Files
DongHyeonkaandClaude Opus 5 b2963105a8 docs(keycloak-session-store): import the session-storage lab as a new project
The keycloak project ended with four open questions that design could not
settle. A two-VM lab was built to answer them by measurement, and this is
that material: 26 experiments, 125 raw command outputs, 22 browser captures.

Follows the import procedure in README.md.

  source/     the originating repository verbatim — 78 documents, 28 SVGs,
              8 manifests, plus .source-revision recording the commit
  final/      the SSOT
    document.md   729 lines written from the 29 experiment documents, not
                  concatenated: what was predicted, what was measured, and
                  where the measurement itself was wrong
    evidence/raw    125 outputs, flattened to <experiment>__<file> because
                    the originals collided (01-baseline.txt appeared three
                    times) and the audit only globs the top level
    evidence/meta   one per raw file; command and exitCode are null and the
                    README says why rather than inventing them
    evidence/browser  22 captures
    assets/       three diagrams through techviz
    .techviz/     their VizSpecs

A separate project rather than an addition to keycloak: the B-layer answers
that project's four questions, but the A, C and D layers are about cluster
failure, SSO and operations, and one document.md should hold one subject.
The four question records there can point here through 관계.

Recorded rather than papered over: only three of the 28 diagrams were
remade. The repository forbids hand-drawn SVG and forbids titles inside the
canvas; all 28 originals carry both, so converting them is redrawing, not
reformatting. They stay in source/ and the gap is written into the document.

verify-pipeline.py passes. audit-records.py reports no issues.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-04 22:51:59 +09:00

156 lines
6.8 KiB
Python
Executable File

#!/usr/bin/env python3
"""기록과 증거가 이 저장소 규칙을 지키는지 전수로 본다.
읽어서 확인할 수 없는 분량이라 기계로 센다. 파서 검사는 check_body.mjs 가 따로 한다.
python3 scripts/audit-records.py [프로젝트 ...]
"""
from __future__ import annotations
import os, re, sys, glob, json, collections
ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
KINDS = {"case": "CASE", "concept": "CONCEPT", "reference": "REFERENCE",
"question": "QUESTION", "decision": "PROJECT_DECISION"}
BODY_KINDS = {"case", "concept"}
PLAIN_FIELDS = {"case": ("문제", "결론", "검증 환경", "재현 조건"),
"reference": ("목적", "규칙", "적용 조건", "예외", "예시"),
"question": ("사실", "가정", "미지수", "제약", "선택지", "다음 검증"),
"decision": ("결정문", "판단 이유", "영향")}
REQUIRED = ("kind", "slug", "title", "topic", "project", "status")
def front_matter(text: str) -> dict:
if not text.startswith("---"):
return {}
end = text.find("\n---", 3)
out = {}
for line in text[3:end].splitlines():
m = re.match(r"^([a-zA-Z_]+):\s*(.*)$", line)
if m:
out[m.group(1)] = m.group(2).strip().strip('"')
return out
def audit_project(project: str) -> dict:
base = os.path.join(ROOT, "docs", project)
studio = os.path.join(base, "tech-log-studio")
ev = os.path.join(base, "final", "evidence")
issues = collections.Counter()
samples = collections.defaultdict(list)
def flag(key: str, detail: str) -> None:
issues[key] += 1
if len(samples[key]) < 3:
samples[key].append(detail)
shown: set[str] = set()
cited: list[str] = []
records = [f for f in sorted(glob.glob(f"{studio}/*/*/*.md"))
if not f.split(os.sep)[-3].startswith("_")]
for f in records:
rel = os.path.relpath(f, ROOT)
parts = f.split(os.sep)
topic_dir, kind_dir = parts[-3], parts[-2]
text = open(f, encoding="utf-8").read()
fm = front_matter(text)
for key in REQUIRED:
if not fm.get(key):
flag(f"칸 없음: {key}", rel)
if kind_dir in KINDS and fm.get("kind") != KINDS[kind_dir]:
flag("kind 와 폴더 불일치", f"{rel} — kind={fm.get('kind')}")
if fm.get("topic") and fm["topic"] != topic_dir:
flag("topic 과 폴더 불일치", f"{rel} — topic={fm.get('topic')}")
has_body = "<!-- body:start -->" in text and "<!-- body:end -->" in text
if kind_dir in BODY_KINDS and not has_body:
flag("본문 마커 없음", rel)
if kind_dir not in BODY_KINDS and has_body:
flag("본문이 없어야 하는 종류에 본문", rel)
# frontmatter 가 가리키는 파일이 실제로 있나
d = os.path.dirname(f)
for m in re.finditer(r"^ file: (\S+)$", text, re.M):
if not os.path.exists(os.path.normpath(os.path.join(d, m.group(1)))):
flag("assets 링크 깨짐", f"{rel}{m.group(1)}")
for m in re.finditer(r"^ - (\.\./\S+)$", text, re.M):
if not os.path.exists(os.path.normpath(os.path.join(d, m.group(1)))):
flag("evidence 링크 깨짐", f"{rel}{m.group(1)}")
# 평문 칸은 백틱·코드펜스가 글자 그대로 보인다
head = text if "<!-- body:start -->" not in text \
else text[:text.index("<!-- body:start -->")]
for field in PLAIN_FIELDS.get(kind_dir, ()):
fm2 = re.search(rf"^## {re.escape(field)}\n(.*?)(?=\n## |\Z)", head, re.M | re.S)
if fm2 and ("`" in fm2.group(1) or "```" in fm2.group(1)):
flag("평문 칸에 마크업", f"{rel}{field}")
# 본문이 부르는 자산이 frontmatter 에 선언돼 있나
declared = set(re.findall(r"^ - key: (\S+)$", text, re.M))
body_keys = set(re.findall(r':::evidence key="([^"]+)"', text))
shown |= {k[:-8] if k.endswith("-diagram") else k for k in body_keys}
# 본문에 싣지 않고 evidence 로만 잇거나 산문에서 이름을 대도 쓰인 것이다
shown |= {os.path.basename(m)[:-4]
for m in re.findall(r"^ - \S+/raw/(\S+\.txt)$", text, re.M)}
cited.append(text)
for key in body_keys:
if key not in declared:
flag("본문 자산이 frontmatter 에 없음", f"{rel}{key}")
# 증거 삼종 — 기록이 화면에 쓰는 증거만 원문·메타·렌더가 다 있어야 한다.
# raw 에만 있는 캡처는 분석 단계 자료다. 실행 메타를 요구하지 않는다.
def stems(sub: str, ext: str) -> set:
return {os.path.splitext(os.path.basename(p))[0]
for p in glob.glob(f"{ev}/{sub}/*{ext}")}
raw, meta, rendered = stems("raw", ".txt"), stems("meta", ".json"), stems("rendered", ".svg")
for s in sorted(rendered - raw):
flag("렌더에 원문 없음", s)
for s in sorted(rendered - meta):
flag("렌더에 실행 메타 없음", s)
blob = "\n".join(cited)
for s in sorted(rendered - shown):
if s not in blob:
flag("아무 기록도 쓰지 않는 렌더", s)
# meta 가 가리키는 파일이 실제로 있나 (final/ 기준 상대경로)
for p in sorted(glob.glob(f"{ev}/meta/*.json")):
try:
m = json.load(open(p, encoding="utf-8"))
except Exception as exc:
flag("meta 파싱 실패", f"{os.path.basename(p)}{exc}")
continue
for field in ("raw", "svg", "rawPath", "presentationPath"):
v = m.get(field)
if v and not os.path.exists(os.path.join(base, "final", v)):
flag("meta 경로 깨짐", f"{os.path.basename(p)}{field}: {v}")
return {"project": project, "records": len(records),
"evidence": {"raw": len(raw), "meta": len(meta), "rendered": len(rendered)},
"issues": issues, "samples": samples}
def main(argv: list[str]) -> int:
projects = argv[1:] or sorted(
os.path.basename(os.path.dirname(p))
for p in glob.glob(os.path.join(ROOT, "docs/*/tech-log-studio")))
total = 0
for project in projects:
r = audit_project(project)
n = sum(r["issues"].values())
total += n
ev = r["evidence"]
print(f"\n{r['project']} — 기록 {r['records']}건 · 원문 {ev['raw']} · 메타 {ev['meta']} · 렌더 {ev['rendered']}")
if not n:
print(" 문제 없음")
continue
for key, count in r["issues"].most_common():
print(f" {count:>5} {key}")
for s in r["samples"][key]:
print(f" {s}")
print(f"\n합계 {total}건")
return 0 if total == 0 else 1
if __name__ == "__main__":
raise SystemExit(main(sys.argv))