stats: Improve JSON output

- For the histogram, dump objects rather than lists, in order to help
  explain what the values represent.
- For the count of allocations by allocator, dump an object mapping
  allocator name to count.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>

src/memray/reporters/stats.py

@@ -5,6 +5,7 @@
 from collections import Counter
 from dataclasses import asdict
 from pathlib import Path
+from typing import Any
 from typing import Dict
 from typing import Iterator
 from typing import List
@@ -37,6 +38,19 @@ def get_histogram_databins(data: Dict[int, int], bins: int) -> List[Tuple[int, i
     return [(steps[b], dist[b]) for b in range(bins)]
 
 
+def describe_histogram_databins(
+    databins: List[Tuple[int, int]]
+) -> List[Dict[str, int]]:
+    ret: List[Dict[str, int]] = []
+    start = 0
+    for i, (end, count) in enumerate(databins):
+        # The max size for the last bucket is inclusive, not exclusive
+        adjustment = 1 if i != len(databins) - 1 else 0
+        ret.append(dict(min_bytes=start, max_bytes=end - adjustment, count=count))
+        start = end
+    return ret
+
+
 def draw_histogram(
     data: Dict[int, int], bins: int, *, hist_scale_factor: int = 25
 ) -> str:
@@ -149,23 +163,25 @@ def _render_to_terminal(self, histogram_params: Dict[str, int]) -> None:
             print(f"\t- {self._format_location(location)} -> {count}")
 
     def _render_to_json(self, histogram_params: Dict[str, int], out_path: Path) -> None:
-        alloc_size_hist = get_histogram_databins(
-            self._stats.allocation_count_by_size, bins=histogram_params["num_bins"]
+        alloc_size_hist = describe_histogram_databins(
+            get_histogram_databins(
+                self._stats.allocation_count_by_size, bins=histogram_params["num_bins"]
+            )
         )
 
         metadata = asdict(self._stats.metadata)
         for name, val in metadata.items():
             if isinstance(val, datetime.datetime):
                 metadata[name] = str(val)
 
-        data = dict(
+        data: Dict[str, Any] = dict(
             total_num_allocations=self._stats.total_num_allocations,
             total_bytes_allocated=self._stats.total_memory_allocated,
             allocation_size_histogram=alloc_size_hist,
-            allocator_type_distribution=[
-                (allocation_type, count)
+            allocator_type_distribution={
+                allocation_type: count
                 for allocation_type, count in self._get_allocator_type_distribution()
-            ],
+            },
             top_allocations_by_size=[
                 {"location": self._format_location(location), "size": size}
                 for location, size in self._get_top_allocations_by_size()

tests/unit/test_stats_reporter.py

@@ -398,23 +398,23 @@ def test_stats_output_json(fake_stats, tmp_path):
         "total_num_allocations": 20,
         "total_bytes_allocated": 3341500,
         "allocation_size_histogram": [
-            (4739, 1),
-            (8986, 0),
-            (17036, 3),
-            (32299, 0),
-            (61237, 1),
-            (116099, 1),
-            (220113, 6),
-            (417312, 0),
-            (791181, 0),
-            (1500000, 2),
-        ],
-        "allocator_type_distribution": [
-            ("MALLOC", 1013),
-            ("REALLOC", 797),
-            ("CALLOC", 152),
-            ("MMAP", 4),
+            {"min_bytes": 0, "max_bytes": 4738, "count": 1},
+            {"min_bytes": 4739, "max_bytes": 8985, "count": 0},
+            {"min_bytes": 8986, "max_bytes": 17035, "count": 3},
+            {"min_bytes": 17036, "max_bytes": 32298, "count": 0},
+            {"min_bytes": 32299, "max_bytes": 61236, "count": 1},
+            {"min_bytes": 61237, "max_bytes": 116098, "count": 1},
+            {"min_bytes": 116099, "max_bytes": 220112, "count": 6},
+            {"min_bytes": 220113, "max_bytes": 417311, "count": 0},
+            {"min_bytes": 417312, "max_bytes": 791180, "count": 0},
+            {"min_bytes": 791181, "max_bytes": 1500000, "count": 2},
         ],
+        "allocator_type_distribution": {
+            "MALLOC": 1013,
+            "REALLOC": 797,
+            "CALLOC": 152,
+            "MMAP": 4,
+        },
         "top_allocations_by_size": [
             {"location": "fake_func:fake.py:5", "size": 5242880},
             {"location": "fake_func2:fake.py:10", "size": 3072},