updated automate test

setup.py

@@ -2,7 +2,7 @@
 
 # Package information
 name = 'multilearn'
-version = '0.0.6'  # Need to increment every time to push to PyPI
+version = '0.0.7'  # Need to increment every time to push to PyPI
 description = 'Multi-task learning with Pytorch.'
 url = 'https://github.com/leschultz/multilearn'
 author = 'Lane E. Schultz'

tests/test_fit.py

@@ -1,9 +1,14 @@
+from sklearn.model_selection import RepeatedKFold
 from sklearn.preprocessing import StandardScaler
 from multilearn import datasets, models, utils
 from torch import optim, nn
 
+import pandas as pd
+import numpy as np
+
 import unittest
 import shutil
+import os
 
 
 class ml_test(unittest.TestCase):
@@ -13,31 +18,109 @@ def test_ml(self):
         save_dir = 'outputs'
         lr = 1e-4
         batch_size = 32
-        n_epochs = 2
-        tasks = ['toy1', 'toy2', 'friedman1']
+        n_epochs = 100
+        patience = 10  # Learning loop patience
+        train_size = 0.8  # Traning fraction
+        val_size = 1.0-train_size  # Validation fraction
+        print_n = n_epochs//10
+
+        # Local data 1
+        local1 = './local1.csv'
+
+        # Create local data 1
+        X1 = np.random.uniform(size=(500, 1))
+        y1 = 10*np.sin(np.pi*X1[:, 0])
+
+        # Save local data 1
+        df = pd.DataFrame(X1)
+        df['target_1'] = y1
+        df.to_csv(local1, index=False)
+
+        # Local data 2
+        local2 = './local2.csv'
+
+        # Create local data 2
+        X2 = np.random.uniform(size=(50, 7))
+        y2 = (
+              10*np.sin(np.pi*X2[:, 0]*X2[:, 1])
+              + 20*(X2[:, 2]-0.5)**2
+              + 10*X2[:, 3]
+              + 5*X2[:, 4]
+              )
+
+        # Save local data 2
+        df = pd.DataFrame(X2)
+        df['target_2'] = y2
+        df.to_csv(local2, index=False)
 
-        # Data
-        X, y = datasets.load(tasks)
-        data = datasets.splitter(X, y, tasks, train_size=1)
+        # Combine data to load
+        locations = [local1, local2]
 
-        for k, v in data.items():
-            data[k]['scaler'] = StandardScaler()
-            data[k]['loss'] = nn.L1Loss()
+        # Load data in dictionary (make sure to keep order for loading items)
+        tasks = ['name1', 'name2']
+        data = datasets.load(
+                             locations,
+                             names=tasks,  # User defined name
+                             targets=['target_1', 'target_2'],  # Target names
+                             drops=[None, ['5', '6']],  # Columns to drop
+                             )
 
-        model = models.MultiNet(tasks=tasks, input_arch={50: 1})
+        # Clean generated csv file
+        [os.remove(i) for i in locations]
+
+        # Scalers and loss corresponding to loaded Xs and ys
+        for key, value in data.items():
+            value['scaler'] = StandardScaler()
+            value['loss'] = nn.L1Loss()
+
+        # A single model that combines nodes during training
+        model = models.MultiNet(
+                                tasks=tasks,
+                                input_arch={100: 1},
+                                mid_arch={100: 1, 50: 1},
+                                out_arch={50: 1, 10: 1}
+                                )
+
+        # The optimizer for the NN model
         optimizer = optim.Adam
 
-        out = utils.train(
-                          model,
-                          optimizer,
-                          data,
-                          n_epochs=n_epochs,
-                          batch_size=batch_size,
-                          lr=lr,
-                          save_dir=save_dir,
-                          )
-
-        print(out['df_loss'])
+        # Do CV to assess
+        utils.cv(
+                 data,
+                 model,
+                 optimizer,
+                 RepeatedKFold(n_repeats=1),
+                 train_size=train_size,
+                 val_size=val_size,
+                 save_dir=save_dir,
+                 lr=lr,
+                 batch_size=batch_size,
+                 n_epochs=n_epochs,
+                 patience=patience,
+                 print_n=print_n,
+                 )
+
+        # Save one model to all data
+        model = utils.full_fit(
+                               data,
+                               model,
+                               optimizer,
+                               train_size=train_size,
+                               val_size=val_size,
+                               save_dir=save_dir,
+                               lr=lr,
+                               batch_size=batch_size,
+                               n_epochs=n_epochs,
+                               patience=patience,
+                               print_n=print_n,
+                               )
+
+        name = tasks[1]
+        X_inference = data[name]['X']  # Data with dropped columns
+
+        print(f'Model used for predicting {name}')
+        print(model.predict(X_inference, name))
+
         shutil.rmtree(save_dir)