bf-infra/bin/gitea-tp-sync.py

#!/usr/bin/env python3
"""
TP36 (reframed) — teilprojekte.md Index/Status → Gitea Issues (admin/bf-infra).

Idempotenter Sync (get-or-create, Match ueber "TPnn" im Titel):
  - Archiv-Header in docs/teilprojekte-archiv.md  -> closed Issues (nur TPs, die in Gitea
    NICHT offen sind; ein reopentes TP mit Archiv-Zeiger bleibt offen).
  - Labels status:* (Marker) + thema:* (Gruppierung) werden vorab get-or-create.

SSoT-Modell (seit TP36-Cutover 2026-09-07): Gitea fuehrt Index+Status — aktive TPs werden
DIREKT als Gitea-Issue angelegt/gepflegt, NICHT mehr aus einer Datei-Tabelle abgeleitet
(die Quick-Index-Tabelle wurde aus teilprojekte.md entfernt). Dieses Werkzeug haelt nur noch
die ARCHIVIERTEN (closed) TPs mit teilprojekte-archiv.md idempotent abgeglichen und fasst
offene Issues nie an. Die per-TP-Prosa bleibt in docs/teilprojekte.md / teilprojekte-archiv.md.
Token: ~/.hermes/.env  (GITEA_ACCESS_TOKEN, GITEA_HOST) — wird NIE ausgegeben.

Aufruf:
  python3 bin/gitea-tp-sync.py            # Dry-Run (default) — zeigt, was passieren wuerde
  python3 bin/gitea-tp-sync.py --commit   # legt Labels + Issues an / gleicht sie ab
"""
import json
import os
import re
import sys
import urllib.request
import urllib.error
from pathlib import Path

REPO = Path(__file__).resolve().parent.parent
ACTIVE = REPO / "docs" / "teilprojekte.md"
ARCHIV = REPO / "docs" / "teilprojekte-archiv.md"
HERMES = Path.home() / ".hermes" / ".env"
OWNER_REPO = "admin/bf-infra"
COMMIT = "--commit" in sys.argv

# Marker -> (status-label, gitea-state)
MARKER = {
    "🔜": ("status:todo", "open"),
    "⏳": ("status:wip", "open"),
    "⏸️": ("status:paused", "open"),
    "💡": ("status:idea", "open"),
    "🟢": ("status:concept", "open"),
    "✅": ("status:done", "closed"),
    "❌": ("status:dropped", "closed"),
}
LABEL_COLORS = {
    "status:todo": "#ededed", "status:wip": "#fbca04", "status:paused": "#c5def5",
    "status:idea": "#d4c5f9", "status:concept": "#bfdadc", "status:done": "#0e8a16",
    "status:dropped": "#e11d21",
    "thema:firewall-dns": "#b60205", "thema:wlan": "#1d76db", "thema:cloudflare-proxy": "#f9a825",
    "thema:nautobot-doku": "#5319e7", "thema:infra": "#0e8a16", "thema:ipv6-matter": "#006b75",
    "thema:identity": "#e99695", "thema:backlog": "#cccccc",
}
# TP-Nummer -> Themen-Label (aus dem Statusbaum-Gruppierung; nur aktive TPs brauchen ein Board-Thema)
THEME = {
    "5.1": "firewall-dns", "23": "firewall-dns",
    "25": "wlan", "28": "wlan", "13": "wlan", "17.1": "wlan", "17.3": "wlan",
    "16": "cloudflare-proxy", "16.2": "cloudflare-proxy", "16.3": "cloudflare-proxy",
    "31": "cloudflare-proxy", "41": "cloudflare-proxy", "14.4": "cloudflare-proxy",
    "42": "cloudflare-proxy", "43": "cloudflare-proxy", "44": "cloudflare-proxy",
    "32": "nautobot-doku", "34": "nautobot-doku", "35": "nautobot-doku",
    "36": "nautobot-doku", "37": "nautobot-doku",
    "30": "infra", "40": "infra", "20.1": "infra", "18": "infra",
    "24": "infra", "38": "infra", "11": "infra",
    "12": "ipv6-matter", "12.1": "ipv6-matter",
    "8": "identity",
    "29": "backlog", "33": "backlog",
}


def hermes_env():
    tok = host = None
    for line in HERMES.read_text().splitlines():
        m = re.match(r"\s*(GITEA_ACCESS_TOKEN|GITEA_HOST)\s*=\s*(.+?)\s*$", line)
        if not m:
            continue
        if m.group(1) == "GITEA_ACCESS_TOKEN":
            tok = m.group(2).strip().strip('"').strip("'")
        else:
            host = m.group(2).strip().strip('"').strip("'")
    if not tok:
        sys.exit("GITEA_ACCESS_TOKEN nicht in %s gefunden." % HERMES)
    host = (host or "https://gitea.emsenhuber.net").rstrip("/")
    if not host.startswith("http"):
        host = "https://" + host
    return tok, host


TOKEN, HOST = hermes_env()
API = HOST + "/api/v1"


def api(method, path, body=None):
    url = API + path
    data = json.dumps(body).encode() if body is not None else None
    req = urllib.request.Request(url, data=data, method=method, headers={
        "Authorization": "token " + TOKEN,
        "Accept": "application/json", "Content-Type": "application/json"})
    try:
        with urllib.request.urlopen(req, timeout=30) as r:
            return r.status, json.load(r)
    except urllib.error.HTTPError as e:
        return e.code, (json.load(e) if e.headers.get("content-type", "").startswith("application/json") else {})


# --------------------------------------------------------------------------- parse
def parse_archiv(exclude_nums):
    """Archiv-Header -> archivierte TPs. exclude_nums = TP-Nummern, die in Gitea OFFEN sind
    (reopente TPs mit Archiv-Zeiger) -> die bleiben offen, nicht als closed abgleichen."""
    out = []
    hdr = re.compile(r"^#{2,4}\s*(TP[\d.\-A-Za-z]+)\s*—\s*(.+?)\s*▸", )
    seen = set()
    for line in ARCHIV.read_text().splitlines():
        m = hdr.match(line)
        if not m:
            continue
        tp, title = m.groups()
        num = tp[2:]
        if num in exclude_nums or num in seen:
            continue
        seen.add(num)
        short = re.sub(r"\*\*|`", "", title).strip()[:80]
        out.append({"tp": tp, "num": num, "marker": "✅",
                    "title": short, "date": "", "src": "archiv"})
    return out


# --------------------------------------------------------------------------- gitea sync
def ensure_labels(existing):
    by_name = {l["name"]: l for l in existing}
    for name, color in LABEL_COLORS.items():
        if name in by_name:
            continue
        if COMMIT:
            api("POST", f"/repos/{OWNER_REPO}/labels",
                {"name": name, "color": color.lstrip("#")})
        print(f"  [{'+' if COMMIT else 'DRY'}] label {name}")
    st, cur = api("GET", f"/repos/{OWNER_REPO}/labels?limit=100")
    return {l["name"]: l["id"] for l in (cur if isinstance(cur, list) else [])}


def all_issues():
    out, page = [], 1
    while True:
        st, d = api("GET", f"/repos/{OWNER_REPO}/issues?state=all&type=issues&limit=50&page={page}")
        if not isinstance(d, list) or not d:
            break
        out += d
        page += 1
    return out


def main():
    print(f"Gitea {HOST}  repo {OWNER_REPO}  mode={'COMMIT' if COMMIT else 'DRY-RUN'}")

    st, existing_labels = api("GET", f"/repos/{OWNER_REPO}/labels?limit=100")
    label_id = ensure_labels(existing_labels if isinstance(existing_labels, list) else [])

    issues = all_issues()
    by_tp = {}
    open_nums = set()
    for iss in issues:
        # Titel ist "TPxx — …"; volle ID inkl. Buchstaben/Bindestrich (TP14.2a, TP-R) bis zum " — "
        m = re.match(r"^(TP[\w.\-]*?)\s+—", iss["title"])
        if m:
            by_tp[m.group(1)] = iss
            if iss.get("state") == "open":
                open_nums.add(m.group(1)[2:])

    # Aktive TPs sind seit dem TP36-Cutover SSoT direkt in Gitea -> nur das Archiv abgleichen,
    # und offene (reopente) TPs davon ausschliessen, damit sie nicht faelschlich geschlossen werden.
    items = parse_archiv(open_nums)
    print(f"Geparst: {len(items)} archivierte TP  (aktive TPs = SSoT in Gitea: "
          f"{len(open_nums)} offen, nicht angetastet)\n")

    created = updated = 0
    for it in items:
        status_label, state = MARKER.get(it["marker"], ("status:todo", "open"))
        labels = [status_label]
        th = THEME.get(it["num"])
        if th:
            labels.append(f"thema:{th}")
        title = f"{it['tp']} — {it['title']}"
        body = (f"Steuerungs-SSoT: Index/Status hier in Gitea; **Prosa/Detail** im Repo unter "
                f"`docs/{'teilprojekte-archiv.md' if it['src']=='archiv' else 'teilprojekte.md'}` "
                f"(Abschnitt `{it['tp']}`).\n\n"
                f"- Status: {it['marker']} ({status_label})\n"
                + (f"- Abschluss: {it['date']}\n" if it["date"] else ""))
        want_ids = [label_id[l] for l in labels if l in label_id]

        cur = by_tp.get(it["tp"])
        if cur is None:
            print(f"  [{'+' if COMMIT else 'DRY'}] NEW  {title}  {labels} state={state}")
            if COMMIT:
                sc, res = api("POST", f"/repos/{OWNER_REPO}/issues",
                              {"title": title, "body": body, "labels": want_ids})
                if state == "closed" and isinstance(res, dict) and res.get("number"):
                    api("PATCH", f"/repos/{OWNER_REPO}/issues/{res['number']}", {"state": "closed"})
            created += 1
        else:
            have = {l["name"] for l in cur.get("labels", [])}
            drift = (have != set(labels)) or (cur.get("state") != state)
            if drift:
                print(f"  [{'~' if COMMIT else 'DRY'}] UPD  #{cur['number']} {title}  "
                      f"{sorted(have)}->{labels} state {cur.get('state')}->{state}")
                if COMMIT:
                    api("PATCH", f"/repos/{OWNER_REPO}/issues/{cur['number']}",
                        {"labels": want_ids, "state": state})
                updated += 1

    print(f"\n{'Angelegt' if COMMIT else 'Wuerde anlegen'}: {created} · "
          f"{'Abgeglichen' if COMMIT else 'Wuerde abgleichen'}: {updated}")
    if not COMMIT:
        print("Dry-Run — nichts geschrieben. Mit --commit ausfuehren.")


if __name__ == "__main__":
    main()