FIX/#199

opteryx/engine/planner/operations/show_columns.py

@@ -32,6 +32,7 @@
 
 MAX_COLLECTOR: int = 17
 
+
 def myhash(any):
     from cityhash import CityHash64
 
@@ -205,21 +206,34 @@ def _extended_collector(pages):
                 [i for i in column_data if i not in (None, numpy.nan)]
             )
 
-            if _type in (OPTERYX_TYPES.VARCHAR, OPTERYX_TYPES.NUMERIC, OPTERYX_TYPES.TIMESTAMP):                
+            if _type in (
+                OPTERYX_TYPES.VARCHAR,
+                OPTERYX_TYPES.NUMERIC,
+                OPTERYX_TYPES.TIMESTAMP,
+            ):
                 # hyperloglog estimates cardinality/uniqueness
                 hll = profile.get("hyperloglog")
                 if hll is None:
                     hll = hyperloglog.HyperLogLogPlusPlus(p=16)
                 [hll.update(value) for value in column_data]
                 profile["hyperloglog"] = hll
-
-            if _type in (OPTERYX_TYPES.BOOLEAN, OPTERYX_TYPES.VARCHAR, OPTERYX_TYPES.NUMERIC, OPTERYX_TYPES.TIMESTAMP):
+
+            if _type in (
+                OPTERYX_TYPES.BOOLEAN,
+                OPTERYX_TYPES.VARCHAR,
+                OPTERYX_TYPES.NUMERIC,
+                OPTERYX_TYPES.TIMESTAMP,
+            ):
                 # counter is used to collect and count unique values
                 counter = profile.get("counter")
                 if counter is None:
                     counter = CountingTree()
                 if len(counter) < MAX_COLLECTOR:
-                    [counter.insert(value) for value in column_data if len(counter) < MAX_COLLECTOR]
+                    [
+                        counter.insert(value)
+                        for value in column_data
+                        if len(counter) < MAX_COLLECTOR
+                    ]
                 profile["counter"] = counter
 
             if _type in (OPTERYX_TYPES.NUMERIC, OPTERYX_TYPES.TIMESTAMP):

opteryx/sketches/__init__.py

@@ -1,3 +1,3 @@
 import pyximport
 
-pyximport.install()
+pyximport.install()

opteryx/third_party/abctree/__init__.py

@@ -2,4 +2,4 @@
 
 pyximport.install()
 
-from .abctree import ABCTree
+from .abctree import ABCTree

opteryx/third_party/hyperloglog/__init__.py

@@ -1,2 +1,2 @@
 from .hyperloglog import HyperLogLog
-from .hyperloglog import HyperLogLogPlusPlus
+from .hyperloglog import HyperLogLogPlusPlus

opteryx/third_party/hyperloglog/hyperloglog.py

@@ -283,7 +283,8 @@ def deserialize(cls, buf):
             )
         except TypeError:
             h.reg = numpy.array(
-                struct.unpack_from("%dB" % h.m, bytearray(buf), offset), dtype=numpy.int8
+                struct.unpack_from("%dB" % h.m, bytearray(buf), offset),
+                dtype=numpy.int8,
             )
         return h
 
@@ -306,7 +307,8 @@ def __setstate__(self, buf):
             )
         except TypeError:
             self.reg = numpy.array(
-                struct.unpack_from("%dB" % self.m, bytearray(buf), offset), dtype=numpy.int8
+                struct.unpack_from("%dB" % self.m, bytearray(buf), offset),
+                dtype=numpy.int8,
             )
 
 
@@ -338,9 +340,7 @@ class HyperLogLogPlusPlus(HyperLogLog):
     _hash_range_byte = 8
 
     def __init__(self, p=8, reg=None, hashfunc=CityHash64):
-        super(HyperLogLogPlusPlus, self).__init__(
-            p=p, reg=reg, hashfunc=hashfunc
-        )
+        super(HyperLogLogPlusPlus, self).__init__(p=p, reg=reg, hashfunc=hashfunc)
 
     def _get_threshold(self, p):
         return _thresholds[p - 4]

opteryx/third_party/hyperloglog/tests/test_hll.py

@@ -6,6 +6,7 @@
 
 import sys
 import os
+
 sys.path.insert(1, os.path.join(sys.path[0], ".."))
 from hyperloglog import HyperLogLog, HyperLogLogPlusPlus
 
@@ -126,9 +127,6 @@ def test_copy(self):
         self.assertEqual(h1.hashfunc, h2.hashfunc)
 
 
-
-
-
 class TestHyperLogLogPlusPlus(TestHyperLogLog):
 
     _class = HyperLogLogPlusPlus

tests/sketches/test_counting_tree.py

@@ -5,6 +5,7 @@
 
 from opteryx.sketches.counting_tree import CountingTree
 
+
 def test_counter():
     pass
 
@@ -13,7 +14,7 @@ def test_counter():
     ct = CountingTree()
     for i in range(20):
         for u in range(i):
-            #ct.insert(i)
+            # ct.insert(i)
             ct.insert(u)
 
-    print(ct)
+    print(ct)