48 lines
1.5 KiB
Python
48 lines
1.5 KiB
Python
import sys, re
|
|
from pathlib import Path
|
|
from datetime import date
|
|
wiki = Path(sys.argv[1])
|
|
report = Path(sys.argv[2])
|
|
threshold = int(sys.argv[3])
|
|
today = date.today()
|
|
stale, fresh, nodate = [], [], []
|
|
for f in sorted(wiki.rglob('*.md')):
|
|
rel = f.relative_to(wiki)
|
|
if str(rel).startswith('raw/'):
|
|
continue
|
|
content = f.read_text(errors='ignore')
|
|
upd = re.search(r'^updated:\s*(\S+)', content[:800], re.M)
|
|
if not upd:
|
|
nodate.append(str(rel))
|
|
continue
|
|
try:
|
|
d = date.fromisoformat(upd.group(1))
|
|
except ValueError:
|
|
nodate.append(str(rel))
|
|
continue
|
|
age = (today - d).days
|
|
(stale if age > threshold else fresh).append((str(rel), age, upd.group(1)))
|
|
stale.sort(key=lambda x: -x[1])
|
|
lines = [
|
|
f"# Wiki freshness audit — {today.isoformat()}",
|
|
f"Threshold: >{threshold} days since `updated`",
|
|
"",
|
|
f"## STALE ({len(stale)})",
|
|
"",
|
|
]
|
|
for rel, age, d in stale:
|
|
lines.append(f"- [{age:>3}d] {rel} (updated {d})")
|
|
lines += [f"", f"## FRESH ({len(fresh)})", ""]
|
|
for rel, age, d in fresh:
|
|
lines.append(f"- [{age:>3}d] {rel}")
|
|
lines += [f"", f"## NO updated DATE ({len(nodate)})", ""]
|
|
for rel in nodate:
|
|
lines.append(f"- {rel}")
|
|
lines += [
|
|
"",
|
|
f"TOTAL: {len(stale)} stale, {len(fresh)} fresh, {len(nodate)} no-date of {len(stale)+len(fresh)+len(nodate)} pages",
|
|
]
|
|
report.parent.mkdir(parents=True, exist_ok=True)
|
|
report.write_text('\n'.join(lines))
|
|
print(f"[2/3] audit: {len(stale)} stale (>={threshold}d), {len(fresh)} fresh, {len(nodate)} no-date -> {report}")
|