Various changes.

[python_utils.git] / histogram.py

diff --git a/histogram.py b/histogram.py
index ca67839da4cb7a7031df835690c30dc9dde15c73..3391b0b3c51d33164b6d173d59a7b4a1b3c9788b 100644 (file)
--- a/histogram.py
+++ b/histogram.py
@@ -4,10 +4,6 @@ import math
 from numbers import Number
 from typing import Generic, Iterable, List, Optional, Tuple, TypeVar
 
-from math_utils import RunningMedian
-from text_utils import bar_graph
-
-
 T = TypeVar("T", bound=Number)
 
 
@@ -18,6 +14,7 @@ class SimpleHistogram(Generic[T]):
     NEGATIVE_INFINITY = -math.inf
 
     def __init__(self, buckets: List[Tuple[T, T]]):
+        from math_utils import RunningMedian
         self.buckets = {}
         for start_end in buckets:
             if self._get_bucket(start_end[0]) is not None:
@@ -69,14 +66,18 @@ class SimpleHistogram(Generic[T]):
             all_true = all_true and self.add_item(item)
         return all_true
 
-    def __repr__(self) -> str:
+    def __repr__(self,
+                 label_formatter='%10s') -> str:
+        from text_utils import bar_graph
+
         max_population: Optional[int] = None
         for bucket in self.buckets:
             pop = self.buckets[bucket]
             if pop > 0:
-                last_bucket_start = bucket[0]
+                last_bucket_start = bucket[0]  # beginning of range
             if max_population is None or pop > max_population:
-                max_population = pop
+                max_population = pop  # bucket with max items
+
         txt = ""
         if max_population is None:
             return txt
@@ -88,17 +89,11 @@ class SimpleHistogram(Generic[T]):
             bar = bar_graph(
                 (pop / max_population),
                 include_text = False,
-                width = 70,
+                width = 58,
                 left_end = "",
                 right_end = "")
-            label = f'{start}..{end}'
-            txt += f'{label:12}: ' + bar + f"({pop}) ({len(bar)})\n"
+            label = f'{label_formatter}..{label_formatter}' % (start, end)
+            txt += f'{label:20}: ' + bar + f"({pop/self.count*100.0:5.2f}% n={pop})\n"
             if start == last_bucket_start:
                 break
-
-        txt = txt + f'''{self.count} item(s)
-{self.maximum} max
-{self.minimum} min
-{self.sigma/self.count:.3f} mean
-{self.median.get_median()} median'''
         return txt