124 lines
3.3 KiB
YAML
124 lines
3.3 KiB
YAML
version: 1.3 # Update für Smart Edge Allocation
|
|
|
|
# --- CHUNKING DEFINITIONEN ---
|
|
# Hier definieren wir die technischen Strategien und den Smart Edge Filter.
|
|
chunking_profiles:
|
|
|
|
# 1. Standard Profile (Sliding Window, KEIN LLM-Filter)
|
|
sliding_short:
|
|
strategy: sliding_window
|
|
enable_smart_edge_allocation: false # Sekundärverfeinerung deaktiviert
|
|
target: 200
|
|
max: 350
|
|
overlap: [30, 50]
|
|
|
|
# 2. Smart Edge Allocation Profile (Sliding Window + LLM-Filter)
|
|
sliding_smart_edges:
|
|
strategy: sliding_window # Primärzerlegung: Sliding Window
|
|
enable_smart_edge_allocation: true # SEKUNDÄRVERFEINERUNG: LLM-Filter aktiv
|
|
target: 400
|
|
max: 600
|
|
overlap: [50, 80]
|
|
|
|
# 3. Strukturierte Profile (By Heading, KEIN LLM-Filter)
|
|
structured_strict:
|
|
strategy: by_heading
|
|
enable_smart_edge_allocation: false
|
|
split_level: 2
|
|
max: 600
|
|
target: 400
|
|
overlap: [50, 80]
|
|
|
|
# 4. Strukturierte Profile (By Heading + LLM-Filter)
|
|
structured_smart_edges:
|
|
strategy: by_heading # Primärzerlegung: Harte Trennung
|
|
enable_smart_edge_allocation: true # SEKUNDÄRVERFEINERUNG: LLM-Filter aktiv
|
|
split_level: 2
|
|
max: 600
|
|
target: 400
|
|
overlap: [50, 80]
|
|
|
|
defaults:
|
|
retriever_weight: 1.0
|
|
chunking_profile: sliding_standard # Fallback Profil
|
|
edge_defaults: []
|
|
|
|
types:
|
|
# --- WISSENSBAUSTEINE ---
|
|
concept:
|
|
chunking_profile: sliding_smart_edges # Nutzt Kantenfilterung
|
|
retriever_weight: 0.60
|
|
edge_defaults: ["references", "related_to"]
|
|
|
|
source:
|
|
chunking_profile: sliding_short # Kein LLM-Filter
|
|
retriever_weight: 0.50
|
|
edge_defaults: []
|
|
|
|
glossary:
|
|
chunking_profile: sliding_short
|
|
retriever_weight: 0.40
|
|
edge_defaults: ["related_to"]
|
|
|
|
# --- IDENTITÄT & PERSÖNLICHKEIT ---
|
|
profile:
|
|
chunking_profile: structured_smart_edges # Strukturiert + Kantenfilterung
|
|
retriever_weight: 0.70
|
|
edge_defaults: ["references", "related_to"]
|
|
|
|
value:
|
|
chunking_profile: structured_smart_edges
|
|
retriever_weight: 1.00
|
|
edge_defaults: ["related_to"]
|
|
|
|
principle:
|
|
chunking_profile: structured_smart_edges
|
|
retriever_weight: 0.95
|
|
edge_defaults: ["derived_from", "references"]
|
|
|
|
belief:
|
|
chunking_profile: sliding_short
|
|
retriever_weight: 0.90
|
|
edge_defaults: ["related_to"]
|
|
|
|
experience:
|
|
chunking_profile: sliding_smart_edges
|
|
retriever_weight: 0.90
|
|
edge_defaults: ["derived_from", "references"]
|
|
|
|
# --- STRATEGIE & ENTSCHEIDUNG ---
|
|
goal:
|
|
chunking_profile: sliding_smart_edges
|
|
retriever_weight: 0.95
|
|
edge_defaults: ["depends_on", "related_to"]
|
|
|
|
decision:
|
|
chunking_profile: structured_smart_edges
|
|
retriever_weight: 1.00
|
|
edge_defaults: ["caused_by", "references"]
|
|
|
|
risk:
|
|
chunking_profile: sliding_short
|
|
retriever_weight: 0.85
|
|
edge_defaults: ["related_to", "blocks"]
|
|
|
|
milestone:
|
|
chunking_profile: sliding_short
|
|
retriever_weight: 0.70
|
|
edge_defaults: ["related_to", "part_of"]
|
|
|
|
# --- OPERATIV ---
|
|
project:
|
|
chunking_profile: sliding_smart_edges
|
|
retriever_weight: 0.97
|
|
edge_defaults: ["references", "depends_on"]
|
|
|
|
task:
|
|
chunking_profile: sliding_short
|
|
retriever_weight: 0.80
|
|
edge_defaults: ["depends_on", "part_of"]
|
|
|
|
journal:
|
|
chunking_profile: sliding_smart_edges # Fließtext + Kantenfilterung
|
|
retriever_weight: 0.80
|
|
edge_defaults: ["references", "related_to"] |