feat(HPO): added run_hpo script (#237)

* feat(HPO): added `run_hpo` script

* fix(Linter): ran

* feat(HPO): removed any reference to sweep (superseeded by optuna)

* fix(HPO): optimize for sharpe

* fix(Config): removed glassnode data, save trials from hpo

* feat(Labelling): added three-balanced method works again

* fix(BetSizing): set the correct class labels

* fix(HPO): powerset should return what's expected, added two new normalization methods

* fix(Linter): ran

* fix(DataLoader): sort the dataframe when fetching data

* fix(Config): only take z-score of other assets

config/preprocess.py

@@ -1,4 +1,3 @@
-from sklearn.model_selection import TimeSeriesSplit
 from .types import Config, RawConfig
 from utils.helpers import flatten
 from feature_extractors.feature_extractor_presets import (
@@ -114,9 +113,12 @@ def __preprocess_data_collections_config(data_dict: dict) -> dict:
     return data_dict
 
 
-def __preprocess_event_filter_config(data_dict: dict) -> dict:
-    data_dict["event_filter"] = eventfilters_map[data_dict["event_filter"]]
-    return data_dict
+def __preprocess_event_filter_config(config_dict: dict) -> dict:
+    config_dict["event_filter"] = eventfilters_map[config_dict["event_filter"]](
+        config_dict["event_filter_multiplier"]
+    )
+    config_dict.pop("event_filter_multiplier")
+    return config_dict
 
 
 def __preprocess_event_labeller_config(config_dict: dict) -> dict:

config/presets.py

@@ -1,4 +1,4 @@
-from .types import RawConfig, Config
+from .types import RawConfig
 
 
 def get_default_config() -> RawConfig:
@@ -22,14 +22,15 @@ def get_default_config() -> RawConfig:
         assets=["fivemin_crypto"],
         target_asset="BTCUSDT",
         other_assets=[],
-        exogenous_data=["daily_glassnode"],
+        exogenous_data=[],
         load_non_target_asset=True,
         own_features=["level_2"],
         other_features=["z_score"],
-        exogenous_features=["z_score"],
+        exogenous_features=[],
         directional_models=classification_models,
         meta_models=meta_models,
         event_filter="cusum_vol",
+        event_filter_multiplier=3.5,
         remove_overlapping_events=False,
         labeling="two_class",
         forecasting_horizon=10,

config/sweep_ensemble.yaml

@@ -1,44 +0,0 @@
-program: run_sweep.py
-method: grid
-project: price-forecasting
-name: Meta labelling
-metric:
-  goal: maximize
-  name: sharpe
-parameters:
-    assets:
-      value: ['daily_crypto']
-    other_assets:
-      value: ['daily_etf']
-    exogenous_data:
-      value: ['daily_glassnode']
-    initial_window_size: 
-      value: 380
-      distribution: categorical
-    n_features_to_select:
-      values: [40, 50, 60]
-      distribution: categorical
-    dimensionality_reduction_ratio:
-      value: 0.5
-    retrain_every: 
-      value: 20
-    scaler:
-      value: 'minmax'
-    no_of_classes:
-      value: 'two'
-    load_non_target_asset: 
-      value: True
-    directional_models: 
-      distribution: categorical
-      values:
-      - ["LDA", "LogisticRegression_two_class", "KNN", "SVC", "CART", "NB", "AB", "RFC", "XGB_two_class", "LGBM", "StaticMom"]
-      - ["LogisticRegression_two_class", "LDA", "NB", "RFC", "XGB_two_class", "LGBM", "StaticMom"]
-      - ["LogisticRegression_two_class", "LDA", "LGBM", "RFC", "XGB_two_class"]
-    meta_models: 
-      value: ["LGBM", "LogisticRegression_two_class"]
-    own_features:
-      value: ['date_days', 'level_2', 'lags_up_to_5']
-    other_features:
-      value: ['level_2', 'lags_up_to_5']
-    exogenous_features:
-      value: ['z_score']

config/types.py

@@ -26,6 +26,7 @@ class RawConfig(BaseModel):
     other_features: list[str]
     exogenous_features: list[str]
     event_filter: Literal["none", "cusum_vol", "cusum_fixed"]
+    event_filter_multiplier: float
     remove_overlapping_events: bool
     labeling: Literal["two_class", "three_class_balanced", "three_class_imbalanced"]
     forecasting_horizon: int