Used isort to sort imports. Also added to the git pre-commit hook.

[python_utils.git] / histogram.py

diff --git a/histogram.py b/histogram.py
index a899fe9c60cde770eda49dcdd1dade1790e9ec08..d45e93f328185f869c2b3af4cc832db280d14428 100644 (file)
--- a/histogram.py
+++ b/histogram.py
@@ -2,46 +2,49 @@
 
 import math
 from numbers import Number
-from typing import Generic, Iterable, List, Optional, Tuple, TypeVar
+from typing import Dict, Generic, Iterable, List, Optional, Tuple, TypeVar
 
-T = TypeVar("T", bound=Number)
+T = TypeVar("T", int, float)
+Bound = int
+Count = int
 
 
 class SimpleHistogram(Generic[T]):
-
     # Useful in defining wide open bottom/top bucket bounds:
     POSITIVE_INFINITY = math.inf
     NEGATIVE_INFINITY = -math.inf
 
-    def __init__(self, buckets: List[Tuple[T, T]]):
+    def __init__(self, buckets: List[Tuple[Bound, Bound]]):
         from math_utils import RunningMedian
 
-        self.buckets = {}
+        self.buckets: Dict[Tuple[Bound, Bound], Count] = {}
         for start_end in buckets:
             if self._get_bucket(start_end[0]) is not None:
                 raise Exception("Buckets overlap?!")
             self.buckets[start_end] = 0
-        self.sigma = 0
-        self.median = RunningMedian()
-        self.maximum = None
-        self.minimum = None
-        self.count = 0
+        self.sigma: float = 0.0
+        self.median: RunningMedian = RunningMedian()
+        self.maximum: Optional[T] = None
+        self.minimum: Optional[T] = None
+        self.count: Count = 0
 
     @staticmethod
     def n_evenly_spaced_buckets(
         min_bound: T,
         max_bound: T,
         n: int,
-    ) -> List[Tuple[T, T]]:
-        ret = []
+    ) -> List[Tuple[int, int]]:
+        ret: List[Tuple[int, int]] = []
         stride = int((max_bound - min_bound) / n)
         if stride <= 0:
             raise Exception("Min must be < Max")
-        for bucket_start in range(min_bound, max_bound, stride):
+        imax = math.ceil(max_bound)
+        imin = math.floor(min_bound)
+        for bucket_start in range(imin, imax, stride):
             ret.append((bucket_start, bucket_start + stride))
         return ret
 
-    def _get_bucket(self, item: T) -> Optional[Tuple[T, T]]:
+    def _get_bucket(self, item: T) -> Optional[Tuple[int, int]]:
         for start_end in self.buckets:
             if start_end[0] <= item < start_end[1]:
                 return start_end
@@ -67,9 +70,10 @@ class SimpleHistogram(Generic[T]):
             all_true = all_true and self.add_item(item)
         return all_true
 
-    def __repr__(self, label_formatter='%10s') -> str:
+    def __repr__(self, *, width: int = 80, label_formatter: str = '%d') -> str:
         from text_utils import bar_graph
 
+        txt = ""
         max_population: Optional[int] = None
         for bucket in self.buckets:
             pop = self.buckets[bucket]
@@ -77,24 +81,56 @@ class SimpleHistogram(Generic[T]):
                 last_bucket_start = bucket[0]  # beginning of range
             if max_population is None or pop > max_population:
                 max_population = pop  # bucket with max items
-
-        txt = ""
         if max_population is None:
             return txt
 
+        max_label_width: Optional[int] = None
+        lowest_start: Optional[int] = None
+        highest_end: Optional[int] = None
         for bucket in sorted(self.buckets, key=lambda x: x[0]):
-            pop = self.buckets[bucket]
             start = bucket[0]
+            if lowest_start is None:
+                lowest_start = start
             end = bucket[1]
+            if highest_end is None or end > highest_end:
+                highest_end = end
+            label = f'[{label_formatter}..{label_formatter}): ' % (start, end)
+            label_width = len(label)
+            if max_label_width is None or label_width > max_label_width:
+                max_label_width = label_width
+            if start == last_bucket_start:
+                break
+        assert max_label_width
+        assert lowest_start
+        assert highest_end
+
+        sigma_label = f'[{label_formatter}..{label_formatter}): ' % (
+            lowest_start,
+            highest_end,
+        )
+        if len(sigma_label) > max_label_width:
+            max_label_width = len(sigma_label)
+        bar_width = width - (max_label_width + 16)
+
+        for bucket in sorted(self.buckets, key=lambda x: x[0]):
+            start = bucket[0]
+            end = bucket[1]
+            label = f'[{label_formatter}..{label_formatter}): ' % (start, end)
+            pop = self.buckets[bucket]
             bar = bar_graph(
                 (pop / max_population),
                 include_text=False,
-                width=58,
+                width=bar_width,
                 left_end="",
                 right_end="",
             )
-            label = f'{label_formatter}..{label_formatter}' % (start, end)
-            txt += f'{label:20}: ' + bar + f"({pop/self.count*100.0:5.2f}% n={pop})\n"
+            txt += label.rjust(max_label_width)
+            txt += bar
+            txt += f"({pop/self.count*100.0:5.2f}% n={pop})\n"
             if start == last_bucket_start:
                 break
+        txt += '-' * width + '\n'
+        txt += sigma_label.rjust(max_label_width)
+        txt += ' ' * (bar_width - 2)
+        txt += f'Σ=(100.00% n={self.count})\n'
         return txt