Implement WP-15c enhancements across graph and retrieval modules, including full metadata support for Super-Edge aggregation and Note-Level Diversity Pooling. Update scoring logic to reflect new edge handling and improve retrieval accuracy. Version updates to reflect these changes.
This commit is contained in:
1 parent
cd5056d4c9
commit
d35bdc64b9
5 files changed
+202
-91
No files matched your search
@@ -1,13 +1,15 @@
|
||||
"""
|
||||
FILE: app/core/graph/graph_db_adapter.py
|
||||
DESCRIPTION: Datenbeschaffung aus Qdrant für den Graphen.
|
||||
AUDIT v1.1.0: Nutzt nun die zentrale database-Infrastruktur für Namen.
|
||||
AUDIT v1.1.1: Volle Unterstützung für WP-15c Metadaten.
|
||||
Stellt sicher, dass 'target_section' und 'provenance' für die
|
||||
Super-Edge-Aggregation im Retriever geladen werden.
|
||||
"""
|
||||
from typing import List, Dict, Optional
|
||||
from qdrant_client import QdrantClient
|
||||
from qdrant_client.http import models as rest
|
||||
|
||||
# ENTSCHEIDENDER FIX: Nutzt die neue Infrastruktur für konsistente Collection-Namen
|
||||
# Nutzt die zentrale Infrastruktur für konsistente Collection-Namen (WP-14)
|
||||
from app.core.database import collection_names
|
||||
|
||||
def fetch_edges_from_qdrant(
|
||||
@@ -19,14 +21,16 @@ def fetch_edges_from_qdrant(
|
||||
) -> List[Dict]:
|
||||
"""
|
||||
Holt Edges aus der Datenbank basierend auf Seed-IDs.
|
||||
Filtert auf source_id, target_id oder note_id.
|
||||
WP-15c: Erhält alle Metadaten für das Note-Level Diversity Pooling.
|
||||
"""
|
||||
if not seeds or limit <= 0:
|
||||
return []
|
||||
|
||||
# Konsistente Namensauflösung via database-Paket
|
||||
# Rückgabe: (notes_col, chunks_col, edges_col)
|
||||
_, _, edges_col = collection_names(prefix)
|
||||
|
||||
# Wir suchen Kanten, bei denen die Seed-IDs entweder Quelle, Ziel oder Kontext-Note sind.
|
||||
seed_conditions = []
|
||||
for field in ("source_id", "target_id", "note_id"):
|
||||
for s in seeds:
|
||||
@@ -35,6 +39,7 @@ def fetch_edges_from_qdrant(
|
||||
)
|
||||
seeds_filter = rest.Filter(should=seed_conditions) if seed_conditions else None
|
||||
|
||||
# Optionaler Filter auf spezifische Kanten-Typen (z.B. für Intent-Routing)
|
||||
type_filter = None
|
||||
if edge_types:
|
||||
type_conds = [
|
||||
@@ -52,6 +57,7 @@ def fetch_edges_from_qdrant(
|
||||
flt = rest.Filter(must=must) if must else None
|
||||
|
||||
# Abfrage via Qdrant Scroll API
|
||||
# WICHTIG: with_payload=True lädt alle Metadaten (target_section, provenance etc.)
|
||||
pts, _ = client.scroll(
|
||||
collection_name=edges_col,
|
||||
scroll_filter=flt,
|
||||
@@ -60,4 +66,6 @@ def fetch_edges_from_qdrant(
|
||||
with_vectors=False,
|
||||
)
|
||||
|
||||
# Wir geben das vollständige Payload zurück, damit der Retriever
|
||||
# alle Signale für die Super-Edge-Aggregation und das Scoring hat.
|
||||
return [dict(p.payload) for p in pts if p.payload]
|
||||
Reference in new issue
Block a user