2026-07-03 13:05:27 +02:00
|
|
|
#!/usr/bin/env python3
|
|
|
|
|
"""
|
|
|
|
|
lint_audit.py - Moteur d'audit (Lint) du knowledge graph Context Continuity.
|
|
|
|
|
|
|
|
|
|
ETAGE 1 : LECTURE SEULE. Detecte et rapporte des anomalies ; ne corrige RIEN
|
|
|
|
|
et n'ecrit RIEN dans Trilium (rapport texte : stdout + fichier horodate dans
|
|
|
|
|
lint_reports/).
|
|
|
|
|
|
|
|
|
|
ORGANISATION PAR DATA QUALITY DIMENSIONS (referentiel autoporteur, extensible).
|
|
|
|
|
Chaque controle declare la dimension qu'il sert ; le rapport est groupe par
|
|
|
|
|
dimension. Ajouter un controle = decorer une fonction avec @check(dimension,...).
|
|
|
|
|
Ajouter une dimension = l'ajouter a DIMENSIONS. Le rapport suit automatiquement.
|
|
|
|
|
|
|
|
|
|
Chaque anomalie porte aussi un bac d'action (prepare l'etage 2) :
|
|
|
|
|
AUTO : correction mecanique sure (accent, casse, enum non ambigue)
|
|
|
|
|
SIGNALER : deterministe mais sans correction evidente -> decision humaine
|
|
|
|
|
JUGEMENT : demande du contexte metier -> revue par l'agent / l'humain
|
|
|
|
|
A l'etage 1, AUCUNE correction n'est appliquee, quel que soit le bac.
|
|
|
|
|
|
|
|
|
|
Usage (sur GrosseBertha, dans le venv) :
|
|
|
|
|
cd ~/App/Context_continuity && source venv/bin/activate
|
|
|
|
|
python3 lint_audit.py
|
|
|
|
|
"""
|
|
|
|
|
|
|
|
|
|
import os
|
|
|
|
|
import json
|
|
|
|
|
import unicodedata
|
|
|
|
|
from datetime import datetime, timezone
|
|
|
|
|
from typing import List, Dict, Optional
|
|
|
|
|
|
|
|
|
|
from trilium_api import (
|
|
|
|
|
search_by_label, get_note, get_note_content,
|
2026-07-17 14:59:33 +02:00
|
|
|
RELATIONS_ONTOLOGIE, TYPES_CANONIQUES, projets_canoniques,
|
2026-07-03 13:05:27 +02:00
|
|
|
)
|
|
|
|
|
|
|
|
|
|
BASE = os.path.expanduser("~/App/Context_continuity")
|
|
|
|
|
IDS = json.load(open(os.path.join(BASE, "trilium_ids.json")))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# ==========================================================================
|
|
|
|
|
# 1. LA TAXONOMIE : data quality dimensions (autoporteur)
|
|
|
|
|
# ==========================================================================
|
|
|
|
|
|
|
|
|
|
# Ordre = ordre d'affichage dans le rapport. Definitions = celles de Bastien
|
|
|
|
|
# (Data Governance), alignees sur les dimensions DAMA standard.
|
|
|
|
|
DIMENSIONS = {
|
|
|
|
|
"completeness": "Les champs obligatoires sont remplis ET les entites attendues sont toutes saisies (exhaustivite).",
|
|
|
|
|
"accuracy": "Le niveau de la donnee reflete la realite (bon rangement, contenu conforme au type, statut a jour).",
|
|
|
|
|
"consistency": "Coherence dans le temps et a travers les entites (vocabulaire courant, informations repetees a jour partout).",
|
|
|
|
|
"validity": "Les valeurs et relations respectent les regles en vigueur (enums, conventions de nommage, ontologie des relations).",
|
|
|
|
|
"uniqueness": "Absence de doublons.",
|
|
|
|
|
"freshness": "Les entites ne portent pas un etat perime (anciennete anormale, non-cloture).",
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
# Controles PREVUS mais NON deterministes : delegues a l'agent reviseur (etage 2).
|
|
|
|
|
# Rendus visibles pour l'honnetete de la couverture.
|
|
|
|
|
CHECKS_DELEGUES = [
|
|
|
|
|
{"dimension": "completeness", "code": "COMP-exhaustivite",
|
|
|
|
|
"libelle": "Entites manquantes vs realite du projet (ex : backlog incomplet)",
|
|
|
|
|
"raison": "Non detectable en deterministe : requiert la comparaison avec la realite (code, conversations). -> agent reviseur."},
|
|
|
|
|
{"dimension": "accuracy", "code": "ACC-contenu-type",
|
|
|
|
|
"libelle": "Le contenu redige d'une note reflete-t-il bien son type",
|
|
|
|
|
"raison": "Requiert un jugement semantique. -> agent reviseur."},
|
|
|
|
|
{"dimension": "consistency", "code": "CONS-chiffres-croises",
|
|
|
|
|
"libelle": "Une information repetee (ex : nombre de tools) est-elle a jour dans toutes les notes",
|
|
|
|
|
"raison": "Partiellement couvert par CONS-vocab ; la verification exhaustive des chiffres croises -> agent reviseur."},
|
|
|
|
|
]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# ==========================================================================
|
|
|
|
|
# 2. REFERENTIEL : ce que le systeme considere comme valide
|
|
|
|
|
# ==========================================================================
|
|
|
|
|
|
|
|
|
|
ENUMS = {
|
|
|
|
|
("projet", "statut"): {"actif", "en-pause", "archive"},
|
|
|
|
|
("conversation", "statut"): {"en-cours", "clos"},
|
|
|
|
|
("backlogItem", "statut"): {"a faire", "en cours", "bloque", "fait", "abandonne"},
|
|
|
|
|
("backlogItem", "priorite"): {"haute", "moyenne", "basse"},
|
|
|
|
|
("decision", "statut"): {"active", "revisee", "annulee"},
|
|
|
|
|
("historiqueItem", "typeHistorique"): {
|
|
|
|
|
"Fait etabli", "Test effectue", "Hypothese invalidee", "Contrainte decouverte",
|
|
|
|
|
},
|
|
|
|
|
("historiqueItem", "encoreValide"): {"true", "false"},
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
LABELS_OBLIGATOIRES = {
|
|
|
|
|
"projet": ["projet", "statut"],
|
|
|
|
|
"conversation": ["projet", "llm", "statut", "date"],
|
|
|
|
|
"backlogItem": ["projet", "statut", "priorite"],
|
|
|
|
|
"decision": ["projet", "statut"],
|
|
|
|
|
"historiqueItem": ["projet", "typeHistorique", "encoreValide"],
|
|
|
|
|
"concept": ["projet"],
|
|
|
|
|
"contexteReprise": ["projet", "llmCible", "version"],
|
|
|
|
|
"composantLogiciel": ["projet"],
|
|
|
|
|
"service": ["projet"],
|
|
|
|
|
"infrastructure": ["projet"],
|
|
|
|
|
"outil": ["projet"],
|
|
|
|
|
"contrainte": ["projet"],
|
|
|
|
|
"principe": ["projet"],
|
|
|
|
|
"convention": ["projet"],
|
|
|
|
|
"methode": ["projet"],
|
|
|
|
|
"version": ["projet"],
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
DOSSIER_ATTENDU = {
|
|
|
|
|
"backlogItem": "Backlog",
|
|
|
|
|
"decision": "Decisions",
|
|
|
|
|
"historiqueItem": "Historique",
|
|
|
|
|
"conversation": "Conversations",
|
|
|
|
|
"concept": "Concepts",
|
|
|
|
|
"contexteReprise": "ContextesReprise",
|
|
|
|
|
"projet": "Projets",
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
TYPES_METIER = list(LABELS_OBLIGATOIRES.keys()) + ["skill", "documentation"]
|
|
|
|
|
|
|
|
|
|
TYPES_DEPRECIES = {"termeGlossaire"}
|
|
|
|
|
MOTS_DEPRECIES = ["type systeme", "type système", "termeGlossaire"]
|
|
|
|
|
|
|
|
|
|
JOURS_STALE_ENCOURS = 90
|
|
|
|
|
JOURS_STALE_CONVO = 30
|
|
|
|
|
|
|
|
|
|
# Domaine -> portee (sous-ensemble ROBUSTE ; classes mappees aux types reels).
|
|
|
|
|
# domaine None = seule la portee est verifiee. Table volontairement partielle :
|
|
|
|
|
# uniquement les relations dont le mapping classe-ontologie -> type est certain.
|
|
|
|
|
RELATION_DP = {
|
|
|
|
|
"impacte": ({"decision", "historiqueItem"}, {"composantLogiciel", "service"}),
|
|
|
|
|
"revise": ({"decision"}, {"decision"}),
|
|
|
|
|
"dependDe": ({"composantLogiciel", "service"}, {"composantLogiciel", "service", "outil"}),
|
|
|
|
|
"aVersion": ({"composantLogiciel", "service"}, {"version"}),
|
|
|
|
|
"implementePar": ({"service"}, {"composantLogiciel"}),
|
|
|
|
|
"contraintPar": ({"decision", "composantLogiciel"}, {"contrainte", "principe"}),
|
|
|
|
|
"illustre": ({"concept", "composantLogiciel", "service"}, {"concept"}),
|
2026-07-17 22:36:43 +02:00
|
|
|
"documentePar": ({"service", "outil", "methode", "composantLogiciel", "historiqueItem", "decision"}, {"skill", "documentation"}),
|
2026-07-03 13:05:27 +02:00
|
|
|
"concerneProjet": (None, {"projet"}),
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# ==========================================================================
|
|
|
|
|
# 3. REGISTRE DES CONTROLES (decorateur autoporteur)
|
|
|
|
|
# ==========================================================================
|
|
|
|
|
|
|
|
|
|
CHECKS: List[dict] = [] # {dimension, code, libelle, fn}
|
|
|
|
|
_CTX = {"dimension": None} # dimension du check en cours d'execution
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def check(dimension: str, code: str, libelle: str):
|
|
|
|
|
"""Enregistre une fonction de controle sous une dimension de qualite."""
|
|
|
|
|
def deco(fn):
|
|
|
|
|
CHECKS.append({"dimension": dimension, "code": code, "libelle": libelle, "fn": fn})
|
|
|
|
|
return fn
|
|
|
|
|
return deco
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# ==========================================================================
|
|
|
|
|
# 4. UTILITAIRES + CHARGEMENT (une passe, cache)
|
|
|
|
|
# ==========================================================================
|
|
|
|
|
|
|
|
|
|
def sans_accents(s: str) -> str:
|
|
|
|
|
return "".join(c for c in unicodedata.normalize("NFD", s)
|
|
|
|
|
if unicodedata.category(c) != "Mn")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
NOTES: Dict[str, dict] = {}
|
|
|
|
|
BY_TYPE: Dict[str, List[str]] = {}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def labels_of(note: dict) -> Dict[str, str]:
|
|
|
|
|
return {a.get("name"): a.get("value", "")
|
|
|
|
|
for a in note.get("attributes", []) if a.get("type") == "label"}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def relations_of(note: dict) -> List[Dict[str, str]]:
|
|
|
|
|
return [{"name": a.get("name"), "value": a.get("value", "")}
|
|
|
|
|
for a in note.get("attributes", []) if a.get("type") == "relation"]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def type_of(note: dict) -> Optional[str]:
|
|
|
|
|
return labels_of(note).get("type")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def age_jours(note: dict) -> Optional[int]:
|
|
|
|
|
dm = note.get("utcDateModified") or note.get("dateModified")
|
|
|
|
|
if not dm:
|
|
|
|
|
return None
|
|
|
|
|
for fmt in ("%Y-%m-%d %H:%M:%S.%fZ", "%Y-%m-%d %H:%M:%SZ",
|
|
|
|
|
"%Y-%m-%dT%H:%M:%S.%fZ", "%Y-%m-%d %H:%M:%S.%f"):
|
|
|
|
|
try:
|
|
|
|
|
d = datetime.strptime(dm, fmt).replace(tzinfo=timezone.utc)
|
|
|
|
|
return (datetime.now(timezone.utc) - d).days
|
|
|
|
|
except Exception:
|
|
|
|
|
continue
|
|
|
|
|
return None
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def charger():
|
|
|
|
|
"""Charge TOUTES les notes portant un label type (recherche #type globale),
|
|
|
|
|
quel que soit le type - y compris les types invalides/inconnus, pour que
|
|
|
|
|
VAL-type puisse les detecter. Indexe par la valeur reelle du label type."""
|
|
|
|
|
try:
|
|
|
|
|
res = search_by_label("type", "") # #type sans valeur = toutes les notes typees
|
|
|
|
|
except Exception as e:
|
|
|
|
|
print(" ! echec recherche #type : %s" % e)
|
|
|
|
|
res = []
|
|
|
|
|
for r in res:
|
|
|
|
|
nid = r.get("noteId")
|
|
|
|
|
if not nid or nid in NOTES:
|
|
|
|
|
continue
|
|
|
|
|
try:
|
|
|
|
|
note = get_note(nid)
|
|
|
|
|
except Exception:
|
|
|
|
|
continue
|
|
|
|
|
NOTES[nid] = note
|
|
|
|
|
tv = type_of(note) or "(sans type)"
|
|
|
|
|
BY_TYPE.setdefault(tv, []).append(nid)
|
|
|
|
|
print(" Charge : %d notes sur %d valeurs de type" % (len(NOTES), len(BY_TYPE)))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# ==========================================================================
|
|
|
|
|
# 5. COLLECTE DES ANOMALIES
|
|
|
|
|
# ==========================================================================
|
|
|
|
|
|
|
|
|
|
ANOMALIES: List[dict] = []
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def signaler(code, bac, note_id, titre, detail, suggestion=""):
|
|
|
|
|
"""Enregistre une anomalie ; la dimension est heritee du check courant."""
|
|
|
|
|
ANOMALIES.append({
|
|
|
|
|
"dimension": _CTX["dimension"], "code": code, "bac": bac,
|
|
|
|
|
"note_id": note_id, "titre": titre, "detail": detail, "suggestion": suggestion,
|
|
|
|
|
})
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# ==========================================================================
|
|
|
|
|
# 6. LES CONTROLES, GROUPES PAR DIMENSION
|
|
|
|
|
# ==========================================================================
|
|
|
|
|
|
|
|
|
|
# ---- COMPLETENESS --------------------------------------------------------
|
|
|
|
|
|
|
|
|
|
@check("completeness", "COMP-labels", "Labels obligatoires presents selon le type")
|
|
|
|
|
def c_labels_obligatoires():
|
|
|
|
|
for typ, obligatoires in LABELS_OBLIGATOIRES.items():
|
|
|
|
|
for nid in BY_TYPE.get(typ, []):
|
|
|
|
|
note = NOTES[nid]
|
|
|
|
|
labels = labels_of(note)
|
|
|
|
|
manquants = [l for l in obligatoires if not labels.get(l)]
|
|
|
|
|
if manquants:
|
|
|
|
|
signaler("COMP-labels", "SIGNALER", nid, note.get("title", "?"),
|
|
|
|
|
"type=%s : labels manquants %s" % (typ, manquants),
|
|
|
|
|
"ajouter les labels manquants")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# ---- ACCURACY ------------------------------------------------------------
|
|
|
|
|
|
|
|
|
|
@check("accuracy", "ACC-rangement", "Note rangee dans le dossier attendu de son type")
|
|
|
|
|
def a_rangement():
|
|
|
|
|
for typ, cle_dossier in DOSSIER_ATTENDU.items():
|
|
|
|
|
dossier_id = IDS.get(cle_dossier)
|
|
|
|
|
if not dossier_id:
|
|
|
|
|
continue
|
|
|
|
|
for nid in BY_TYPE.get(typ, []):
|
|
|
|
|
note = NOTES[nid]
|
|
|
|
|
parents = note.get("parentNoteIds", [])
|
|
|
|
|
if parents and dossier_id not in parents:
|
|
|
|
|
signaler("ACC-rangement", "JUGEMENT", nid, note.get("title", "?"),
|
|
|
|
|
"type=%s hors du dossier %s (parents=%s)" % (typ, cle_dossier, parents),
|
|
|
|
|
"verifier le rangement (clone legitime possible)")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
@check("accuracy", "ACC-decision", "Statut des decisions actives reflete la realite (v1 vs v2)")
|
|
|
|
|
def a_decisions_obsoletes():
|
|
|
|
|
SEUIL = 12
|
|
|
|
|
par_projet = {}
|
|
|
|
|
for nid in BY_TYPE.get("decision", []):
|
|
|
|
|
labels = labels_of(NOTES[nid])
|
|
|
|
|
if labels.get("statut") == "active":
|
|
|
|
|
par_projet.setdefault(labels.get("projet", "?"), []).append(nid)
|
|
|
|
|
for proj, ids in par_projet.items():
|
|
|
|
|
if len(ids) >= SEUIL:
|
|
|
|
|
signaler("ACC-decision", "JUGEMENT", ids[0], "(projet %s)" % proj,
|
|
|
|
|
"%d decisions actives sur %s : possibles v1 non revisees" % (len(ids), proj),
|
|
|
|
|
"revue : tisser ~revise, passer les v1 obsoletes en revisee")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# ---- CONSISTENCY ---------------------------------------------------------
|
|
|
|
|
|
|
|
|
|
@check("consistency", "CONS-vocab", "Vocabulaire courant (types et termes non deprecies)")
|
|
|
|
|
def cons_vocabulaire():
|
|
|
|
|
# Types deprecies
|
|
|
|
|
for typ in TYPES_DEPRECIES:
|
|
|
|
|
for nid in BY_TYPE.get(typ, []):
|
|
|
|
|
signaler("CONS-type", "SIGNALER", nid, NOTES[nid].get("title", "?"),
|
|
|
|
|
"type deprecie : %s" % typ, "migrer vers concept")
|
|
|
|
|
# Mots deprecies dans le contenu des skills / documentation
|
|
|
|
|
for typ in ("skill", "documentation"):
|
|
|
|
|
for nid in BY_TYPE.get(typ, []):
|
|
|
|
|
try:
|
|
|
|
|
contenu = get_note_content(nid)
|
|
|
|
|
except Exception:
|
|
|
|
|
continue
|
|
|
|
|
trouves = [m for m in MOTS_DEPRECIES if m in contenu]
|
|
|
|
|
if trouves:
|
|
|
|
|
signaler("CONS-vocab", "JUGEMENT", nid, NOTES[nid].get("title", "?"),
|
|
|
|
|
"mots deprecies dans le contenu : %s" % trouves,
|
|
|
|
|
"reformuler (faux positif possible si mention explicative legitime)")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# ---- VALIDITY ------------------------------------------------------------
|
|
|
|
|
|
|
|
|
|
@check("validity", "VAL-enum", "Valeurs de labels conformes aux enumerations permises")
|
|
|
|
|
def v_enums():
|
|
|
|
|
for (typ, label), valides in ENUMS.items():
|
|
|
|
|
for nid in BY_TYPE.get(typ, []):
|
|
|
|
|
note = NOTES[nid]
|
|
|
|
|
val = labels_of(note).get(label)
|
|
|
|
|
if val is None or val in valides:
|
|
|
|
|
continue
|
|
|
|
|
titre = note.get("title", "?")
|
|
|
|
|
cible = None
|
|
|
|
|
for v in valides:
|
|
|
|
|
if sans_accents(val.lower()) == sans_accents(v.lower()):
|
|
|
|
|
cible = v
|
|
|
|
|
break
|
|
|
|
|
if cible is not None:
|
|
|
|
|
signaler("VAL-enum", "AUTO", nid, titre,
|
|
|
|
|
"%s=%r (casse/accent)" % (label, val), "corriger en %r" % cible)
|
|
|
|
|
else:
|
|
|
|
|
signaler("VAL-enum", "SIGNALER", nid, titre,
|
|
|
|
|
"%s=%r hors valeurs permises %s" % (label, val, sorted(valides)),
|
|
|
|
|
"corriger manuellement")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
@check("validity", "VAL-type", "Label type appartient a la liste canonique de l ontologie")
|
|
|
|
|
def v_type():
|
|
|
|
|
for nid, note in NOTES.items():
|
|
|
|
|
tv = type_of(note)
|
|
|
|
|
if tv is None or tv == "container" or tv in TYPES_CANONIQUES:
|
|
|
|
|
continue # container = dossier structurel, hors perimetre
|
|
|
|
|
titre = note.get("title", "?")
|
|
|
|
|
cible = None
|
|
|
|
|
for t in TYPES_CANONIQUES:
|
|
|
|
|
if sans_accents(tv.lower()) == sans_accents(t.lower()):
|
|
|
|
|
cible = t
|
|
|
|
|
break
|
|
|
|
|
if cible is not None:
|
|
|
|
|
signaler("VAL-type", "AUTO", nid, titre,
|
|
|
|
|
"type=%r (casse)" % tv, "corriger en %r" % cible)
|
|
|
|
|
else:
|
|
|
|
|
signaler("VAL-type", "SIGNALER", nid, titre,
|
|
|
|
|
"type=%r hors liste canonique" % tv, "corriger ou ajouter a l ontologie")
|
|
|
|
|
|
|
|
|
|
|
2026-07-17 14:59:33 +02:00
|
|
|
@check("validity", "VAL-nommage", "Label projet appartient au referentiel (notes de type projet)")
|
2026-07-03 13:05:27 +02:00
|
|
|
def v_nommage():
|
2026-07-17 14:59:33 +02:00
|
|
|
canon = projets_canoniques()
|
|
|
|
|
if not canon:
|
|
|
|
|
return # referentiel illisible : ne pas produire de faux positifs
|
|
|
|
|
import unicodedata
|
|
|
|
|
def norm(s):
|
|
|
|
|
s = "".join(ch for ch in unicodedata.normalize("NFD", s) if unicodedata.category(ch) != "Mn")
|
|
|
|
|
return s.lower().replace(" ", "").replace("_", "").replace("-", "")
|
2026-07-03 13:05:27 +02:00
|
|
|
for nid, note in NOTES.items():
|
|
|
|
|
val = labels_of(note).get("projet")
|
2026-07-17 14:59:33 +02:00
|
|
|
if not val or val in canon:
|
|
|
|
|
continue
|
|
|
|
|
# Chercher un canonique proche (faute de casse/espace/accent) -> AUTO
|
|
|
|
|
cible = None
|
|
|
|
|
for cand in canon:
|
|
|
|
|
if norm(cand) == norm(val):
|
|
|
|
|
cible = cand
|
|
|
|
|
break
|
|
|
|
|
if cible is not None:
|
|
|
|
|
signaler("VAL-nommage", "AUTO", nid, note.get("title", "?"),
|
|
|
|
|
"projet=%r non canonique" % val, "corriger en %r" % cible)
|
|
|
|
|
else:
|
2026-07-03 13:05:27 +02:00
|
|
|
signaler("VAL-nommage", "SIGNALER", nid, note.get("title", "?"),
|
2026-07-17 14:59:33 +02:00
|
|
|
"projet=%r hors referentiel" % val,
|
|
|
|
|
"corriger ou creer la note de type projet correspondante")
|
2026-07-03 13:05:27 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
@check("validity", "VAL-relation", "Relations valides (ontologie, cible existante, domaine/portee)")
|
|
|
|
|
def v_relations():
|
|
|
|
|
for nid, note in NOTES.items():
|
|
|
|
|
typ_src = type_of(note)
|
|
|
|
|
titre = note.get("title", "?")
|
|
|
|
|
for rel in relations_of(note):
|
|
|
|
|
nom, cible = rel["name"], rel["value"]
|
|
|
|
|
if nom not in RELATIONS_ONTOLOGIE:
|
|
|
|
|
signaler("VAL-rel-nom", "SIGNALER", nid, titre,
|
|
|
|
|
"relation ~%s hors ontologie" % nom, "renommer ou retirer")
|
|
|
|
|
continue
|
|
|
|
|
cible_note = NOTES.get(cible)
|
|
|
|
|
if cible_note is None:
|
|
|
|
|
try:
|
|
|
|
|
cible_note = get_note(cible)
|
|
|
|
|
except Exception:
|
|
|
|
|
cible_note = None
|
|
|
|
|
if not cible_note:
|
|
|
|
|
signaler("VAL-rel-cible", "SIGNALER", nid, titre,
|
|
|
|
|
"~%s pointe vers une note absente (%s)" % (nom, cible),
|
|
|
|
|
"corriger ou retirer la relation")
|
|
|
|
|
continue
|
|
|
|
|
dp = RELATION_DP.get(nom)
|
|
|
|
|
if dp:
|
|
|
|
|
dom, por = dp
|
|
|
|
|
typ_cible = type_of(cible_note)
|
|
|
|
|
if dom is not None and typ_src not in dom:
|
|
|
|
|
signaler("VAL-rel-domaine", "JUGEMENT", nid, titre,
|
|
|
|
|
"~%s : source type=%s hors domaine %s" % (nom, typ_src, sorted(dom)),
|
|
|
|
|
"verifier la pertinence")
|
|
|
|
|
if por is not None and typ_cible not in por:
|
|
|
|
|
signaler("VAL-rel-portee", "JUGEMENT", nid, titre,
|
|
|
|
|
"~%s : cible type=%s hors portee %s" % (nom, typ_cible, sorted(por)),
|
|
|
|
|
"verifier la pertinence")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# ---- UNIQUENESS ----------------------------------------------------------
|
|
|
|
|
|
|
|
|
|
@check("uniqueness", "UNIQ-doublon", "Absence de notes identiques (titre + type + projet)")
|
|
|
|
|
def u_doublons():
|
|
|
|
|
vus = {}
|
|
|
|
|
for nid, note in NOTES.items():
|
|
|
|
|
cle = ((note.get("title") or "").strip().lower(),
|
|
|
|
|
type_of(note), labels_of(note).get("projet", ""))
|
|
|
|
|
vus.setdefault(cle, []).append(nid)
|
|
|
|
|
for (titre, typ, proj), ids in vus.items():
|
|
|
|
|
if len(ids) > 1:
|
|
|
|
|
signaler("UNIQ-doublon", "SIGNALER", ids[0], titre or "(sans titre)",
|
|
|
|
|
"%d notes identiques (type=%s projet=%s) : %s" % (len(ids), typ, proj, ids),
|
|
|
|
|
"fusionner ou supprimer les doublons")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# ---- FRESHNESS -----------------------------------------------------------
|
|
|
|
|
|
|
|
|
|
@check("freshness", "FRESH-staleness", "Entites sans etat perime (anciennete, non-cloture)")
|
|
|
|
|
def f_staleness():
|
|
|
|
|
for nid in BY_TYPE.get("backlogItem", []):
|
|
|
|
|
note = NOTES[nid]
|
|
|
|
|
if labels_of(note).get("statut") == "en cours":
|
|
|
|
|
age = age_jours(note)
|
|
|
|
|
if age is not None and age > JOURS_STALE_ENCOURS:
|
|
|
|
|
signaler("FRESH-backlog", "JUGEMENT", nid, note.get("title", "?"),
|
|
|
|
|
"backlog 'en cours' inchange depuis %d j" % age, "verifier si toujours actif")
|
|
|
|
|
for nid in BY_TYPE.get("conversation", []):
|
|
|
|
|
note = NOTES[nid]
|
|
|
|
|
if labels_of(note).get("statut") == "en-cours":
|
|
|
|
|
age = age_jours(note)
|
|
|
|
|
if age is not None and age > JOURS_STALE_CONVO:
|
|
|
|
|
signaler("FRESH-conversation", "JUGEMENT", nid, note.get("title", "?"),
|
|
|
|
|
"conversation 'en-cours' non cloturee depuis %d j" % age, "cloturer si terminee")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# ==========================================================================
|
|
|
|
|
# 7. ORCHESTRATION + RAPPORT (groupes par dimension)
|
|
|
|
|
# ==========================================================================
|
|
|
|
|
|
|
|
|
|
def executer():
|
|
|
|
|
print("Chargement des notes...")
|
|
|
|
|
charger()
|
|
|
|
|
print("Execution des controles par dimension...")
|
|
|
|
|
ordre = {d: i for i, d in enumerate(DIMENSIONS)}
|
|
|
|
|
for chk in sorted(CHECKS, key=lambda c: ordre.get(c["dimension"], 99)):
|
|
|
|
|
_CTX["dimension"] = chk["dimension"]
|
|
|
|
|
try:
|
|
|
|
|
chk["fn"]()
|
|
|
|
|
except Exception as e:
|
|
|
|
|
print(" ! controle %s (%s) a echoue : %s" % (chk["code"], chk["dimension"], e))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def rapport() -> str:
|
|
|
|
|
date = datetime.now().strftime("%Y-%m-%d %H:%M")
|
|
|
|
|
L = ["# RAPPORT LINT - Context Continuity - %s" % date,
|
|
|
|
|
"# ETAGE 1 (lecture seule) - groupe par data quality dimension", ""]
|
|
|
|
|
|
|
|
|
|
# Inventaire
|
|
|
|
|
L.append("## Inventaire")
|
|
|
|
|
L.append("- Notes chargees : %d" % len(NOTES))
|
|
|
|
|
for t in sorted(BY_TYPE):
|
|
|
|
|
if BY_TYPE[t]:
|
|
|
|
|
L.append(" - %s : %d" % (t, len(BY_TYPE[t])))
|
|
|
|
|
L.append("")
|
|
|
|
|
|
|
|
|
|
# Synthese
|
|
|
|
|
par_dim = {d: 0 for d in DIMENSIONS}
|
|
|
|
|
par_bac = {"AUTO": 0, "SIGNALER": 0, "JUGEMENT": 0}
|
|
|
|
|
for a in ANOMALIES:
|
|
|
|
|
par_dim[a["dimension"]] = par_dim.get(a["dimension"], 0) + 1
|
|
|
|
|
par_bac[a["bac"]] = par_bac.get(a["bac"], 0) + 1
|
|
|
|
|
L.append("## Synthese : %d anomalies" % len(ANOMALIES))
|
|
|
|
|
L.append(" Par dimension : " + ", ".join("%s=%d" % (d, par_dim[d]) for d in DIMENSIONS))
|
|
|
|
|
L.append(" Par bac d'action : " + ", ".join("%s=%d" % (b, par_bac[b]) for b in ("AUTO", "SIGNALER", "JUGEMENT")))
|
|
|
|
|
L.append("")
|
|
|
|
|
|
|
|
|
|
# Detail par dimension
|
|
|
|
|
for dim, definition in DIMENSIONS.items():
|
|
|
|
|
anos = [a for a in ANOMALIES if a["dimension"] == dim]
|
|
|
|
|
actifs = [c for c in CHECKS if c["dimension"] == dim]
|
|
|
|
|
delegues = [c for c in CHECKS_DELEGUES if c["dimension"] == dim]
|
|
|
|
|
L.append("## %s (%d anomalie%s)" % (dim.upper(), len(anos), "s" if len(anos) != 1 else ""))
|
|
|
|
|
L.append(" %s" % definition)
|
|
|
|
|
L.append(" Controles deterministes : " + ", ".join(c["code"] for c in actifs))
|
|
|
|
|
for d in delegues:
|
|
|
|
|
L.append(" [delegue agent] %s : %s" % (d["code"], d["libelle"]))
|
|
|
|
|
if anos:
|
|
|
|
|
L.append("")
|
|
|
|
|
for a in anos:
|
|
|
|
|
L.append(" - [%s | %s] %s (%s)" % (a["code"], a["bac"], a["titre"], a["note_id"]))
|
|
|
|
|
L.append(" %s" % a["detail"])
|
|
|
|
|
if a["suggestion"]:
|
|
|
|
|
L.append(" -> %s" % a["suggestion"])
|
|
|
|
|
L.append("")
|
|
|
|
|
return "\n".join(L)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def main():
|
|
|
|
|
executer()
|
|
|
|
|
txt = rapport()
|
|
|
|
|
print("\n" + txt)
|
|
|
|
|
rep_dir = os.path.join(BASE, "lint_reports")
|
|
|
|
|
os.makedirs(rep_dir, exist_ok=True)
|
|
|
|
|
chemin = os.path.join(rep_dir, "lint_%s.txt" % datetime.now().strftime("%Y%m%d_%H%M%S"))
|
|
|
|
|
with open(chemin, "w", encoding="utf-8") as f:
|
|
|
|
|
f.write(txt)
|
|
|
|
|
print("\nRapport ecrit : %s" % chemin)
|
|
|
|
|
if len(ANOMALIES) > 100:
|
|
|
|
|
print("\n!!! ALERTE : %d anomalies (>100). Investiguer avant toute correction." % len(ANOMALIES))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
if __name__ == "__main__":
|
|
|
|
|
main()
|