# ruff: noqa
# mypy: ignore-errors
"""Cross-Domain MCAP Convergence Analysis — where does the knowledge model converge? (Phase Ω, pause)

After Automotive the user paused on adding domains to ask a deeper question. NOT "which MCAPs occur most
often?" (frequency deceives — a generic `document_changes` may be in 40 sources but is not the product
core) but "which MCAPs CARRY the largest part of the system?". The answer is a deterministic MCAP IMPACT
SCORE (no AI), an internal engineering tool, computed by aggregating over the EXISTING data only.

  Impact(MCAP) = distinct Requirement Sources + distinct Target Types + distinct Domains
               + distinct Journeys + Regulatory Leverage + Business Leverage

Four reports, all pure aggregation (no new runtime, no new architecture):
  1. Core      — highest impact (the cross-cutting nodes that carry the system)
  2. Emerging  — span >= 2 domains (bridges across requirement worlds)
  3. Isolated  — only one source/journey/domain (specialised, OR convergence not yet recognised)
  4. Suspicious— probably cut too coarse (generic) or too fine (one hyper-specific occurrence)

Non-runtime -> no deploy.
Run:  cd backend-compliance && PYTHONPATH=. python3 reference_scenarios/mcap_convergence_analysis.py
"""
from __future__ import annotations

import os
import yaml

OUT = []


def w(s=""):
    OUT.append(s)


_HERE = os.path.dirname(__file__)
_TP = os.path.join(_HERE, "..", "knowledge", "transition_patterns")

# pattern -> (domain, default target_type, default sources, source_type-of-default)
PATTERN_META = {
    "transition_pattern_iso27001_to_cra_maschinenvo_v1.yaml": ("industrial_automation", "regulation", ["CRA", "MaschinenVO"]),
    "transition_pattern_iso27001_to_cra_v1.yaml":             ("industrial_automation", "regulation", ["CRA"]),
    "transition_pattern_iso9001_to_cra_v1.yaml":              ("industrial_automation", "regulation", ["CRA"]),
    "transition_pattern_isms_to_tisax_v1.yaml":               ("automotive", "certification", ["TISAX"]),
    "transition_pattern_iso14001_to_environmental_v1.yaml":   ("environmental", "regulation",
        ["REACH", "RoHS", "Batterieverordnung", "Wasserrecht", "Abwasservorschriften", "Energiemanagement", "Kreislaufwirtschaft", "Emissionsschutz"]),
}

# capability -> dict of sets we aggregate
idx = {}


def _ent(cap):
    return idx.setdefault(cap, {"sources": set(), "types": set(), "domains": set(), "journeys": set(),
                                "regs": set(), "markets": set()})


def _add(cap, sources, stype, domain, journey):
    e = _ent(cap)
    e["sources"] |= set(sources)
    e["types"].add(stype)
    e["domains"].add(domain)
    e["journeys"].add(journey)
    if stype == "regulation":
        e["regs"] |= set(sources)
    if stype == "contract":
        e["markets"] |= set(sources)


# ── A) transition patterns: each pattern is a journey with a target/domain ───────────────────
# IMPORTANT (anti-frequency-deception): a `likely_covered` cap is PROVIDED BY the source cert (one
# certification source), NOT required by every target regulation — attributing all target sources to it
# would inflate management caps on raw frequency alone. Only `delta` caps name their real target sources.
for fname, (domain, ttype, default_sources) in PATTERN_META.items():
    p = yaml.safe_load(open(os.path.join(_TP, fname), encoding="utf-8"))
    journey = p.get("id", fname)
    cert = (p.get("transition_goal", {}).get("from", {}) or {}).get("standard", journey)
    for a in p.get("likely_covered", []):
        _add(a["capability"], [cert], "certification", domain, journey)   # provided by the cert
    for d in p.get("delta_requirements", []):
        srcs = d.get("covers_targets") or default_sources                 # required by these target sources
        _add(d["capability"], srcs, ttype, domain, journey)

# ── B) automotive multi-source data: precise per-source attribution ──────────────────────────
A = yaml.safe_load(open(os.path.join(_HERE, "..", "knowledge", "domains", "automotive", "source_capabilities.yaml"), encoding="utf-8"))
for s in A["sources"]:
    for cap in s["requires"]:
        _add(cap, [s["id"]], s["type"], "automotive", "automotive_ecu")

# ── Impact score (deterministic) ─────────────────────────────────────────────────────────────
def impact(e):
    return len(e["sources"]) + len(e["types"]) + len(e["domains"]) + len(e["journeys"]) + len(e["regs"]) + len(e["markets"])


scored = sorted(idx.items(), key=lambda kv: (-impact(kv[1]), kv[0]))
GENERIC = ("document_", "manage_", "control_", "conduct_", "operate_", "run_", "assign_", "plan_", "approve_")

w("# Cross-Domain MCAP Convergence Analysis — wo konvergiert das Wissensmodell?")
w("")
w('_Nicht „welche MCAPs kommen am häufigsten vor?" (Häufigkeit täuscht), sondern „welche MCAPs TRAGEN den größten Teil des Systems?". Deterministischer **Impact-Score** (kein ML), internes Engineering-Werkzeug, reine Aggregation über vorhandene Daten (5 Transition Patterns + 7 Automotive-Quellen). Non-runtime, keine echten Namen._')
w("")
w("## Impact-Score (deterministisch)")
w("> `Impact = distinct Sources + distinct Target-Types + distinct Domains + distinct Journeys + Regulatory Leverage + Business Leverage`")
w("- %d distinct Capabilities (MCAP-Kandidaten) über alle Quellen aggregiert." % len(idx))
w("")

# ── 1. Core MCAPs ─────────────────────────────────────────────────────────
w("## 1. Core MCAPs — höchster Impact (die tragenden Knoten)")
w("| Capability | Impact | Sources | Types | Domains | Journeys |")
w("|---|---:|---:|---:|---:|---:|")
for cap, e in scored[:10]:
    w("| `%s` | **%d** | %d | %d | %d | %d |" % (cap, impact(e), len(e["sources"]), len(e["types"]), len(e["domains"]), len(e["journeys"])))
w("")
w('→ Hoher Impact = ein Knoten verbindet viele Quellen ÜBER Typen/Domänen/Journeys hinweg — nicht „in 40 Dokumenten einer Normenfamilie".')
w("")

# ── 2. Emerging MCAPs (cross-domain bridges) ──────────────────────────────
emerging = [(c, e) for c, e in scored if len(e["domains"]) >= 2]
w("## 2. Emerging MCAPs — verbinden ≥2 Domänen (Brücken zwischen Anforderungswelten)")
for cap, e in emerging[:8]:
    w("- `%s` — %d Domänen (%s), %d Typen." % (cap, len(e["domains"]), ", ".join(sorted(e["domains"])), len(e["types"])))
w('- _(Echtes „Wachstum über Zeit" braucht historische Snapshots — hier Proxy = Domänen-Spannweite jetzt.)_')
w("")

# ── 3. Isolated MCAPs ─────────────────────────────────────────────────────
isolated = [(c, e) for c, e in scored if len(e["sources"]) == 1 and len(e["journeys"]) == 1]
w("## 3. Isolated MCAPs — nur 1 Quelle/Journey (Review: spezialisiert ODER Konvergenz übersehen?)")
w("- %d Stück, u. a.: %s." % (len(isolated), ", ".join("`%s`" % c for c, _ in isolated[:8])))
w("")

# ── 4. Suspicious MCAPs (abstraction level) ───────────────────────────────
too_coarse = [(c, e) for c, e in scored if c.startswith(GENERIC) and len(e["types"]) <= 1 and len(e["sources"]) >= 2]
too_fine = [(c, e) for c, e in isolated if len(c) >= 34]
w("## 4. Suspicious MCAPs — Abstraktionsgrad-Verdacht (Experten-Review)")
w("- **Evtl. zu grob** (generisches Verb, breit aber nur 1 Typ): %s." % (", ".join("`%s`" % c for c, _ in too_coarse[:6]) or "—"))
w("- **Evtl. zu fein** (isoliert + sehr spezifischer Name): %s." % (", ".join("`%s`" % c for c, _ in too_fine[:6]) or "—"))
w("- Die Analyse sagt damit nicht nur WELCHE MCAPs wichtig sind, sondern auch, ob sie auf dem **richtigen Abstraktionsniveau** definiert sind.")
w("")

# ── Befund ─────────────────────────────────────────────────────────────────
core_cut = [c for c, e in scored if impact(e) >= 8]
cross = [c for c, e in scored if len(e["domains"]) >= 2]
w("## Befund")
w("")
w('> **Ein Kern beginnt sich zu zeigen:** %d von %d Capabilities erreichen Impact ≥ 8 (tragende Knoten), %d verbinden ≥2 Domänen. Bislang ist das Wissensmodell noch jung (5 Patterns + 1 Automotive-Profil), aber die Methode steht: sobald Medical/Payment/weitere Domänen als DATEN hinzukommen, zeigt dieselbe Aggregation, ob sich der erwartete stabile Kern von 30–50 hochkonvergenten MCAPs bildet — der gemeinsame Strukturkern hinter sehr unterschiedlichen Anforderungswelten. Das ist ein tieferer Wertnachweis als „eine weitere Norm unterstützt". Reine Aggregation, 0 Runtime, 0 neue Architektur.' % (len(core_cut), len(idx), len(cross)))
w("")

print("\n".join(OUT))