PERF: fix easily fixed issues in asv benchmarks

- eval.eval_frame_*: fix up local variables
- groupby: range() -> list(range()) and using np.random.choice
- plotting: ensure matplotlib Agg backend is used
- packers: fix data file paths

Author: pv

asv_bench/benchmarks/eval.py

@@ -3,192 +3,36 @@
 import pandas.computation.expressions as expr
 
 
-class eval_frame_add_all_threads(object):
+class eval_frame(object):
     goal_time = 0.2
 
-    def setup(self):
-        self.df = DataFrame(np.random.randn(20000, 100))
-        self.df2 = DataFrame(np.random.randn(20000, 100))
-        self.df3 = DataFrame(np.random.randn(20000, 100))
-        self.df4 = DataFrame(np.random.randn(20000, 100))
-
-    def time_eval_frame_add_all_threads(self):
-        pd.eval('df + df2 + df3 + df4')
-
-
-class eval_frame_add_one_thread(object):
-    goal_time = 0.2
-
-    def setup(self):
-        self.df = DataFrame(np.random.randn(20000, 100))
-        self.df2 = DataFrame(np.random.randn(20000, 100))
-        self.df3 = DataFrame(np.random.randn(20000, 100))
-        self.df4 = DataFrame(np.random.randn(20000, 100))
-        expr.set_numexpr_threads(1)
-
-    def time_eval_frame_add_one_thread(self):
-        pd.eval('df + df2 + df3 + df4')
-
-
-class eval_frame_add_python(object):
-    goal_time = 0.2
-
-    def setup(self):
-        self.df = DataFrame(np.random.randn(20000, 100))
-        self.df2 = DataFrame(np.random.randn(20000, 100))
-        self.df3 = DataFrame(np.random.randn(20000, 100))
-        self.df4 = DataFrame(np.random.randn(20000, 100))
-
-    def time_eval_frame_add_python(self):
-        pd.eval('df + df2 + df3 + df4', engine='python')
-
-
-class eval_frame_add_python_one_thread(object):
-    goal_time = 0.2
-
-    def setup(self):
-        self.df = DataFrame(np.random.randn(20000, 100))
-        self.df2 = DataFrame(np.random.randn(20000, 100))
-        self.df3 = DataFrame(np.random.randn(20000, 100))
-        self.df4 = DataFrame(np.random.randn(20000, 100))
-        expr.set_numexpr_threads(1)
-
-    def time_eval_frame_add_python_one_thread(self):
-        pd.eval('df + df2 + df3 + df4', engine='python')
-
-
-class eval_frame_and_all_threads(object):
-    goal_time = 0.2
-
-    def setup(self):
-        self.df = DataFrame(np.random.randn(20000, 100))
-        self.df2 = DataFrame(np.random.randn(20000, 100))
-        self.df3 = DataFrame(np.random.randn(20000, 100))
-        self.df4 = DataFrame(np.random.randn(20000, 100))
-
-    def time_eval_frame_and_all_threads(self):
-        pd.eval('(df > 0) & (df2 > 0) & (df3 > 0) & (df4 > 0)')
-
-
-class eval_frame_and_python_one_thread(object):
-    goal_time = 0.2
-
-    def setup(self):
-        self.df = DataFrame(np.random.randn(20000, 100))
-        self.df2 = DataFrame(np.random.randn(20000, 100))
-        self.df3 = DataFrame(np.random.randn(20000, 100))
-        self.df4 = DataFrame(np.random.randn(20000, 100))
-        expr.set_numexpr_threads(1)
-
-    def time_eval_frame_and_python_one_thread(self):
-        pd.eval('(df > 0) & (df2 > 0) & (df3 > 0) & (df4 > 0)', engine='python')
-
-
-class eval_frame_and_python(object):
-    goal_time = 0.2
-
-    def setup(self):
-        self.df = DataFrame(np.random.randn(20000, 100))
-        self.df2 = DataFrame(np.random.randn(20000, 100))
-        self.df3 = DataFrame(np.random.randn(20000, 100))
-        self.df4 = DataFrame(np.random.randn(20000, 100))
-
-    def time_eval_frame_and_python(self):
-        pd.eval('(df > 0) & (df2 > 0) & (df3 > 0) & (df4 > 0)', engine='python')
-
-
-class eval_frame_chained_cmp_all_threads(object):
-    goal_time = 0.2
-
-    def setup(self):
-        self.df = DataFrame(np.random.randn(20000, 100))
-        self.df2 = DataFrame(np.random.randn(20000, 100))
-        self.df3 = DataFrame(np.random.randn(20000, 100))
-        self.df4 = DataFrame(np.random.randn(20000, 100))
-
-    def time_eval_frame_chained_cmp_all_threads(self):
-        pd.eval('df < df2 < df3 < df4')
-
-
-class eval_frame_chained_cmp_python_one_thread(object):
-    goal_time = 0.2
+    params = [['numexpr', 'python'], [1, 'all']]
+    param_names = ['engine', 'threads']
 
-    def setup(self):
-        self.df = DataFrame(np.random.randn(20000, 100))
-        self.df2 = DataFrame(np.random.randn(20000, 100))
-        self.df3 = DataFrame(np.random.randn(20000, 100))
-        self.df4 = DataFrame(np.random.randn(20000, 100))
-        expr.set_numexpr_threads(1)
-
-    def time_eval_frame_chained_cmp_python_one_thread(self):
-        pd.eval('df < df2 < df3 < df4', engine='python')
-
-
-class eval_frame_chained_cmp_python(object):
-    goal_time = 0.2
-
-    def setup(self):
+    def setup(self, engine, threads):
         self.df = DataFrame(np.random.randn(20000, 100))
         self.df2 = DataFrame(np.random.randn(20000, 100))
         self.df3 = DataFrame(np.random.randn(20000, 100))
         self.df4 = DataFrame(np.random.randn(20000, 100))
 
-    def time_eval_frame_chained_cmp_python(self):
-        pd.eval('df < df2 < df3 < df4', engine='python')
+        if threads == 1:
+            expr.set_numexpr_threads(1)
 
+    def time_add(self, engine, threads):
+        df, df2, df3, df4 = self.df, self.df2, self.df3, self.df4
+        pd.eval('df + df2 + df3 + df4', engine=engine)
 
-class eval_frame_mult_all_threads(object):
-    goal_time = 0.2
+    def time_and(self, engine, threads):
+        df, df2, df3, df4 = self.df, self.df2, self.df3, self.df4
+        pd.eval('(df > 0) & (df2 > 0) & (df3 > 0) & (df4 > 0)', engine=engine)
 
-    def setup(self):
-        self.df = DataFrame(np.random.randn(20000, 100))
-        self.df2 = DataFrame(np.random.randn(20000, 100))
-        self.df3 = DataFrame(np.random.randn(20000, 100))
-        self.df4 = DataFrame(np.random.randn(20000, 100))
-
-    def time_eval_frame_mult_all_threads(self):
-        pd.eval('df * df2 * df3 * df4')
-
-
-class eval_frame_mult_one_thread(object):
-    goal_time = 0.2
-
-    def setup(self):
-        self.df = DataFrame(np.random.randn(20000, 100))
-        self.df2 = DataFrame(np.random.randn(20000, 100))
-        self.df3 = DataFrame(np.random.randn(20000, 100))
-        self.df4 = DataFrame(np.random.randn(20000, 100))
-        expr.set_numexpr_threads(1)
-
-    def time_eval_frame_mult_one_thread(self):
-        pd.eval('df * df2 * df3 * df4')
-
-
-class eval_frame_mult_python(object):
-    goal_time = 0.2
-
-    def setup(self):
-        self.df = DataFrame(np.random.randn(20000, 100))
-        self.df2 = DataFrame(np.random.randn(20000, 100))
-        self.df3 = DataFrame(np.random.randn(20000, 100))
-        self.df4 = DataFrame(np.random.randn(20000, 100))
-
-    def time_eval_frame_mult_python(self):
-        pd.eval('df * df2 * df3 * df4', engine='python')
-
-
-class eval_frame_mult_python_one_thread(object):
-    goal_time = 0.2
-
-    def setup(self):
-        self.df = DataFrame(np.random.randn(20000, 100))
-        self.df2 = DataFrame(np.random.randn(20000, 100))
-        self.df3 = DataFrame(np.random.randn(20000, 100))
-        self.df4 = DataFrame(np.random.randn(20000, 100))
-        expr.set_numexpr_threads(1)
+    def time_chained_cmp(self, engine, threads):
+        df, df2, df3, df4 = self.df, self.df2, self.df3, self.df4
+        pd.eval('df < df2 < df3 < df4', engine=engine)
 
-    def time_eval_frame_mult_python_one_thread(self):
-        pd.eval('df * df2 * df3 * df4', engine='python')
+    def time_mult(self, engine, threads):
+        df, df2, df3, df4 = self.df, self.df2, self.df3, self.df4
+        pd.eval('df * df2 * df3 * df4', engine=engine)
 
 
 class query_datetime_index(object):
@@ -203,6 +47,7 @@ def setup(self):
         self.df = DataFrame({'a': np.random.randn(self.N), }, index=self.index)
 
     def time_query_datetime_index(self):
+        ts = self.ts
         self.df.query('index < @ts')
 
 
@@ -218,6 +63,7 @@ def setup(self):
         self.df = DataFrame({'dates': self.s.values, })
 
     def time_query_datetime_series(self):
+        ts = self.ts
         self.df.query('dates < @ts')
 
 
@@ -236,4 +82,5 @@ def setup(self):
         self.max_val = self.df['a'].max()
 
     def time_query_with_boolean_selection(self):
-        self.df.query('(a >= @min_val) & (a <= @max_val)')
+        min_val, max_val = self.min_val, self.max_val
+        self.df.query('(a >= @min_val) & (a <= @max_val)')

asv_bench/benchmarks/groupby.py

@@ -254,7 +254,7 @@ def setup(self):
         self.offsets[(np.random.rand(self.n) > 0.5)] = np.timedelta64('nat')
         self.value2 = np.random.randn(self.n)
         self.value2[(np.random.rand(self.n) > 0.5)] = np.nan
-        self.obj = tm.choice(list('ab'), size=self.n).astype(object)
+        self.obj = np.random.choice(list('ab'), size=self.n).astype(object)
         self.obj[(np.random.randn(self.n) > 0.5)] = np.nan
         self.df = DataFrame({'key1': np.random.randint(0, 500, size=self.n),
                              'key2': np.random.randint(0, 100, size=self.n),
@@ -651,7 +651,7 @@ class groupby_sum_multiindex(object):
 
     def setup(self):
         self.N = 50
-        self.df = DataFrame({'A': (range(self.N) * 2), 'B': range((self.N * 2)), 'C': 1, }).set_index(['A', 'B'])
+        self.df = DataFrame({'A': (list(range(self.N)) * 2), 'B': list(range((self.N * 2))), 'C': 1, }).set_index(['A', 'B'])
 
     def time_groupby_sum_multiindex(self):
         self.df.groupby(level=[0, 1]).sum()
@@ -673,9 +673,9 @@ def setup(self):
         self.secid_min = int('10000000', 16)
         self.secid_max = int('F0000000', 16)
         self.step = ((self.secid_max - self.secid_min) // (self.n_securities - 1))
-        self.security_ids = map((lambda x: hex(x)[2:10].upper()), range(self.secid_min, (self.secid_max + 1), self.step))
+        self.security_ids = map((lambda x: hex(x)[2:10].upper()), list(range(self.secid_min, (self.secid_max + 1), self.step)))
         self.data_index = MultiIndex(levels=[self.dates.values, self.security_ids],
-                                     labels=[[i for i in range(self.n_dates) for _ in range(self.n_securities)], (range(self.n_securities) * self.n_dates)],
+                                     labels=[[i for i in range(self.n_dates) for _ in range(self.n_securities)], (list(range(self.n_securities)) * self.n_dates)],
                                      names=['date', 'security_id'])
         self.n_data = len(self.data_index)
         self.columns = Index(['factor{}'.format(i) for i in range(1, (self.n_columns + 1))])

asv_bench/benchmarks/packers.py

@@ -321,7 +321,9 @@ def remove(self, f):
 class packers_read_sas7bdat(object):
 
     def setup(self):
-        self.f = 'data/test1.sas7bdat'
+        self.f = os.path.join(os.path.dirname(__file__), '..', '..',
+                              'pandas', 'io', 'tests', 'sas', 'data',
+                              'test1.sas7bdat')
 
     def time_packers_read_sas7bdat(self):
         pd.read_sas(self.f, format='sas7bdat')
@@ -330,7 +332,9 @@ def time_packers_read_sas7bdat(self):
 class packers_read_xport(object):
 
     def setup(self):
-        self.f = 'data/paxraw_d_short.xpt'
+        self.f = os.path.join(os.path.dirname(__file__), '..', '..',
+                              'pandas', 'io', 'tests', 'sas', 'data',
+                              'paxraw_d_short.xpt')
 
     def time_packers_read_xport(self):
         pd.read_sas(self.f, format='xport')

asv_bench/benchmarks/plotting.py

@@ -11,17 +11,22 @@ class plot_timeseries_period(object):
     goal_time = 0.2
 
     def setup(self):
+        import matplotlib
+        matplotlib.use('Agg')
         self.N = 2000
         self.M = 5
         self.df = DataFrame(np.random.randn(self.N, self.M), index=date_range('1/1/1975', periods=self.N))
 
     def time_plot_timeseries_period(self):
         self.df.plot()
 
+
 class plot_andrews_curves(object):
     goal_time = 0.6
 
     def setup(self):
+        import matplotlib
+        matplotlib.use('Agg')
         self.N = 500
         self.M = 10
         data_dict = {x: np.random.randn(self.N) for x in range(self.M)}