added sinle model training api

alphamind/strategy/strategy.py

@@ -74,6 +74,7 @@ class Strategy(object):
         self.total_data = None
         self.index_return = None
         self.risk_models = None
+        self.alpha_models = None
 
     def prepare_backtest_data(self):
         total_factors = self.engine.fetch_factor_range(self.universe,
@@ -121,6 +122,25 @@ class Strategy(object):
                                                                     offset=1).set_index('trade_date')
         self.total_data = total_data
 
+    def prepare_backtest_models(self):
+        if self.total_data is None:
+            self.prepare_backtest_data()
+        total_data_groups = self.total_data.groupby('trade_date')
+        if self.dask_client is None:
+            models = {}
+            for ref_date, _ in total_data_groups:
+                models[ref_date] = train_model(ref_date.strftime('%Y-%m-%d'), self.alpha_model, self.data_meta)
+        else:
+            def worker(parameters):
+                new_model = train_model(parameters[0].strftime('%Y-%m-%d'), parameters[1], parameters[2])
+                return parameters[0], new_model
+
+            l = self.dask_client.map(worker, [(d[0], self.alpha_model, self.data_meta) for d in total_data_groups])
+            results = self.dask_client.gather(l)
+            models = dict(results)
+        self.alpha_models = models
+        alpha_logger.info("alpha models training finished ...")
+
     @staticmethod
     def _create_lu_bounds(running_setting, codes, benchmark_w):
 
@@ -169,22 +189,12 @@ class Strategy(object):
         executor = copy.deepcopy(running_setting.executor)
         positions = pd.DataFrame()
 
-        if self.dask_client is None:
-            models = {}
-            for ref_date, _ in total_data_groups:
-                models[ref_date] = train_model(ref_date.strftime('%Y-%m-%d'), self.alpha_model, self.data_meta)
-        else:
-            def worker(parameters):
-                new_model = train_model(parameters[0].strftime('%Y-%m-%d'), parameters[1], parameters[2])
-                return parameters[0], new_model
-
-            l = self.dask_client.map(worker, [(d[0], self.alpha_model, self.data_meta) for d in total_data_groups])
-            results = self.dask_client.gather(l)
-            models = dict(results)
+        if self.alpha_models is None:
+            self.prepare_backtest_models()
 
         for ref_date, this_data in total_data_groups:
             risk_model = self.risk_models[ref_date]
-            new_model = models[ref_date]
+            new_model = self.alpha_models[ref_date]
             codes = this_data.code.values.tolist()
 
             if previous_pos.empty:
@@ -249,7 +259,6 @@ class Strategy(object):
         ret_df = ret_df.shift(1)
         ret_df.iloc[0] = 0.
         ret_df['excess_return'] = ret_df['returns'] - ret_df['benchmark_returns'] * ret_df['leverage']
-
         return ret_df, positions
 
     def _calculate_pos(self, running_setting, er, data, constraints, benchmark_w, lbound, ubound, risk_model,