virgo-modules 0.3.2__tar.gz → 0.3.5__tar.gz

{virgo_modules-0.3.2 → virgo_modules-0.3.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: virgo_modules
-Version: 0.3.2
+Version: 0.3.5
 Summary: data processing and statistical modeling using stock market data
 Home-page: https://github.com/miguelmayhem92/virgo_module
 Author: Miguel Mayhuire

{virgo_modules-0.3.2 → virgo_modules-0.3.5}/setup.py

@@ -5,7 +5,7 @@ with open("virgo_app/README.md", "r") as f:
 
 setup(
     name="virgo_modules",
-    version="0.3.2",
+    version="0.3.5",
     description="data processing and statistical modeling using stock market data",
     package_dir={"": "virgo_app"},
     packages=find_packages(where="virgo_app"),

{virgo_modules-0.3.2 → virgo_modules-0.3.5}/virgo_app/virgo_modules/src/ticketer_source.py

@@ -1400,6 +1400,39 @@ class stock_eda_panel(object):
         if save_features:
             self.signals.append(signal_feature_name)
             self.signals.append(order_feature_name)
+    
+    def get_order_feature_nosignal(self,feature_name, save_features=False):
+        """
+        perform a feature that captures number of steps after the end of a signal
+
+        Parameters
+        ----------
+        feature_name (str): name of the feature
+        save_features (boolean): True to save feature configuration and feature names
+        
+        Returns
+        -------
+        None
+        """
+        order_feature_name = f'order_signal_{feature_name}'
+        ns_order_feature_name = f'ns_order_{feature_name}'
+        self.df = self.df.sort_values('Date')
+        self.df['lag_'] = self.df[order_feature_name].shift(1)
+        self.df['flag'] = np.where((self.df[order_feature_name] == 0) & (self.df['lag_']!=0),1,np.nan)
+        self.df = self.df.drop(columns=['lag_'])
+        self.df['order_'] = self.df.sort_values('Date').groupby(['flag']).cumcount() + 1
+        self.df['order_'] = self.df['order_'].fillna(method='ffill')
+        self.df['order_'] = np.where(self.df[order_feature_name]==0,self.df['order_'],0)
+        self.df = self.df.drop(columns=['flag'])
+        self.df['order_'] = self.df.sort_values('Date').groupby(['order_']).cumcount() + 1
+        norm_list = [f'norm_{feature_name}', f'z_{feature_name}', feature_name]
+        for norm_feature in norm_list:
+            self.df['order_'] = np.sign(self.df[norm_feature])*self.df['order_']
+            break
+        self.df['order_'] = np.where(self.df[order_feature_name]==0,self.df['order_'],0)
+        self.df = self.df.rename(columns={'order_':ns_order_feature_name})
+        if save_features:
+            self.signals.append(ns_order_feature_name)
 
     def compute_last_signal(self,feature, save_features = False):
         """
@@ -1815,7 +1848,7 @@ class stock_eda_panel(object):
         self.target.append(f'mean_target')
         self.settings_target_lasts = {'steps':steps, 'type':'regression'}
 
-    def get_categorical_targets(self, horizon, flor_loss, top_gain):
+    def get_categorical_targets(self, horizon, flor_loss, top_gain, min_pos=1 , min_negs=1):
         """
         produce binary target return taking future prices. it produce two targets, one for high returns and another for low returns
 
@@ -1824,6 +1857,8 @@ class stock_eda_panel(object):
         horizon (int): number of lags and steps for future returns
         flor_loss (float): min loss return
         top_gain (float): max gain return
+        min_pos (int): minimun number of positives to count in a window for target_up
+        min_negs (int): minimun number of negatives to count in a window for target_down
 
         Returns
         -------
@@ -1841,7 +1876,7 @@ class stock_eda_panel(object):
             self.df[f'target_{i}'] = np.where(self.df[f'target_{i}'] >= top_gain,1,0)
             columns.append(f'target_{i}')
         self.df[f'target_up'] = self.df[columns].sum(axis=1)
-        self.df[f'target_up'] = np.where(self.df[f'target_up'] >=1,1,0 )
+        self.df[f'target_up'] = np.where(self.df[f'target_up'] >=min_pos,1,0 )
         self.df = self.df.drop(columns = columns)
 
         for i in range(1,horizon+1):
@@ -1851,7 +1886,7 @@ class stock_eda_panel(object):
             self.df[f'target_{i}'] = np.where(self.df[f'target_{i}'] <= flor_loss,1,0)
             columns.append(f'target_{i}')
         self.df[f'target_down'] = self.df[columns].sum(axis=1)
-        self.df[f'target_down'] = np.where(self.df[f'target_down'] >= 1,1,0 )
+        self.df[f'target_down'] = np.where(self.df[f'target_down'] >= min_negs,1,0 )
         self.df = self.df.drop(columns = columns)
 
         self.targets.append('target_up')

{virgo_modules-0.3.2 → virgo_modules-0.3.5}/virgo_app/virgo_modules.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: virgo-modules
-Version: 0.3.2
+Version: 0.3.5
 Summary: data processing and statistical modeling using stock market data
 Home-page: https://github.com/miguelmayhem92/virgo_module
 Author: Miguel Mayhuire