Add initial support for multiverse generation

.github/workflows/main.yml

@@ -55,8 +55,7 @@ jobs:
       - name: Run tests
         run: |
           source .venv/bin/activate
-          pytest tests/
-          pytest tests/ --ignore-glob='test_graph_vis.py' .--cov=./ --cov-report=xml
+          pytest tests/ --ignore=tests/test_graph_vis.py --cov-report=xml
       - name: Upload coverage data to coveralls.io
         run: |
           python -m pip install coveralls==2.2

examples/Animal_Science/model_data_path.py

@@ -12,6 +12,7 @@ def fit_model():
 
 
     model = Lmer(formula='Weight ~ Time + (1|Litter) + (1|Pig) + (1|Time)', family="gaussian", data=df)
+
     print(model.fit())
     return model
 
@@ -21,6 +22,7 @@ def fit_model():
 # Read more here: https://sscc.wisc.edu/sscc/pubs/RegressionDiagnostics.html
 def show_model_diagnostics(model): 
 
+    plt.axhline(y=0, color='r', linestyle='-')
     plt.scatter(model.fits, model.residuals)
     plt.title("Fitted values vs. Residuals")
     plt.xlabel("fitted values")

examples/Animal_Science/model_df.py

@@ -14,7 +14,8 @@ def fit_model():
     df = pd.read_csv('/Users/emjun/Git/tisane/data.csv') # Make sure that the data path is correct
 
 
-    model = Lmer(formula='Weight ~ Time + (1|Litter) + (1|Pig) + (1|Time)', family="gaussian", data=df)
+    model = Lmer(formula='Weight ~ Time + (1|Pig) + (1|Time) + (1|Litter)', family="gaussian", data=df)
+
     print(model.fit())
     return model
 
@@ -24,6 +25,7 @@ def fit_model():
 # Read more here: https://sscc.wisc.edu/sscc/pubs/RegressionDiagnostics.html
 def show_model_diagnostics(model): 
 
+    plt.axhline(y=0, color='r', linestyle='-')
     plt.scatter(model.fits, model.residuals)
     plt.title("Fitted values vs. Residuals")
     plt.xlabel("fitted values")

examples/Animal_Science/model_no_data.py

@@ -15,7 +15,8 @@ def fit_model():
     # df = <your pandas Dataframe>
 
 
-    model = Lmer(formula='Weight ~ Time + (1|Pig) + (1|Litter) + (1|Time)', family="gaussian", data=df)
+    model = Lmer(formula='Weight ~ Time + (1|Litter) + (1|Time) + (1|Pig)', family="gaussian", data=df)
+
     print(model.fit())
     return model
 
@@ -25,6 +26,7 @@ def fit_model():
 # Read more here: https://sscc.wisc.edu/sscc/pubs/RegressionDiagnostics.html
 def show_model_diagnostics(model): 
 
+    plt.axhline(y=0, color='r', linestyle='-')
     plt.scatter(model.fits, model.residuals)
     plt.title("Fitted values vs. Residuals")
     plt.xlabel("fitted values")

examples/Exercise/tisane_generated_files/model_data_path.py

@@ -13,6 +13,7 @@ def fit_model():
 
 
     model = smf.glm(formula='endurance ~ age + exercise', data=df, family=sm.families.Gaussian(sm.families.links.identity()))
+
     res = model.fit()
     print(res.summary())
     return model
@@ -26,6 +27,7 @@ def show_model_diagnostics(model):
     res = model.fit()
     plt.clf()
     plt.grid(True)
+    plt.axhline(y=0, color='r', linestyle='-')
     plt.plot(res.predict(linear=True), res.resid_pearson, 'o')
     plt.xlabel("Linear predictor")
     plt.ylabel("Residual")

examples/Exercise/tisane_generated_files/model_df.py

@@ -16,6 +16,7 @@ def fit_model():
 
 
     model = smf.glm(formula='endurance ~ age + exercise', data=df, family=sm.families.Gaussian(sm.families.links.identity()))
+
     res = model.fit()
     print(res.summary())
     return model
@@ -29,6 +30,7 @@ def show_model_diagnostics(model):
     res = model.fit()
     plt.clf()
     plt.grid(True)
+    plt.axhline(y=0, color='r', linestyle='-')
     plt.plot(res.predict(linear=True), res.resid_pearson, 'o')
     plt.xlabel("Linear predictor")
     plt.ylabel("Residual")

examples/Exercise/tisane_generated_files/model_no_data.py

@@ -17,6 +17,7 @@ def fit_model():
 
 
     model = smf.glm(formula='endurance ~ age + exercise', data=df, family=sm.families.Gaussian(sm.families.links.identity()))
+
     res = model.fit()
     print(res.summary())
     return model
@@ -30,6 +31,7 @@ def show_model_diagnostics(model):
     res = model.fit()
     plt.clf()
     plt.grid(True)
+    plt.axhline(y=0, color='r', linestyle='-')
     plt.plot(res.predict(linear=True), res.resid_pearson, 'o')
     plt.xlabel("Linear predictor")
     plt.ylabel("Residual")

examples/Group_Exercise/tisane_generated_files/model_data_path.py

@@ -12,6 +12,7 @@ def fit_model():
 
 
     model = Lmer(formula='pounds_lost ~ motivation + treatment + (1|group)', family="gaussian", data=df)
+
     print(model.fit())
     return model
 
@@ -21,6 +22,7 @@ def fit_model():
 # Read more here: https://sscc.wisc.edu/sscc/pubs/RegressionDiagnostics.html
 def show_model_diagnostics(model): 
 
+    plt.axhline(y=0, color='r', linestyle='-')
     plt.scatter(model.fits, model.residuals)
     plt.title("Fitted values vs. Residuals")
     plt.xlabel("fitted values")

examples/Group_Exercise/tisane_generated_files/model_df.py

@@ -15,6 +15,7 @@ def fit_model():
 
 
     model = Lmer(formula='pounds_lost ~ motivation + treatment + (1|group)', family="gaussian", data=df)
+
     print(model.fit())
     return model
 
@@ -24,6 +25,7 @@ def fit_model():
 # Read more here: https://sscc.wisc.edu/sscc/pubs/RegressionDiagnostics.html
 def show_model_diagnostics(model): 
 
+    plt.axhline(y=0, color='r', linestyle='-')
     plt.scatter(model.fits, model.residuals)
     plt.title("Fitted values vs. Residuals")
     plt.xlabel("fitted values")

examples/Group_Exercise/tisane_generated_files/model_no_data.py

@@ -16,6 +16,7 @@ def fit_model():
 
 
     model = Lmer(formula='pounds_lost ~ motivation + treatment + (1|group)', family="gaussian", data=df)
+
     print(model.fit())
     return model
 
@@ -25,6 +26,7 @@ def fit_model():
 # Read more here: https://sscc.wisc.edu/sscc/pubs/RegressionDiagnostics.html
 def show_model_diagnostics(model): 
 
+    plt.axhline(y=0, color='r', linestyle='-')
     plt.scatter(model.fits, model.residuals)
     plt.title("Fitted values vs. Residuals")
     plt.xlabel("fitted values")