boc/projects/landvex/arkiv/byggscript/konsolidera_v2.py

#!/usr/bin/env python3
# Landvex IOL — våg 5: konsolidering v2 (Tier 1 + Tier 2), validering, grafexport
import json, re
from pathlib import Path

OUT = Path("/mnt/user-data/outputs")
PREDIKAT = {"monteras_pa","star_pa","ansluts_med","matas_fran","ingar_i","del_av",
            "ersatter","ersatts_av","kompatibel_med","forvaxlas_med","tillverkas_av","foljer_standard","variant_av"}
VERIF = {"obekraftad","kallbelagd","faltverifierad","tillverkarbekraftad"}
DOMANER = {"TRP","ELN","VAT","TEL","BYG","PRK","JVG","HMN","FLG"}
ID_RE = re.compile(r"^LVX-(TRP|ELN|VAT|TEL|BYG|PRK|JVG|HMN|FLG)-[0-9]{4}$|^LVX-P-[0-9]{7}$")

def load(name):
    with open(OUT / name, encoding="utf-8") as f:
        return json.load(f)

poster = {}
for fn in ["lvx-klassposter-master-v1.json","lvx-klassposter-tier2-trp.json","lvx-klassposter-tier2-ovriga.json"]:
    for p in load(fn)["poster"]:
        poster[p["lvx_id"]] = p

def multi_lvx(mal):
    return " / " in mal and all(ID_RE.match(x.strip()) for x in mal.split(" / "))

olosta = {}
for p in poster.values():
    for rel in p.get("relationer", []):
        mal = rel.get("mal","")
        if rel.get("predikat") == "foljer_standard" or multi_lvx(mal) or ID_RE.match(mal):
            continue
        olosta[mal] = olosta.get(mal, 0) + 1

modeller = {}
for fn in ["lvx-produktmodeller-vag2.json","lvx-produktmodeller-vag3.json"]:
    for m in load(fn)["poster"]:
        modeller[m["lvx_id"]] = m

fel = []
def kolla(rec, typ):
    rid = rec.get("lvx_id","?")
    if not ID_RE.match(rid): fel.append(f"{rid}: ogiltigt ID-format")
    if rec.get("doman") not in DOMANER: fel.append(f"{rid}: ogiltig domän")
    namn = rec.get("namn",{})
    if not (namn.get("sv") and namn.get("en")): fel.append(f"{rid}: namn.sv/en saknas")
    if rec.get("posttyp") not in {"objektklass","produktmodell"}: fel.append(f"{rid}: posttyp")
    prov = rec.get("proveniens",{})
    if prov.get("verifieringsniva") not in VERIF: fel.append(f"{rid}: verifieringsnivå")
    if not prov.get("skapad"): fel.append(f"{rid}: proveniens.skapad saknas")
    for pr in rec.get("problem",[]):
        if not pr.get("kalla"): fel.append(f"{rid}: problem utan källa")
    for rel in rec.get("relationer",[]):
        if rel.get("predikat") not in PREDIKAT: fel.append(f"{rid}: okänt predikat {rel.get('predikat')}")
    if typ=="modell" and not rec.get("parent"): fel.append(f"{rid}: parent saknas")
    if prov.get("verifieringsniva")=="kallbelagd" and not prov.get("kallor"):
        fel.append(f"{rid}: källbelagd utan kallor")

for p in poster.values(): kolla(p,"klass")
for m in modeller.values(): kolla(m,"modell")

kanter = []
def kant(s, pred, obj, niva, beskr=""):
    typ = "lvx" if ID_RE.match(obj or "") else ("standard" if pred=="foljer_standard" else "text")
    kanter.append({"subjekt": s, "predikat": pred, "objekt": obj, "objekt_typ": typ,
                   "niva": niva, **({"beskrivning": beskr} if beskr else {})})

def emit(s, pred, mal, niva, beskr=""):
    if multi_lvx(mal):
        for x in mal.split(" / "):
            kant(s, pred, x.strip(), niva, beskr)
    else:
        kant(s, pred, mal, niva, beskr)

for p in poster.values():
    for rel in p.get("relationer", []):
        emit(p["lvx_id"], rel["predikat"], rel["mal"], "klass", rel.get("beskrivning",""))
    for st in p.get("standarder", []):
        if st.get("beteckning"):
            kant(p["lvx_id"], "foljer_standard", st["beteckning"], "klass", st.get("roll",""))

for m in modeller.values():
    kant(m["lvx_id"], "tillhor_klass", m["parent"], "modell")
    hist = m.get("historik", {})
    for e in hist.get("ersatter", []): kant(m["lvx_id"], "ersatter", e, "modell")
    for e in hist.get("ersatts_av", []): kant(m["lvx_id"], "ersatts_av", e, "modell")
    for rel in m.get("relationer", []):
        emit(m["lvx_id"], rel["predikat"], rel["mal"], "modell", rel.get("beskrivning",""))

lvx_k = sum(1 for k in kanter if k["objekt_typ"]=="lvx")
std_k = sum(1 for k in kanter if k["objekt_typ"]=="standard")
verif_count = {}
for p in poster.values():
    v = p["proveniens"]["verifieringsniva"]
    verif_count[v] = verif_count.get(v,0)+1

master = {
  "beskrivning": "Landvex IOL — konsoliderade klassposter v2: Tier 1 + Tier 2 (37 klasser). Ersätter master-v1 och de två tier2-filerna som arbetskopia.",
  "genererad": "2026-07-04",
  "antal_poster": len(poster),
  "verifieringsfordelning": verif_count,
  "poster": [poster[k] for k in sorted(poster)]
}
with open(OUT/"lvx-klassposter-master-v2.json","w",encoding="utf-8") as f:
    json.dump(master,f,ensure_ascii=False,indent=2)

graf = {
  "beskrivning": "Landvex IOL — grafexport v2. Tier 1 + Tier 2-klasser samt produktmodeller. objekt_typ: lvx=upplöst nod, standard=standardbeteckning, text=olöst platshållare (framtida klass).",
  "genererad": "2026-07-04",
  "noder": {"objektklasser": len(poster), "produktmodeller": len(modeller)},
  "statistik": {"kanter_totalt": len(kanter), "upplosta_lvx_kanter": lvx_k, "standardkanter": std_k, "textnoder_olosta": len(olosta)},
  "olosta_platshallare": [{"text": k, "forekomster": v} for k,v in sorted(olosta.items())],
  "kanter": kanter
}
with open(OUT/"lvx-graf-kanter-v2.json","w",encoding="utf-8") as f:
    json.dump(graf,f,ensure_ascii=False,indent=2)

print(f"Klassposter: {len(poster)}  |  fördelning: {verif_count}")
print(f"Produktmodeller: {len(modeller)}")
print(f"Grafkanter: {len(kanter)} totalt | {lvx_k} LVX-upplösta | {std_k} standardkanter")
print(f"Olösta textnoder: {len(olosta)}")
for k,v in sorted(olosta.items()): print(f"  - {k} ({v})")
print(f"Valideringsfel: {len(fel)}")
for e in fel: print(f"  ! {e}")