add support for pandas DataFrame (keras-team#8199)

* add support for pandas DataFrame

* multiple updates according to @fchollet's review

* DataFrame should be handled correctly if list/dict is passed as model inputs/outputs

keras/engine/training.py

@@ -60,6 +60,9 @@ def _standardize_input_data(data, names, shapes=None,
     if data is None:
         return [None for _ in range(len(names))]
     if isinstance(data, dict):
+        for key, value in data.items():
+            if value.__class__.__name__ == 'DataFrame':
+                data[key] = value.values
         arrays = []
         for name in names:
             if name not in data:
@@ -68,6 +71,9 @@ def _standardize_input_data(data, names, shapes=None,
                                  str(names))
             arrays.append(data[name])
     elif isinstance(data, list):
+        for key, value in enumerate(data):
+            if value.__class__.__name__ == 'DataFrame':
+                data[key] = value.values
         if len(data) != len(names):
             if data and hasattr(data[0], 'shape'):
                 raise ValueError('Error when checking model ' +
@@ -95,6 +101,9 @@ def _standardize_input_data(data, names, shapes=None,
                         'The list you passed was: ' +
                         str(data)[:200])
         arrays = data
+    elif data.__class__.__name__ == 'DataFrame':
+        # test if data is a DataFrame, without pandas installed
+        data = data.values
     else:
         if not hasattr(data, 'shape'):
             raise TypeError('Error when checking model ' +

setup.py

@@ -20,7 +20,8 @@
           'tests': ['pytest',
                     'pytest-pep8',
                     'pytest-xdist',
-                    'pytest-cov'],
+                    'pytest-cov',
+                    'pandas'],
       },
       classifiers=[
           'Development Status :: 5 - Production/Stable',

tests/keras/engine/test_training.py

@@ -1,5 +1,6 @@
 import pytest
 import numpy as np
+import pandas as pd
 from numpy.testing import assert_allclose
 import sys
 import scipy.sparse as sparse
@@ -106,9 +107,13 @@ def test_model_methods():
 
     input_a_np = np.random.random((10, 3))
     input_b_np = np.random.random((10, 3))
+    input_a_df = pd.DataFrame(input_a_np)
+    input_b_df = pd.DataFrame(input_b_np)
 
     output_a_np = np.random.random((10, 4))
     output_b_np = np.random.random((10, 3))
+    output_a_df = pd.DataFrame(output_a_np)
+    output_b_df = pd.DataFrame(output_b_np)
 
     # training/testing doesn't work before compiling.
     with pytest.raises(RuntimeError):
@@ -124,6 +129,8 @@ def test_model_methods():
                                [output_a_np, output_b_np])
     out = model.train_on_batch({'input_a': input_a_np, 'input_b': input_b_np},
                                {'dense_1': output_a_np, 'dropout': output_b_np})
+    out = model.train_on_batch([input_a_df, input_b_df],
+                               [output_a_df, output_b_df])
 
     # test fit
     out = model.fit([input_a_np, input_b_np],
@@ -133,6 +140,8 @@ def test_model_methods():
     out = model.fit({'input_a': input_a_np, 'input_b': input_b_np},
                     {'dense_1': output_a_np, 'dropout': output_b_np},
                     epochs=1, batch_size=4)
+    out = model.fit([input_a_df, input_b_df],
+                    [output_a_df, output_b_df], epochs=1, batch_size=4)
 
     # test validation_split
     out = model.fit([input_a_np, input_b_np],
@@ -165,10 +174,13 @@ def test_model_methods():
                               [output_a_np, output_b_np])
     out = model.test_on_batch({'input_a': input_a_np, 'input_b': input_b_np},
                               {'dense_1': output_a_np, 'dropout': output_b_np})
+    out = model.test_on_batch([input_a_df, input_b_df],
+                              [output_a_df, output_b_df])
 
     # predict_on_batch
     out = model.predict_on_batch([input_a_np, input_b_np])
     out = model.predict_on_batch({'input_a': input_a_np, 'input_b': input_b_np})
+    out = model.predict_on_batch([input_a_df, input_b_df])
 
     # predict, evaluate
     input_a_np = np.random.random((10, 3))
@@ -178,7 +190,9 @@ def test_model_methods():
     output_b_np = np.random.random((10, 3))
 
     out = model.evaluate([input_a_np, input_b_np], [output_a_np, output_b_np], batch_size=4)
+    out = model.evaluate([input_a_df, input_b_df], [output_a_df, output_b_df], batch_size=4)
     out = model.predict([input_a_np, input_b_np], batch_size=4)
+    out = model.predict([input_a_df, input_b_df], batch_size=4)
 
     # with sample_weight
     input_a_np = np.random.random((10, 3))