1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316 | from __future__ import annotations
import argparse
from pathlib import Path
import sys
from pytest_park.__about__ import __version__
from pytest_park.core import (
attach_profiler_data,
build_method_statistics,
build_overview_statistics,
compare_method_to_all_prior_runs,
compare_runs,
list_methods,
select_candidate_run,
select_latest_and_previous_runs,
select_reference_run,
summarize_groups,
)
from pytest_park.data import BenchmarkLoadError, ProfilerLoadError, load_benchmark_folder, load_profiler_folder
from pytest_park.ui import serve_dashboard
def main(argv: list[str] | None = None) -> int:
parser = _build_parser()
args = parser.parse_args(argv)
if args.command is None:
return _run_interactive(parser)
if args.command == "version":
print(f"pytest-park v{__version__}")
return 0
try:
runs = load_benchmark_folder(
Path(args.benchmark_folder),
original_postfix=args.original_postfix,
reference_postfix=args.reference_postfix,
)
except BenchmarkLoadError as exc:
parser.error(str(exc))
if args.profiler_folder:
try:
profiler_by_run = load_profiler_folder(Path(args.profiler_folder))
except ProfilerLoadError as exc:
parser.error(str(exc))
attach_profiler_data(runs, profiler_by_run)
if args.command == "load":
_cmd_load(runs)
return 0
if args.command == "analyze":
_cmd_analyze(runs, args.grouping, args.distinct_param, args.method)
return 0
if args.command == "compare":
_cmd_compare(runs, args.reference, args.candidate, args.grouping, args.distinct_param, args.method)
return 0
if args.command == "serve":
serve_dashboard(
benchmark_folder=args.benchmark_folder,
reference=args.reference,
group_by=args.grouping,
distinct_params=args.distinct_param,
original_postfix=args.original_postfix,
reference_postfix=args.reference_postfix,
host=args.host,
port=args.port,
)
return 0
parser.error("Unknown command")
return 1
def _build_parser() -> argparse.ArgumentParser:
parser = argparse.ArgumentParser(prog="pytest-park")
subparsers = parser.add_subparsers(dest="command", required=False)
subparsers.add_parser("version")
shared = argparse.ArgumentParser(add_help=False)
shared.add_argument("benchmark_folder", help="Folder containing pytest-benchmark JSON artifacts")
shared.add_argument("--profiler-folder", default=None, help="Optional folder containing profiler JSON artifacts")
shared.add_argument(
"--original-postfix",
default="",
help="Optional postfix used in original/candidate method names (for name normalization)",
)
shared.add_argument(
"--reference-postfix",
default="",
help="Optional postfix used in reference method names (for name normalization)",
)
load_parser = subparsers.add_parser("load", parents=[shared])
load_parser.set_defaults(command="load")
analyze_parser = subparsers.add_parser("analyze", parents=[shared])
analyze_parser.add_argument("--grouping", action="append", default=[], help="Grouping token, repeatable")
analyze_parser.add_argument("--group-by", dest="grouping", action="append", help=argparse.SUPPRESS)
analyze_parser.add_argument(
"--distinct-param", action="append", default=[], help="Distinct parameter key, repeatable"
)
analyze_parser.add_argument("--method", default=None, help="Optional method/benchmark name for details")
analyze_parser.set_defaults(command="analyze")
compare_parser = subparsers.add_parser("compare", parents=[shared])
compare_parser.add_argument("--reference", default=None, help="Reference run_id or tag (defaults to oldest run)")
compare_parser.add_argument("--candidate", default=None, help="Candidate run_id or tag")
compare_parser.add_argument("--grouping", action="append", default=[], help="Grouping token, repeatable")
compare_parser.add_argument("--group-by", dest="grouping", action="append", help=argparse.SUPPRESS)
compare_parser.add_argument(
"--distinct-param", action="append", default=[], help="Distinct parameter key, repeatable"
)
compare_parser.add_argument("--method", default=None, help="Optional method/benchmark name for details")
compare_parser.set_defaults(command="compare")
serve_parser = subparsers.add_parser("serve", parents=[shared])
serve_parser.add_argument("--reference", default=None, help="Default reference run_id or tag")
serve_parser.add_argument("--grouping", action="append", default=[], help="Grouping token, repeatable")
serve_parser.add_argument("--group-by", dest="grouping", action="append", help=argparse.SUPPRESS)
serve_parser.add_argument(
"--distinct-param", action="append", default=[], help="Distinct parameter key, repeatable"
)
serve_parser.add_argument("--host", default="127.0.0.1")
serve_parser.add_argument("--port", type=int, default=8080)
serve_parser.set_defaults(command="serve")
parser.set_defaults(command=None)
return parser
def _cmd_load(runs) -> None:
case_count = sum(len(run.cases) for run in runs)
print(f"Loaded {len(runs)} runs and {case_count} benchmark cases")
for run in runs:
print(f"- {run.run_id} ({len(run.cases)} cases) from {run.source_file}")
def _cmd_analyze(runs, grouping: list[str], distinct_params: list[str], method: str | None) -> None:
reference_run, candidate_run = select_latest_and_previous_runs(runs)
deltas = compare_runs(reference_run, candidate_run, grouping or None, distinct_params or None)
_print_overview(deltas, reference_run.run_id, candidate_run.run_id)
methods_to_inspect = [method] if method else list_methods(runs)
for m in methods_to_inspect:
_print_method_details(runs, reference_run, candidate_run, deltas, m, distinct_params)
def _cmd_compare(
runs,
reference: str | None,
candidate: str | None,
grouping: list[str],
distinct_params: list[str],
method: str | None,
) -> None:
if reference is None and candidate is None:
reference_run, candidate_run = select_latest_and_previous_runs(runs)
elif candidate and reference is None:
candidate_run = select_reference_run(runs, candidate)
reference_run = _select_previous_run(runs, candidate_run)
else:
reference_run = select_reference_run(runs, reference) if reference else select_latest_and_previous_runs(runs)[0]
candidate_run = select_candidate_run(runs, candidate, reference_run)
deltas = compare_runs(reference_run, candidate_run, grouping or None, distinct_params or None)
_print_overview(deltas, reference_run.run_id, candidate_run.run_id)
methods_to_inspect = [method] if method else list_methods(runs)
for m in methods_to_inspect:
_print_method_details(runs, reference_run, candidate_run, deltas, m, distinct_params)
def _print_overview(deltas, reference_id: str, candidate_id: str) -> None:
summaries = summarize_groups(deltas)
overview = build_overview_statistics(deltas)
print(f"Compared run {candidate_id} against reference {reference_id}")
print(
"Accumulated: "
f"count={overview['count']}, avg_delta={overview['avg_delta_pct']:.2f}%, "
f"median_delta={overview['median_delta_pct']:.2f}%, avg_speedup={overview['avg_speedup']:.3f}, "
f"improved={overview['improved']}, regressed={overview['regressed']}, unchanged={overview['unchanged']}"
)
for summary in summaries:
print(
f"- {summary.label}: count={summary.count}, avg_delta={summary.average_delta_pct:.2f}%, "
f"median_delta={summary.median_delta_pct:.2f}%, improved={summary.improvements}, regressed={summary.regressions}"
)
def _print_method_details(runs, reference_run, candidate_run, deltas, method: str, distinct_params: list[str]) -> None:
method_stats = build_method_statistics(deltas, method)
if not method_stats:
print(f"No matching method data found for: {method}")
return
print(
f"Method {method}: count={method_stats['count']}, avg_delta={method_stats['avg_delta_pct']:.2f}%, "
f"median_delta={method_stats['median_delta_pct']:.2f}%, avg_speedup={method_stats['avg_speedup']:.3f}"
)
history = compare_method_to_all_prior_runs(runs, candidate_run, method, distinct_params or None)
for item in history:
print(
f" current={item['candidate_run_id']} vs={item['reference_run_id']} distinct={item['distinct']} "
f"mean={float(item['mean']):.6f} vs_ref={float(item['reference_mean']):.6f} "
f"delta={float(item['delta_pct']):.2f}%"
)
def _select_previous_run(runs, candidate_run):
candidate_index = next((index for index, run in enumerate(runs) if run.run_id == candidate_run.run_id), None)
if candidate_index is None:
raise ValueError(f"Candidate run not found: {candidate_run.run_id}")
if candidate_index == 0:
raise ValueError("No previous reference run available before candidate run")
return runs[candidate_index - 1]
def _run_interactive(parser: argparse.ArgumentParser) -> int:
if not sys.stdin.isatty():
parser.print_help()
print("No command provided. Run in a TTY for interactive mode.")
return 2
print("pytest-park interactive mode")
print("1) analyze")
print("2) compare")
print("3) serve (quick defaults)")
print("4) serve (custom options)")
print("5) load")
print("6) version")
try:
selection = input("Choose command [1-6]: ").strip()
except EOFError:
return 1
command_map = {
"1": "analyze",
"2": "compare",
"3": "serve_quick",
"4": "serve",
"5": "load",
"6": "version",
}
command = command_map.get(selection)
if command is None:
print("Invalid selection")
return 1
if command == "version":
return main(["version"])
try:
benchmark_folder = input("Benchmark folder [./benchmarks]: ").strip() or "./benchmarks"
except EOFError:
return 1
if command == "serve_quick":
return main(["serve", benchmark_folder])
command_args: list[str] = [command, benchmark_folder]
if command in {"analyze", "compare", "serve"}:
group_by = _read_csv_prompt("Group-by tokens (comma separated, optional): ")
for token in group_by:
command_args.extend(["--grouping", token])
distinct = _read_csv_prompt("Distinct params (comma separated, optional): ")
for token in distinct:
command_args.extend(["--distinct-param", token])
if command in {"analyze", "compare"}:
method = _read_optional_prompt("Method (optional): ")
if method:
command_args.extend(["--method", method])
if command == "compare":
reference = _read_optional_prompt("Reference run/tag (optional): ")
if reference:
command_args.extend(["--reference", reference])
candidate = _read_optional_prompt("Candidate run/tag (optional): ")
if candidate:
command_args.extend(["--candidate", candidate])
if command == "serve":
reference = _read_optional_prompt("Reference run/tag (optional): ")
if reference:
command_args.extend(["--reference", reference])
host = _read_optional_prompt("Host [127.0.0.1]: ") or "127.0.0.1"
port = _read_optional_prompt("Port [8080]: ") or "8080"
command_args.extend(["--host", host, "--port", port])
return main(command_args)
def _read_optional_prompt(prompt: str) -> str:
try:
return input(prompt).strip()
except EOFError:
return ""
def _read_csv_prompt(prompt: str) -> list[str]:
value = _read_optional_prompt(prompt)
if not value:
return []
return [token.strip() for token in value.split(",") if token.strip()]
|