WP24c - Agentic Edge Validation & Chunk-Aware Multigraph-System (v4.5.8) #22

Merged
Lars merged 71 commits from WP24c into main 2026-01-12 10:53:20 +01:00
2 changed files with 25 additions and 18 deletions
Showing only changes of commit 4318395c83 - Show all commits

View File

@ -4,7 +4,6 @@ DESCRIPTION: Datenbank-Schnittstelle für Note-Metadaten und Artefakt-Prüfung.
WP-14: Umstellung auf zentrale database-Infrastruktur.
WP-24c: Implementierung der herkunftsbasierten Lösch-Logik (Origin-Purge).
Verhindert das versehentliche Löschen von inversen Kanten beim Re-Import.
VERSION v2.2.0: Integration der Authority-Prüfung für Point-IDs.
VERSION: 2.2.0 (WP-24c: Protected Purge & Authority Lookup)
STATUS: Active
"""
@ -50,6 +49,7 @@ def is_explicit_edge_present(client: QdrantClient, prefix: str, edge_id: str) ->
"""
_, _, edges_col = collection_names(prefix)
try:
# retrieve erwartet eine Liste von IDs
res = client.retrieve(
collection_name=edges_col,
ids=[edge_id],
@ -83,7 +83,7 @@ def purge_artifacts(client: QdrantClient, prefix: str, note_id: str):
# Dies umfasst:
# - Alle ausgehenden Kanten (A -> B)
# - Alle inversen Kanten, die diese Note in anderen Notizen "deponiert" hat (B -> A)
# Fremde virtuelle Kanten (C -> A) bleiben erhalten, da deren origin_note_id == C ist.
# Fremde inverse Kanten (C -> A) bleiben erhalten.
edges_filter = rest.Filter(must=[
rest.FieldCondition(key="origin_note_id", match=rest.MatchValue(value=note_id))
])

View File

@ -5,9 +5,9 @@ DESCRIPTION: Der zentrale IngestionService (Orchestrator).
WP-25a: Integration der Mixture of Experts (MoE) Architektur.
WP-15b: Two-Pass Workflow mit globalem Kontext-Cache.
WP-20/22: Cloud-Resilienz und Content-Lifecycle integriert.
AUDIT v3.2.0: Fix für KeyError 'target_id', TypeError (Sync-Check)
und Business-Centric Logging.
VERSION: 3.2.0 (WP-24c: Stability & Full Feature Set)
AUDIT v3.2.1: Fix für ID-Kanonisierung in Phase 1 & 2,
robuster Smart-Edge-Logger und Business-Logging.
VERSION: 3.2.1 (WP-24c: Canonical Authority Protection)
STATUS: Active
"""
import logging
@ -56,8 +56,8 @@ class IngestionService:
from app.config import get_settings
self.settings = get_settings()
# --- LOGGING CLEANUP ---
# Unterdrückt Bibliotheks-Lärm in Konsole und Datei
# --- LOGGING CLEANUP (Business Focus) ---
# Unterdrückt Bibliotheks-Lärm in Konsole und Datei (via tee)
logging.getLogger("httpx").setLevel(logging.WARNING)
logging.getLogger("httpcore").setLevel(logging.WARNING)
logging.getLogger("qdrant_client").setLevel(logging.WARNING)
@ -99,18 +99,18 @@ class IngestionService:
if not text or len(text.strip()) < 2:
return False
# Symmetrie-Filter gegen Typ-Strings
# Nur System-Kanten (Symmetrie) filtern wir gegen die Typ-Blacklist
if provenance != "explicit":
blacklisted = {"insight", "event", "source", "task", "project", "person", "concept", "related_to", "referenced_by"}
if text.lower().strip() in blacklisted:
return False
if len(text) > 150: return False # Wahrscheinlich kein Titel
if len(text) > 150: return False # Vermutlich ein ganzer Satz
return True
async def run_batch(self, file_paths: List[str], vault_root: str) -> List[Dict[str, Any]]:
"""
WP-15b: Implementiert den Two-Pass Ingestion Workflow.
WP-15b: Two-Pass Ingestion Workflow.
Pass 1: Pre-Scan füllt den Context-Cache.
Pass 2: Verarbeitung nutzt den Cache für die semantische Prüfung.
"""
@ -122,7 +122,7 @@ class IngestionService:
# Übergabe der Registry für dynamische Scan-Tiefe
ctx = pre_scan_markdown(path, registry=self.registry)
if ctx:
# Mehrfache Indizierung für robusten Look-up
# Mehrfache Indizierung für robusten Look-up (ID, Titel, Dateiname)
self.batch_cache[ctx.note_id] = ctx
self.batch_cache[ctx.title] = ctx
fname = os.path.splitext(os.path.basename(path))[0]
@ -174,6 +174,7 @@ class IngestionService:
)
note_id = note_pl["note_id"]
# BUSINESS LOG: Aktuelle Notiz
logger.info(f"📄 Bearbeite: '{note_id}' (Typ: {note_type})")
old_payload = None if force_replace else fetch_note_payload(self.client, self.prefix, note_id)
@ -208,9 +209,9 @@ class IngestionService:
is_valid = await validate_edge_candidate(
ch.text, cand, self.batch_cache, self.llm, profile_name="ingest_validator"
)
# Fix v3.2.0: Sicherer Zugriff via .get() verhindert Crash bei fehlender target_id
t_id = cand.get('target_id') or cand.get('note_id') or "Unknown"
logger.info(f" 🧠 [SMART EDGE] {t_id} -> {'✅ OK' if is_valid else '❌ SKIP'}")
# Fix v3.2.1: Robuste ID-Auflösung für den Logger
t_label = cand.get('target_id') or cand.get('note_id') or cand.get('to') or "Unknown"
logger.info(f" 🧠 [SMART EDGE] {t_label} -> {'✅ OK' if is_valid else '❌ SKIP'}")
if is_valid: new_pool.append(cand)
else:
new_pool.append(cand)
@ -245,8 +246,12 @@ class IngestionService:
e["virtual"] = False
e["confidence"] = e.get("confidence", 1.0)
# Registrierung für Batch-Authority
edge_id = _mk_edge_id(resolved_kind, note_id, target_raw, e.get("scope", "note"))
# Fix v3.2.1: Kanonisierung der Target-ID vor der Registrierung!
# Nur wenn wir hier die echte Note-ID nutzen, erkennt Phase 2 die Kollision.
t_ctx = self.batch_cache.get(target_raw)
t_canonical = t_ctx.note_id if t_ctx else target_raw
edge_id = _mk_edge_id(resolved_kind, note_id, t_canonical, e.get("scope", "note"))
self.processed_explicit_ids.add(edge_id)
final_edges.append(e)
@ -260,11 +265,12 @@ class IngestionService:
target_id = target_ctx.note_id if target_ctx else target_raw
if (inv_kind and target_id and target_id != note_id and self._is_valid_note_id(target_id, provenance="structure")):
# ID der potenziellen virtuellen Kante
potential_id = _mk_edge_id(inv_kind, target_id, note_id, e.get("scope", "note"))
is_in_batch = potential_id in self.processed_explicit_ids
# Real-Time DB Check (Ohne 'await', da sync)
# Real-Time DB Check (Sync)
is_in_db = False
if not is_in_batch:
is_in_db = is_explicit_edge_present(self.client, self.prefix, potential_id)
@ -312,6 +318,7 @@ class IngestionService:
"""Erstellt eine Note aus einem Textstream."""
target_path = os.path.join(vault_root, folder, filename)
os.makedirs(os.path.dirname(target_path), exist_ok=True)
with open(target_path, "w", encoding="utf-8") as f: f.write(markdown_content)
with open(target_path, "w", encoding="utf-8") as f:
f.write(markdown_content)
await asyncio.sleep(0.1)
return await self.process_file(file_path=target_path, vault_root=vault_root, apply=True, force_replace=True, purge_before=True)