pydantic/benchmarks/run.py

import json
from operator import itemgetter

import os
import random
import string
import sys
from datetime import datetime

from devtools import debug
from functools import partial
from pathlib import Path
from statistics import StatisticsError, mean
from statistics import stdev as stdev_

from test_pydantic import TestPydantic

try:
    from test_trafaret import TestTrafaret
except Exception:
    print('WARNING: unable to import TestTrafaret')
    TestTrafaret = None

try:
    from test_drf import TestDRF
except Exception:
    print('WARNING: unable to import TestDRF')
    TestDRF = None

try:
    from test_marshmallow import TestMarshmallow
except Exception:
    print('WARNING: unable to import TestMarshmallow')
    TestMarshmallow = None


try:
    from test_valideer import TestValideer
except Exception:
    print('WARNING: unable to import TestValideer')
    TestValideer = None

try:
    from test_cattrs import TestCAttrs
except Exception:
    print('WARNING: unable to import TestCAttrs')
    TestCAttrs = None

try:
    from test_cerberus import TestCerberus
except Exception:
    print('WARNING: unable to import TestCerberus')
    TestCerberus = None

try:
    from test_voluptuous import TestVoluptuous
except Exception as e:
    print('WARNING: unable to import TestVoluptuous')
    TestVoluptuous = None

PUNCTUATION = ' \t\n!"#$%&\'()*+,-./'
LETTERS = string.ascii_letters
UNICODE = '\xa0\xad¡¢£¤¥¦§¨©ª«¬ ®¯°±²³´µ¶·¸¹º»¼½¾¿ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖ×ØÙÚÛÜÝÞßàáâãäåæçèéêëìíîïðñòóôõö÷øùúûüýþÿ'
ALL = PUNCTUATION * 5 + LETTERS * 20 + UNICODE
random = random.SystemRandom()

# in order of performance for csv
other_tests = [TestCAttrs, TestValideer, TestMarshmallow, TestVoluptuous, TestTrafaret, TestDRF, TestCerberus]
active_other_tests = [t for t in other_tests if t is not None]


class GenerateData:
    def __init__(self):
        pass


def rand_string(min_length, max_length, corpus=ALL):
    return ''.join(random.choices(corpus, k=random.randrange(min_length, max_length)))


MISSING = object()


def null_missing_v(f, null_chance=0.2, missing_chance=None):
    r = random.random()
    if random.random() < null_chance:
        return None
    missing_chance = null_chance if missing_chance is None else missing_chance
    if r < (null_chance + missing_chance):
        return MISSING
    return f()


def null_missing_string(*args, **kwargs):
    f = partial(rand_string, *args)
    return null_missing_v(f, **kwargs)


def rand_email():
    if random.random() < 0.2:
        c1, c2 = UNICODE, LETTERS
    else:
        c1, c2 = LETTERS, LETTERS
    return f'{rand_string(10, 50, corpus=c1)}@{rand_string(10, 50, corpus=c2)}.{rand_string(2, 5, corpus=c2)}'


def null_missing_email():
    return null_missing_v(rand_email)


def rand_date():
    r = random.randrange
    return f'{r(1900, 2020)}-{r(0, 12)}-{r(0, 32)}T{r(0, 24)}:{r(0, 60)}:{r(0, 60)}'


def remove_missing(d):
    if isinstance(d, dict):
        return {k: remove_missing(v) for k, v in d.items() if v is not MISSING}
    elif isinstance(d, list):
        return [remove_missing(d_) for d_ in d]
    else:
        return d


def generate_case():
    return remove_missing(dict(
        id=random.randrange(1, 2000),
        client_name=null_missing_string(10, 280, null_chance=0.05, missing_chance=0.05),
        sort_index=random.random() * 200,
        # client_email=null_missing_email(),  # email checks differ with different frameworks
        client_phone=null_missing_string(5, 15),
        location=dict(
            latitude=random.random() * 180 - 90,
            longitude=random.random() * 180,
        ),
        contractor=str(random.randrange(-100, 2000)),
        upstream_http_referrer=null_missing_string(10, 1050),
        grecaptcha_response=null_missing_string(10, 1050, null_chance=0.05, missing_chance=0.05),
        last_updated=rand_date(),
        skills=[dict(
            subject=null_missing_string(5, 20, null_chance=0.01, missing_chance=0),
            subject_id=i,
            category=rand_string(5, 20),
            qual_level=rand_string(5, 20),
            qual_level_id=random.randrange(2000),
            qual_level_ranking=random.random() * 20
        ) for i in range(random.randrange(1, 5))]
    ))

THIS_DIR = Path(__file__).parent.resolve()


def stdev(d):
    try:
        return stdev_(d)
    except StatisticsError:
        return 0

def run_tests(classes, cases, repeats, json=False):
    if json:
        classes = [c for c in classes if hasattr(c, 'to_json')]
    lpad = max(len(t.package) for t in classes) + 4
    print(f'testing {", ".join(t.package for t in classes)}, {repeats} times each')
    results = []
    csv_results = []

    for test_class in classes:
        times = []
        p = test_class.package
        for i in range(repeats):
            count, pass_count = 0, 0
            test = test_class(True)
            models = []
            if json:
                models = [m for passed, m in (test.validate(c) for c in cases) if passed]
            start = datetime.now()
            for j in range(3):
                if json:
                    for model in models:
                        test.to_json(model)
                        pass_count += 1
                        count += 1
                else:
                    for case in cases:
                        passed, result = test.validate(case)
                        pass_count += passed
                        count += 1
            time = (datetime.now() - start).total_seconds()
            success = pass_count / count * 100
            print(f'{p:>{lpad}} ({i+1:>{len(str(repeats))}}/{repeats}) time={time:0.3f}s, success={success:0.2f}%')
            times.append(time)
        print(f'{p:>{lpad}} best={min(times):0.3f}s, avg={mean(times):0.3f}s, stdev={stdev(times):0.3f}s')
        model_count = 3 * len(cases)
        avg = mean(times) / model_count * 1e6
        sd = stdev(times) / model_count * 1e6
        results.append(f'{p:>{lpad}} best={min(times) / model_count * 1e6:0.3f}μs/iter '
                       f'avg={avg:0.3f}μs/iter stdev={sd:0.3f}μs/iter version={test_class.version}')
        csv_results.append([p, test_class.version, avg])
        print()

    return results, csv_results

def main():
    json_path = THIS_DIR / 'cases.json'
    if not json_path.exists():
        print('generating test cases...')
        cases = [generate_case() for _ in range(2000)]
        with json_path.open('w') as f:
            json.dump(cases, f, indent=2, sort_keys=True)
    else:
        with json_path.open() as f:
            cases = json.load(f)

    tests = [TestPydantic]
    if 'pydantic-only' not in sys.argv:
        tests += active_other_tests

    repeats = int(os.getenv('BENCHMARK_REPEATS', '5'))
    test_json = 'TEST_JSON' in os.environ
    results, csv_results = run_tests(tests, cases, repeats, test_json)

    for r in results:
        print(r)

    if 'SAVE' in os.environ:
        save_md(csv_results)


def save_md(data):
    headings = 'Package', 'Version', 'Relative Performance', 'Mean validation time'
    rows = [headings, ['---' for _ in headings]]

    first_avg = None
    for package, version, avg in sorted(data, key=itemgetter(2)):
        if first_avg:
            relative = f'{avg / first_avg:0.1f}x slower'
        else:
            relative = ''
            first_avg = avg
        rows.append([package, f'`{version}`', relative, f'{avg:0.1f}μs'])

    table = '\n'.join(' | '.join(row) for row in rows)
    text = f"""\
[//]: <> (Generated with benchmarks/run.py, DO NOT EDIT THIS FILE DIRECTLY, instead run `SAVE=1 python ./run.py`.)

{table}
"""
    (Path(__file__).parent / '..' / 'docs' / '.benchmarks_table.md').write_text(text)


def diff():
    json_path = THIS_DIR / 'cases.json'
    with json_path.open() as f:
        cases = json.load(f)

    allow_extra = True
    pydantic = TestPydantic(allow_extra)
    others = [t(allow_extra) for t in active_other_tests]

    for case in cases:
        pydantic_passed, pydantic_result = pydantic.validate(case)
        for other in others:
            other_passed, other_result = other.validate(case)
            if other_passed != pydantic_passed:
                print(f'⨯ pydantic {pydantic_passed} != {other.package} {other_passed}')
                debug(case, pydantic_result, other_result)
                return
    print('✓ data passes match for all packages')


if __name__ == '__main__':
    if 'diff' in sys.argv:
        diff()
    else:
        main()

    # if None in other_tests:
    #     print('not all libraries could be imported!')
    #     sys.exit(1)