2 # -*- coding: utf-8 -*-
5 from numbers import Number
6 from typing import Dict, Generic, Iterable, List, Optional, Tuple, TypeVar
8 T = TypeVar("T", int, float)
13 class SimpleHistogram(Generic[T]):
14 # Useful in defining wide open bottom/top bucket bounds:
15 POSITIVE_INFINITY = math.inf
16 NEGATIVE_INFINITY = -math.inf
18 def __init__(self, buckets: List[Tuple[Bound, Bound]]):
19 from math_utils import RunningMedian
21 self.buckets: Dict[Tuple[Bound, Bound], Count] = {}
22 for start_end in buckets:
23 if self._get_bucket(start_end[0]) is not None:
24 raise Exception("Buckets overlap?!")
25 self.buckets[start_end] = 0
26 self.sigma: float = 0.0
27 self.median: RunningMedian = RunningMedian()
28 self.maximum: Optional[T] = None
29 self.minimum: Optional[T] = None
33 def n_evenly_spaced_buckets(
37 ) -> List[Tuple[int, int]]:
38 ret: List[Tuple[int, int]] = []
39 stride = int((max_bound - min_bound) / n)
41 raise Exception("Min must be < Max")
42 imax = math.ceil(max_bound)
43 imin = math.floor(min_bound)
44 for bucket_start in range(imin, imax, stride):
45 ret.append((bucket_start, bucket_start + stride))
48 def _get_bucket(self, item: T) -> Optional[Tuple[int, int]]:
49 for start_end in self.buckets:
50 if start_end[0] <= item < start_end[1]:
54 def add_item(self, item: T) -> bool:
55 bucket = self._get_bucket(item)
59 self.buckets[bucket] += 1
61 self.median.add_number(item)
62 if self.maximum is None or item > self.maximum:
64 if self.minimum is None or item < self.minimum:
68 def add_items(self, lst: Iterable[T]) -> bool:
71 all_true = all_true and self.add_item(item)
74 def __repr__(self, *, width: int = 80, label_formatter: str = '%d') -> str:
75 from text_utils import bar_graph
78 max_population: Optional[int] = None
79 for bucket in self.buckets:
80 pop = self.buckets[bucket]
82 last_bucket_start = bucket[0] # beginning of range
83 if max_population is None or pop > max_population:
84 max_population = pop # bucket with max items
85 if max_population is None:
88 max_label_width: Optional[int] = None
89 lowest_start: Optional[int] = None
90 highest_end: Optional[int] = None
91 for bucket in sorted(self.buckets, key=lambda x: x[0]):
93 if lowest_start is None:
96 if highest_end is None or end > highest_end:
98 label = f'[{label_formatter}..{label_formatter}): ' % (start, end)
99 label_width = len(label)
100 if max_label_width is None or label_width > max_label_width:
101 max_label_width = label_width
102 if start == last_bucket_start:
104 assert max_label_width is not None
105 assert lowest_start is not None
106 assert highest_end is not None
108 sigma_label = f'[{label_formatter}..{label_formatter}): ' % (
112 if len(sigma_label) > max_label_width:
113 max_label_width = len(sigma_label)
114 bar_width = width - (max_label_width + 16)
116 for bucket in sorted(self.buckets, key=lambda x: x[0]):
119 label = f'[{label_formatter}..{label_formatter}): ' % (start, end)
120 pop = self.buckets[bucket]
122 (pop / max_population),
128 txt += label.rjust(max_label_width)
130 txt += f"({pop/self.count*100.0:5.2f}% n={pop})\n"
131 if start == last_bucket_start:
133 txt += '-' * width + '\n'
134 txt += sigma_label.rjust(max_label_width)
135 txt += ' ' * (bar_width - 2)
136 txt += f'Σ=(100.00% n={self.count})\n'