fixing various bugs in update

Author: Zsailer

epistasis/models/nonlinear.py

@@ -184,12 +184,14 @@ def fit(self, X=None, y=None, sample_weight=None, use_widgets=False, **kwargs):
         # Fit with an additive model
         self.Additive = EpistasisLinearRegression(order=1, model_type=self.model_type)
         self.Additive.add_gpm(self.gpm)
-        self.Additive.Xfit = X[:,:self.Additive.gpm.length+1]
+        self.Additive.Xfit = X[:,:self.Additive.gpm.binary.length+1]
+        self.Additive.Xpredict = self.Additive.Xfit
 
         # Prepare a high-order model
         self.Linear = EpistasisLinearRegression(order=self.order, model_type=self.model_type)
         self.Linear.add_gpm(self.gpm)
         self.Linear.Xfit = X
+        self.Linear.Xpredict = self.Linear.Xfit
 
         ## Use widgets to guess the value?
         if use_widgets:

epistasis/plot/correlation.py

@@ -27,7 +27,7 @@ def resid(y_pred, y_obs, fig=None, ax=None, figsize=None):
         ax = fig.add_subplot(111)
     residuals = y_obs - y_pred
     # Make residual plot
-    ax.plot(y_pred, residuals, '.')
+    ax.plot(residuals, '.')
     ax.hlines(0,min(y_pred), max(y_pred), linestyle="--", color="gray")
     ax.spines["top"].set_visible(False)
     ax.spines["right"].set_visible(False)

epistasis/simulate/linear.py

@@ -54,9 +54,9 @@ def __init__(self, wildtype, mutations,
     @property
     def p_additive(self):
         """Get the additive phenotypes"""
-        orders = self.epistasis.getorder
-        sites = list(orders[0].sites) + list(orders[1].sites)
-        vals = list(orders[0].values) + list(orders[1].values)
+        orders = self.epistasis.get_orders(0,1)
+        sites = orders.sites
+        vals = orders.values
         x = get_model_matrix(self.binary.genotypes, sites, model_type=self.model_type)
         return np.dot(x, vals)
 

epistasis/stats.py

@@ -40,7 +40,7 @@ def gmean(x):
     """Calculate a geometric mean with zero and negative values.
 
     Following the gmean calculation from this paper:
-    
+
     Habib, Elsayed AE. "Geometric mean for negative and zero values."
     International Journal of Research and Reviews in Applied Sciences 11 (2012): 419-432.
     """
@@ -195,10 +195,7 @@ def pearson(y_obs, y_pred):
     xbar = np.mean(y_obs)
     ybar = np.mean(y_pred)
 
-    terms = np.zeros(len(x), dtype=float)
-
-    for i in range(len(x)):
-        terms[i] = (x[i] - xbar) * (y[i] - ybar)
+    terms = (x - xbar) * (y - ybar)
 
     numerator = sum(terms)