# -*- coding: utf-8 -*-
"""Map each benchmark run to the commits that preceded it.

Turns "v6 regressed" into "v6 regressed, and here is what changed before it".
Read-only against the windows-agent repo.
"""
import datetime
import io
import json
import os
import re
import subprocess

REPO = r'C:\xampp\htdocs\windows-agent'
SRC = os.path.join(REPO, 'bench', 'results')

out = subprocess.run(['git', '-C', REPO, 'log', '--date=iso-strict',
                      '--pretty=%H\t%ad\t%s'], capture_output=True)
commits = []
for line in out.stdout.decode('utf-8', 'replace').splitlines():
    if not line.strip():
        continue
    h, d, s = line.split('\t', 2)
    commits.append((datetime.datetime.fromisoformat(d).replace(tzinfo=None), h[:7], s))
commits.sort(key=lambda c: c[0])

runs = []
for fn in sorted(os.listdir(SRC)):
    if not fn.endswith('.json'):
        continue
    d = json.load(io.open(os.path.join(SRC, fn), encoding='utf-8'))
    m = re.match(r'(\d{4}-\d{2}-\d{2})_(\d{2})(\d{2})(\d{2})_(.+)\.json', fn)
    start = datetime.datetime.fromisoformat(
        '%sT%s:%s:%s' % (m.group(1), m.group(2), m.group(3), m.group(4)))
    fin = datetime.datetime.fromisoformat(d['finished_at'])
    v = int(re.search(r'v(\d+)', m.group(5)).group(1))
    rs = d['results']
    execs = sum(len(t['runs']) if isinstance(t.get('runs'), list) else 1 for t in rs)
    runs.append(dict(v=v, label=m.group(5), start=start, fin=fin, overall=d['overall'],
                     cost=d['estimated_cost_usd'], execs=execs))
runs.sort(key=lambda r: r['start'])

print('=' * 100)
print('WHAT CHANGED BEFORE EACH BENCHMARK RUN')
print('=' * 100)
prev_end = datetime.datetime(2000, 1, 1)
for i, r in enumerate(runs):
    between = [c for c in commits if prev_end < c[0] <= r['start']]
    delta = ''
    if i:
        delta = '  (%+.1f vs v%d)' % (r['overall'] - runs[i - 1]['overall'], runs[i - 1]['v'])
    print('\nv%-3d %-22s %s  score %.1f  $%.5f/exec%s'
          % (r['v'], r['label'], r['start'].strftime('%m-%d %H:%M'), r['overall'],
             r['cost'] / r['execs'], delta))
    if between:
        for t, h, s in between:
            print('       %s  %s  %s' % (t.strftime('%m-%d %H:%M'), h, s[:78]))
    else:
        print('       (no commits between the previous run and this one '
              '-- run made from uncommitted working-tree changes)')
    prev_end = r['start']

after = [c for c in commits if c[0] > runs[-1]['start']]
if after:
    print('\nafter the last run:')
    for t, h, s in after:
        print('       %s  %s  %s' % (t.strftime('%m-%d %H:%M'), h, s[:78]))