) * 1.5) + config['startup_candle_count'] = max_period + logger.info(f'FreqAI auto-downloader using {max_period} startup candles.') + + additional_seconds = max_period * max_tf_seconds + + startts = int( + time + - config["freqai"].get("train_period_days", 0) * 86400 + - additional_seconds + ) + stopts = int(time) + data_load_timerange = TimeRange('date', 'date', startts, stopts) + + return data_load_timerange + + +# Keep below for when we wish to download heterogeneously lengthed data for FreqAI. +# def download_all_data_for_training(dp: DataProvider, config: dict) -> None: +# """ +# Called only once upon start of bot to download the necessary data for +# populating indicators and training a FreqAI model. +# :param timerange: TimeRange = The full data timerange for populating the indicators +# and training the model. +# :param dp: DataProvider instance attached to the strategy +# """ + +# if dp._exchange is not None: +# markets = [p for p, m in dp._exchange.markets.items() if market_is_active(m) +# or config.get('include_inactive')] +# else: +# # This should not occur: +# raise OperationalException('No exchange object found.') + +# all_pairs = dynamic_expand_pairlist(config, markets) + +# if not dp._exchange: +# # Not realistic - this is only called in live mode. +# raise OperationalException("Dataprovider did not have an exchange attached.") + +# time = datetime.now(tz=timezone.utc).timestamp() + +# for tf in config["freqai"]["feature_parameters"].get("include_timeframes"): +# timerange = TimeRange() +# timerange.startts = int(time) +# timerange.stopts = int(time) +# startup_candles = dp.get_required_startup(str(tf)) +# tf_seconds = timeframe_to_seconds(str(tf)) +# timerange.subtract_start(tf_seconds * startup_candles) +# new_pairs_days = int((timerange.stopts - timerange.startts) / 86400) +# # FIXME: now that we are looping on `refresh_backtest_ohlcv_data`, the function +# # redownloads the funding rate for each pair. +# refresh_backtest_ohlcv_data( +# dp._exchange, +# pairs=all_pairs, +# timeframes=[tf], +# datadir=config["datadir"], +# timerange=timerange, +# new_pairs_days=new_pairs_days, +# erase=False, +# data_format=config.get("dataformat_ohlcv", "json"), +# trading_mode=config.get("trading_mode", "spot"), +# prepend=config.get("prepend_data", False), +# ) diff --git a/freqtrade/optimize/backtesting.py b/freqtrade/optimize/backtesting.py index d65a75503..77bf3d8ad 100644 --- a/freqtrade/optimize/backtesting.py +++ b/freqtrade/optimize/backtesting.py @@ -212,21 +212,12 @@ class Backtesting: """ self.progress.init_step(BacktestState.DATALOAD, 1) - if self.config.get('freqai', {}).get('enabled', False): - startup_candles = int(self.config.get('freqai', {}).get('startup_candles', 0)) - if not startup_candles: - raise OperationalException('FreqAI backtesting module requires user set ' - 'startup_candles in config.') - self.required_startup += int(self.config.get('freqai', {}).get('startup_candles', 0)) - logger.info(f'Increasing startup_candle_count for freqai to {self.required_startup}') - self.config['startup_candle_count'] = self.required_startup - data = history.load_data( datadir=self.config['datadir'], pairs=self.pairlists.whitelist, timeframe=self.timeframe, timerange=self.timerange, - startup_candles=self.required_startup, + startup_candles=self.dataprovider.get_required_startup(self.timeframe), fail_without_data=True, data_format=self.config.get('dataformat_ohlcv', 'json'), candle_type=self.config.get('candle_type_def', CandleType.SPOT) diff --git a/freqtrade/strategy/interface.py b/freqtrade/strategy/interface.py index db16ac383..c7ea95bda 100644 --- a/freqtrade/strategy/interface.py +++ b/freqtrade/strategy/interface.py @@ -148,10 +148,19 @@ class IStrategy(ABC, HyperStrategyMixin): def load_freqAI_model(self) -> None: if self.config.get('freqai', {}).get('enabled', False): # Import here to avoid importing this if freqAI is disabled + from freqtrade.freqai.utils import download_all_data_for_training from freqtrade.resolvers.freqaimodel_resolver import FreqaiModelResolver - self.freqai = FreqaiModelResolver.load_freqaimodel(self.config) self.freqai_info = self.config["freqai"] + + # download the desired data in dry/live + if self.config.get('runmode') in (RunMode.DRY_RUN, RunMode.LIVE): + logger.info( + "Downloading all training data for all pairs in whitelist and " + "corr_pairlist, this may take a while if the data is not " + "already on disk." + ) + download_all_data_for_training(self.dp, self.config) else: # Gracious failures if freqAI is disabled but "start" is called. class DummyClass(): diff --git a/freqtrade/templates/FreqaiExampleStrategy.py b/freqtrade/templates/FreqaiExampleStrategy.py index 5810e7881..aa584bfbc 100644 --- a/freqtrade/templates/FreqaiExampleStrategy.py +++ b/freqtrade/templates/FreqaiExampleStrategy.py @@ -43,7 +43,8 @@ class FreqaiExampleStrategy(IStrategy): process_only_new_candles = True stoploss = -0.05 use_exit_signal = True - startup_candle_count: int = 300 + # this is the maximum period fed to talib (timeframe independent) + startup_candle_count: int = 20 can_short = False linear_roi_offset = DecimalParameter( diff --git a/freqtrade/templates/FreqaiHybridExampleStrategy.py b/freqtrade/templates/FreqaiHybridExampleStrategy.py index 0a91455f5..5d1e149dd 100644 --- a/freqtrade/templates/FreqaiHybridExampleStrategy.py +++ b/freqtrade/templates/FreqaiHybridExampleStrategy.py @@ -45,7 +45,6 @@ class FreqaiExampleHybridStrategy(IStrategy): "weight_factor": 0.9, "principal_component_analysis": false, "use_SVM_to_remove_outliers": true, - "indicator_max_period_candles": 20, "indicator_periods_candles": [10, 20] }, "data_split_parameters": { diff --git a/tests/freqai/conftest.py b/tests/freqai/conftest.py index dd148da77..db499631b 100644 --- a/tests/freqai/conftest.py +++ b/tests/freqai/conftest.py @@ -45,7 +45,6 @@ def freqai_conf(default_conf, tmpdir): "principal_component_analysis": False, "use_SVM_to_remove_outliers": True, "stratify_training_data": 0, - "indicator_max_period_candles": 10, "indicator_periods_candles": [10], }, "data_split_parameters": {"test_size": 0.33, "random_state": 1}, diff --git a/tests/freqai/test_freqai_backtesting.py b/tests/freqai/test_freqai_backtesting.py index 273791609..ea127fa99 100644 --- a/tests/freqai/test_freqai_backtesting.py +++ b/tests/freqai/test_freqai_backtesting.py @@ -48,10 +48,4 @@ def test_freqai_backtest_load_data(freqai_conf, mocker, caplog): assert log_has_re('Increasing startup_candle_count for freqai to.*', caplog) - del freqai_conf['freqai']['startup_candles'] - backtesting = Backtesting(freqai_conf) - with pytest.raises(OperationalException, - match=r'FreqAI backtesting module.*startup_candles in config.'): - backtesting.load_bt_data() - Backtesting.cleanup()