X-Git-Url: https://wannabe.guru.org/gitweb/?a=blobdiff_plain;f=histogram.py;h=52a0d1fad558a493c6e303abdd07a6933053a045;hb=532df2c5b57c7517dfb3dddd8c1358fbadf8baf3;hp=4aa47490122481852ae03e264795509a8794e54e;hpb=36fea7f15ed17150691b5b3ead75450e575229ef;p=python_utils.git

diff --git a/histogram.py b/histogram.py
index 4aa4749..52a0d1f 100644
--- a/histogram.py
+++ b/histogram.py
@@ -1,47 +1,69 @@
 #!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+
+# © Copyright 2021-2022, Scott Gasch
+
+"""A text-based simple histogram helper class."""
 
 import math
-from numbers import Number
-from typing import Generic, Iterable, List, Optional, Tuple, TypeVar
+from dataclasses import dataclass
+from typing import Dict, Generic, Iterable, List, Optional, Tuple, TypeVar
+
+T = TypeVar("T", int, float)
+Bound = int
+Count = int
+
+
+@dataclass
+class BucketDetails:
+    """A collection of details about the internal histogram buckets."""
 
-T = TypeVar("T", bound=Number)
+    num_populated_buckets: int = 0
+    max_population: Optional[int] = None
+    last_bucket_start: Optional[int] = None
+    lowest_start: Optional[int] = None
+    highest_end: Optional[int] = None
+    max_label_width: Optional[int] = None
 
 
 class SimpleHistogram(Generic[T]):
+    """A simple histogram."""
 
     # 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
+    def __init__(self, buckets: List[Tuple[Bound, Bound]]):
+        from math_utils import NumericPopulation
 
-        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.stats: NumericPopulation = NumericPopulation()
+        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
@@ -54,7 +76,7 @@ class SimpleHistogram(Generic[T]):
         self.count += 1
         self.buckets[bucket] += 1
         self.sigma += item
-        self.median.add_number(item)
+        self.stats.add_number(item)
         if self.maximum is None or item > self.maximum:
             self.maximum = item
         if self.minimum is None or item < self.minimum:
@@ -67,38 +89,70 @@ class SimpleHistogram(Generic[T]):
             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]
+    def get_bucket_details(self, label_formatter: str) -> BucketDetails:
+        details = BucketDetails()
+        for (start, end), pop in sorted(self.buckets.items(), key=lambda x: x[0]):
             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
+                details.num_populated_buckets += 1
+                details.last_bucket_start = start
+                if details.max_population is None or pop > details.max_population:
+                    details.max_population = pop
+                if details.lowest_start is None or start < details.lowest_start:
+                    details.lowest_start = start
+                if details.highest_end is None or end > details.highest_end:
+                    details.highest_end = end
+                label = f'[{label_formatter}..{label_formatter}): ' % (start, end)
+                label_width = len(label)
+                if details.max_label_width is None or label_width > details.max_label_width:
+                    details.max_label_width = label_width
+        return details
 
+    def __repr__(self, *, width: int = 80, label_formatter: str = '%d') -> str:
+        from text_utils import bar_graph
+
+        details = self.get_bucket_details(label_formatter)
         txt = ""
-        if max_population is None:
+        if details.num_populated_buckets == 0:
             return txt
+        assert details.max_label_width is not None
+        assert details.lowest_start is not None
+        assert details.highest_end is not None
+        assert details.max_population is not None
+        sigma_label = f'[{label_formatter}..{label_formatter}): ' % (
+            details.lowest_start,
+            details.highest_end,
+        )
+        if len(sigma_label) > details.max_label_width:
+            details.max_label_width = len(sigma_label)
+        bar_width = width - (details.max_label_width + 17)
 
-        for bucket in sorted(self.buckets, key=lambda x: x[0]):
-            pop = self.buckets[bucket]
-            start = bucket[0]
-            end = bucket[1]
+        for (start, end), pop in sorted(self.buckets.items(), key=lambda x: x[0]):
+            if start < details.lowest_start:
+                continue
+            label = f'[{label_formatter}..{label_formatter}): ' % (start, end)
             bar = bar_graph(
-                (pop / max_population),
+                (pop / details.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"
-            )
-            if start == last_bucket_start:
+            txt += label.rjust(details.max_label_width)
+            txt += bar
+            txt += f"({pop/self.count*100.0:5.2f}% n={pop})\n"
+            if start == details.last_bucket_start:
                 break
+        txt += '-' * width + '\n'
+        txt += sigma_label.rjust(details.max_label_width)
+        txt += ' ' * (bar_width - 2)
+        txt += f'     pop(Σn)={self.count}\n'
+        txt += ' ' * (bar_width + details.max_label_width - 2)
+        txt += f'     mean(x̄)={self.stats.get_mean():.3f}\n'
+        txt += ' ' * (bar_width + details.max_label_width - 2)
+        txt += f' median(p50)={self.stats.get_median():.3f}\n'
+        txt += ' ' * (bar_width + details.max_label_width - 2)
+        txt += f'    mode(Mo)={self.stats.get_mode()[0]:.3f}\n'
+        txt += ' ' * (bar_width + details.max_label_width - 2)
+        txt += f'    stdev(σ)={self.stats.get_stdev():.3f}\n'
+        txt += '\n'
         return txt