fix fpmax issue (#570) with fptrees that contain no nodes

docs/sources/CHANGELOG.md

@@ -7,6 +7,31 @@ The CHANGELOG for the current development version is available at
 
 ---
 
+### Version 0.18.0 (TBD)
+
+##### Downloads
+
+- [Source code (zip)](https://github.com/rasbt/mlxtend/archive/v0.18.0.zip)
+
+- [Source code (tar.gz)](https://github.com/rasbt/mlxtend/archive/v0.18.0.tar.gz)
+
+##### New Features
+
+- -
+
+##### Changes
+
+- -
+
+##### Bug Fixes
+
+- Behavior of `fpgrowth` and `apriori` consistent for edgecases such as `min_support=0`. ([#573](https://github.com/rasbt/mlxtend/pull/550) via [Steve Harenberg](https://github.com/harenbergsd))
+- `fpmax` returns an empty data frame now instead of raising an error if the frequent itemset set is empty. ([#573](https://github.com/rasbt/mlxtend/pull/550) via [Steve Harenberg](https://github.com/harenbergsd))
+
+
+
+
+
 ### Version 0.17.0 (07/19/2019)
 
 ##### Downloads

mlxtend/frequent_patterns/apriori.py

@@ -143,6 +143,11 @@ def _support(_x, _n_rows, _is_sparse):
         out = (np.sum(_x, axis=0) / _n_rows)
         return np.array(out).reshape(-1)
 
+    if min_support <= 0.:
+        raise ValueError('`min_support` must be a positive '
+                         'number within the interval `(0, 1]`. '
+                         'Got %s.' % min_support)
+
     idxs = np.where((df.values != 1) & (df.values != 0))
     if len(idxs[0]) > 0:
         val = df.values[idxs[0][0], idxs[1][0]]

mlxtend/frequent_patterns/fpcommon.py

@@ -123,6 +123,8 @@ def insert_itemset(self, itemset, count=1):
         count : int
             The number of occurrences of the itemset.
         """
+        self.root.count += count
+
         if len(itemset) == 0:
             return
 
@@ -162,7 +164,7 @@ def print_status(self, count, colnames):
 
 
 class FPNode(object):
-    def __init__(self, item, count=1, parent=None):
+    def __init__(self, item, count=0, parent=None):
         self.item = item
         self.count = count
         self.parent = parent

mlxtend/frequent_patterns/fpgrowth.py

@@ -60,6 +60,11 @@ def fpgrowth(df, min_support=0.5, use_colnames=False, max_len=None, verbose=0):
     """
     fpc.valid_input_check(df)
 
+    if min_support <= 0.:
+        raise ValueError('`min_support` must be a positive '
+                         'number within the interval `(0, 1]`. '
+                         'Got %s.' % min_support)
+
     colname_map = None
     if use_colnames:
         colname_map = {idx: item for idx, item in enumerate(df.columns)}

mlxtend/frequent_patterns/fpmax.py

@@ -61,6 +61,11 @@ def fpmax(df, min_support=0.5, use_colnames=False, max_len=None, verbose=0):
     """
     fpc.valid_input_check(df)
 
+    if min_support <= 0.:
+        raise ValueError('`min_support` must be a positive '
+                         'number within the interval `(0, 1]`. '
+                         'Got %s.' % min_support)
+
     colname_map = None
     if use_colnames:
         colname_map = {idx: item for idx, item in enumerate(df.columns)}
@@ -78,14 +83,16 @@ def fpmax_step(tree, minsup, mfit, colnames, max_len, verbose):
     count = 0
     items = list(tree.nodes.keys())
     largest_set = sorted(tree.cond_items+items, key=mfit.rank.get)
+    if len(largest_set) == 0:
+        return
     if tree.is_path():
         if not mfit.contains(largest_set):
             count += 1
             largest_set.reverse()
             mfit.cache = largest_set
             mfit.insert_itemset(largest_set)
             if max_len is None or len(largest_set) <= max_len:
-                support = min([tree.nodes[i][0].count for i in items])
+                support = tree.root.count
                 yield support, largest_set
 
     if verbose:

mlxtend/frequent_patterns/tests/test_apriori.py

@@ -6,25 +6,36 @@
 
 import unittest
 import numpy as np
-from mlxtend.frequent_patterns.tests.test_fpbase import FPTestAll
+from test_fpbase import FPTestEdgeCases, FPTestErrors, \
+                        FPTestEx1All, FPTestEx2All, FPTestEx3All
 from mlxtend.frequent_patterns import apriori
 
 
 def apriori_wrapper_low_memory(*args, **kwargs):
     return apriori(*args, **kwargs, low_memory=True)
 
 
-class TestApriori(unittest.TestCase, FPTestAll):
+class TestEdgeCases(unittest.TestCase, FPTestEdgeCases):
     def setUp(self):
-        FPTestAll.setUp(self, apriori)
+        FPTestEdgeCases.setUp(self, apriori)
 
 
-class TestAprioriLowMemory(unittest.TestCase, FPTestAll):
+class TestErrors(unittest.TestCase, FPTestErrors):
     def setUp(self):
-        FPTestAll.setUp(self, apriori_wrapper_low_memory)
+        FPTestErrors.setUp(self, apriori)
 
 
-class TestAprioriBinaryInput(unittest.TestCase, FPTestAll):
+class TestApriori(unittest.TestCase, FPTestEx1All):
+    def setUp(self):
+        FPTestEx1All.setUp(self, apriori)
+
+
+class TestAprioriLowMemory(unittest.TestCase, FPTestEx1All):
+    def setUp(self):
+        FPTestEx1All.setUp(self, apriori_wrapper_low_memory)
+
+
+class TestAprioriBoolInput(unittest.TestCase, FPTestEx1All):
     def setUp(self):
         one_ary = np.array(
             [[False, False, False, True, False, True, True, True, True,
@@ -37,4 +48,14 @@ def setUp(self):
               True, True],
              [False, True, False, True, True, True, False, False, True,
               False, False]])
-        FPTestAll.setUp(self, apriori, one_ary=one_ary)
+        FPTestEx1All.setUp(self, apriori, one_ary=one_ary)
+
+
+class TestEx2(unittest.TestCase, FPTestEx2All):
+    def setUp(self):
+        FPTestEx2All.setUp(self, apriori)
+
+
+class TestEx3(unittest.TestCase, FPTestEx3All):
+    def setUp(self):
+        FPTestEx3All.setUp(self, apriori)

mlxtend/frequent_patterns/tests/test_fpbase.py

@@ -7,6 +7,7 @@
 import numpy as np
 from numpy.testing import assert_array_equal
 from mlxtend.utils import assert_raises
+from mlxtend.preprocessing import TransactionEncoder
 import pandas as pd
 import sys
 from contextlib import contextmanager
@@ -24,30 +25,37 @@ def captured_output():
         sys.stdout, sys.stderr = old_out, old_err
 
 
-class FPTestBase(object):
+class FPTestEdgeCases(object):
     """
-    Base testing class for frequent pattern mining. This class should include
-    setup and tests common to all methods (e.g., error for improper input)
+    Base class for testing edge cases for pattern mining.
     """
 
-    def setUp(self, fpalgo, one_ary=None):
-        if one_ary is None:
-            self.one_ary = np.array(
-               [[0, 0, 0, 1, 0, 1, 1, 1, 1, 0, 1],
-                [0, 0, 1, 1, 0, 1, 0, 1, 1, 0, 1],
-                [1, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0],
-                [0, 1, 0, 0, 0, 1, 1, 0, 0, 1, 1],
-                [0, 1, 0, 1, 1, 1, 0, 0, 1, 0, 0]])
+    def setUp(self, fpalgo):
+        self.fpalgo = fpalgo
 
-        else:
-            self.one_ary = one_ary
+    def test_empty(self):
+        df = pd.DataFrame([[]])
+        res_df = self.fpalgo(df)
+        expect = pd.DataFrame([], columns=['support', 'itemsets'])
+        compare_dataframes(res_df, expect)
 
+
+class FPTestErrors(object):
+    """
+    Base class for testing expected errors for pattern mining.
+    """
+
+    def setUp(self, fpalgo):
+        self.one_ary = np.array(
+            [[0, 0, 0, 1, 0, 1, 1, 1, 1, 0, 1],
+             [0, 0, 1, 1, 0, 1, 0, 1, 1, 0, 1],
+             [1, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0],
+             [0, 1, 0, 0, 0, 1, 1, 0, 0, 1, 1],
+             [0, 1, 0, 1, 1, 1, 0, 0, 1, 0, 0]])
         self.cols = ['Apple', 'Corn', 'Dill', 'Eggs', 'Ice cream',
                      'Kidney Beans', 'Milk',
                      'Nutmeg', 'Onion', 'Unicorn', 'Yogurt']
-
         self.df = pd.DataFrame(self.one_ary, columns=self.cols)
-
         self.fpalgo = fpalgo
 
     def test_itemsets_type(self):
@@ -84,6 +92,31 @@ def test_sparsedataframe_notzero_column(self):
                       '`df.columns = [str(i) for i in df.columns`].',
                       self.fpalgo, dfs)
 
+
+class FPTestEx1(object):
+    """
+    Base class for testing frequent pattern mining on a small example.
+    """
+
+    def setUp(self, fpalgo, one_ary=None):
+        if one_ary is None:
+            self.one_ary = np.array(
+                [[0, 0, 0, 1, 0, 1, 1, 1, 1, 0, 1],
+                 [0, 0, 1, 1, 0, 1, 0, 1, 1, 0, 1],
+                 [1, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0],
+                 [0, 1, 0, 0, 0, 1, 1, 0, 0, 1, 1],
+                 [0, 1, 0, 1, 1, 1, 0, 0, 1, 0, 0]])
+        else:
+            self.one_ary = one_ary
+
+        self.cols = ['Apple', 'Corn', 'Dill', 'Eggs', 'Ice cream',
+                     'Kidney Beans', 'Milk',
+                     'Nutmeg', 'Onion', 'Unicorn', 'Yogurt']
+
+        self.df = pd.DataFrame(self.one_ary, columns=self.cols)
+
+        self.fpalgo = fpalgo
+
     def test_frozenset_selection(self):
         res_df = self.fpalgo(self.df, use_colnames=True)
         assert res_df.values.shape == self.fpalgo(self.df).values.shape
@@ -117,9 +150,9 @@ def test_with_fill_values(fill_value):
         test_with_fill_values(False)
 
 
-class FPTestAll(FPTestBase):
+class FPTestEx1All(FPTestEx1):
     def setUp(self, fpalgo, one_ary=None):
-        FPTestBase.setUp(self, fpalgo, one_ary=one_ary)
+        FPTestEx1.setUp(self, fpalgo, one_ary=one_ary)
 
     def test_default(self):
         res_df = self.fpalgo(self.df)
@@ -162,27 +195,62 @@ def test_low_memory_flag(self):
             assert True
 
 
-class FPTestMaximal(FPTestBase):
-    def setUp(self, fpalgo, one_ary=None):
-        FPTestBase.setUp(self, fpalgo, one_ary=one_ary)
+class FPTestEx2(object):
+    """
+    Base class for testing frequent pattern mining on a small example.
+    """
 
-    def test_default(self):
-        res_df = self.fpalgo(self.df)
-        expect = pd.DataFrame([[0.6, frozenset([5, 6])],
-                               [0.6, frozenset([5, 10])],
-                               [0.6, frozenset([3, 5, 8])]],
+    def setUp(self):
+        database = [['a'], ['b'], ['c', 'd'], ['e']]
+        te = TransactionEncoder()
+        te_ary = te.fit(database).transform(database)
+
+        self.df = pd.DataFrame(te_ary, columns=te.columns_)
+
+
+class FPTestEx2All(FPTestEx2):
+    def setUp(self, fpalgo):
+        self.fpalgo = fpalgo
+        FPTestEx2.setUp(self)
+
+    def test_output(self):
+        res_df = self.fpalgo(self.df, min_support=0.001, use_colnames=True)
+        expect = pd.DataFrame([[0.25, frozenset(['a'])],
+                               [0.25, frozenset(['b'])],
+                               [0.25, frozenset(['c'])],
+                               [0.25, frozenset(['d'])],
+                               [0.25, frozenset(['e'])],
+                               [0.25, frozenset(['c', 'd'])]],
                               columns=['support', 'itemsets'])
 
         compare_dataframes(res_df, expect)
 
-    def test_max_len(self):
-        res_df1 = self.fpalgo(self.df)
-        max_len = np.max(res_df1['itemsets'].apply(len))
-        assert max_len == 3
 
-        res_df2 = self.fpalgo(self.df, max_len=2)
-        max_len = np.max(res_df2['itemsets'].apply(len))
-        assert max_len == 2
+class FPTestEx3(object):
+    """
+    Base class for testing frequent pattern mining on a small example.
+    """
+
+    def setUp(self):
+        database = [['a'], ['b'], ['c', 'd'], ['e']]
+        te = TransactionEncoder()
+        te_ary = te.fit(database).transform(database)
+
+        self.df = pd.DataFrame(te_ary, columns=te.columns_)
+
+
+class FPTestEx3All(FPTestEx3):
+    def setUp(self, fpalgo):
+        self.fpalgo = fpalgo
+        FPTestEx3.setUp(self)
+
+    def test_output3(self):
+        assert_raises(ValueError,
+                      '`min_support` must be a positive '
+                      'number within the interval `(0, 1]`. Got 0.0.',
+                      self.fpalgo,
+                      self.df,
+                      min_support=0.)
 
 
 def compare_dataframes(df1, df2):

mlxtend/frequent_patterns/tests/test_fpgrowth.py

@@ -1,15 +1,26 @@
 import unittest
 import numpy as np
-from mlxtend.frequent_patterns.tests.test_fpbase import FPTestAll
+from test_fpbase import FPTestEdgeCases, FPTestErrors, \
+                        FPTestEx1All, FPTestEx2All, FPTestEx3All
 from mlxtend.frequent_patterns import fpgrowth
 
 
-class TestFPGrowth(unittest.TestCase, FPTestAll):
+class TestEdgeCases(unittest.TestCase, FPTestEdgeCases):
     def setUp(self):
-        FPTestAll.setUp(self, fpgrowth)
+        FPTestEdgeCases.setUp(self, fpgrowth)
 
 
-class TestFPGrowth2(unittest.TestCase, FPTestAll):
+class TestErrors(unittest.TestCase, FPTestErrors):
+    def setUp(self):
+        FPTestErrors.setUp(self, fpgrowth)
+
+
+class TestEx1(unittest.TestCase, FPTestEx1All):
+    def setUp(self):
+        FPTestEx1All.setUp(self, fpgrowth)
+
+
+class TestEx1BoolInput(unittest.TestCase, FPTestEx1All):
     def setUp(self):
         one_ary = np.array(
             [[False, False, False, True, False, True, True, True, True,
@@ -22,4 +33,14 @@ def setUp(self):
               True, True],
              [False, True, False, True, True, True, False, False, True,
               False, False]])
-        FPTestAll.setUp(self, fpgrowth, one_ary=one_ary)
+        FPTestEx1All.setUp(self, fpgrowth, one_ary=one_ary)
+
+
+class TestEx2(unittest.TestCase, FPTestEx2All):
+    def setUp(self):
+        FPTestEx2All.setUp(self, fpgrowth)
+
+
+class TestEx3(unittest.TestCase, FPTestEx3All):
+    def setUp(self):
+        FPTestEx3All.setUp(self, fpgrowth)