"""Audit docs repo git history: .bru counts per commit + did the 6 missing files ever exist?"""
import json
import subprocess
import sys

if sys.stdout.encoding.lower() != "utf-8":
    sys.stdout.reconfigure(encoding="utf-8")

REPO = r"E:\code\crm-api-docs"

def git(*args):
    r = subprocess.run(["git", "-C", REPO, "-c", "core.quotepath=false"] + list(args),
                       capture_output=True)
    if r.returncode != 0:
        raise RuntimeError(r.stderr.decode("utf-8", "replace"))
    return r.stdout.decode("utf-8", "replace")

commits = [line.split("|")[0] for line in git("log", "--format=%h|%ci|%s|%d").splitlines()]
print(f"commits on HEAD branch: {len(commits)}")

target_frag = ("战略协议", "batch-update")
for c in commits:
    files = git("ls-tree", "-r", c, "--name-only").splitlines()
    bru = [f for f in files if f.endswith(".bru")]
    hit = [f for f in bru if any(t in f for t in target_frag)]
    print(f"{c}: {len(bru)} .bru, agreement/batch-update hits: {len(hit)}")
    for h in hit:
        print(f"    {h}")

print("--- fsck dangling/unreachable (lost commits check) ---")
fsck = subprocess.run(["git", "-C", REPO, "fsck", "--lost-found"], capture_output=True)
out = fsck.stdout.decode("utf-8", "replace") + fsck.stderr.decode("utf-8", "replace")
print(out[:2000] if out.strip() else "(no output)")

print("--- all refs ---")
print(git("for-each-ref", "--format=%(refname:short) %(objectname:short)").strip() or "(none)")