#!/usr/bin/env python3 import math from numbers import Number from typing import Generic, Iterable, List, Optional, Tuple, TypeVar T = TypeVar("T", bound=Number) 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]]): from math_utils import RunningMedian self.buckets = {} 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 @staticmethod def n_evenly_spaced_buckets( min_bound: T, max_bound: T, n: int, ) -> List[Tuple[T, T]]: ret = [] 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): ret.append((bucket_start, bucket_start + stride)) return ret def _get_bucket(self, item: T) -> Optional[Tuple[T, T]]: for start_end in self.buckets: if start_end[0] <= item < start_end[1]: return start_end return None def add_item(self, item: T) -> bool: bucket = self._get_bucket(item) if bucket is None: return False self.count += 1 self.buckets[bucket] += 1 self.sigma += item self.median.add_number(item) if self.maximum is None or item > self.maximum: self.maximum = item if self.minimum is None or item < self.minimum: self.minimum = item return True def add_items(self, lst: Iterable[T]) -> bool: all_true = True for item in lst: all_true = all_true and self.add_item(item) return all_true 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] # 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 for bucket in sorted(self.buckets, key=lambda x: x[0]): pop = self.buckets[bucket] start = bucket[0] end = bucket[1] bar = bar_graph( (pop / max_population), include_text=False, width=58, 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" ) if start == last_bucket_start: break return txt