komplett neues WP20 deployment

This commit is contained in:
Lars committed 2025-12-23 18:51:12 +01:00
1 parent f1bfa40b5b
commit 0157faab89
7 files changed
+203 -137

No files matched your search

+22 -10
View File
@@ -2,8 +2,11 @@
FILE: app/config.py
DESCRIPTION: Zentrale Pydantic-Konfiguration. Enthält Parameter für Qdrant,
Embeddings, Ollama, Google GenAI und OpenRouter.
VERSION: 0.6.0 (WP-20 Hybrid & OpenRouter Integration)
WP-20: Optimiert für Hybrid-Cloud Modus und Vektor-Synchronisation.
WP-22: Integration von Change-Detection und Vocab-Paths.
VERSION: 0.6.2
STATUS: Active
DEPENDENCIES: os, functools, pathlib
"""
from __future__ import annotations
import os
@@ -15,22 +18,27 @@ class Settings:
QDRANT_URL: str = os.getenv("QDRANT_URL", "http://127.0.0.1:6333")
QDRANT_API_KEY: str | None = os.getenv("QDRANT_API_KEY")
COLLECTION_PREFIX: str = os.getenv("MINDNET_PREFIX", "mindnet")
VECTOR_SIZE: int = int(os.getenv("MINDNET_VECTOR_SIZE", "384"))
# WP-22: Vektor-Dimension muss mit dem Embedding-Modell (nomic) übereinstimmen
VECTOR_SIZE: int = int(os.getenv("VECTOR_DIM", "768"))
DISTANCE: str = os.getenv("MINDNET_DISTANCE", "Cosine")
# --- Lokale Embeddings ---
# --- Lokale Embeddings (Ollama & Sentence-Transformers) ---
# Modell für die Vektorisierung (z.B. nomic-embed-text)
EMBEDDING_MODEL: str = os.getenv("MINDNET_EMBEDDING_MODEL", "nomic-embed-text")
# Legacy Fallback Modellname
MODEL_NAME: str = os.getenv("MINDNET_MODEL", "sentence-transformers/all-MiniLM-L6-v2")
# --- WP-20 Hybrid LLM Provider ---
# Optionen: "ollama" | "gemini" | "openrouter"
# Erlaubt: "ollama" | "gemini" | "openrouter"
MINDNET_LLM_PROVIDER: str = os.getenv("MINDNET_LLM_PROVIDER", "ollama").lower()
# Google AI Studio
# Google AI Studio (Direkt)
GOOGLE_API_KEY: str | None = os.getenv("GOOGLE_API_KEY")
GEMINI_MODEL: str = os.getenv("MINDNET_GEMINI_MODEL", "gemini-1.5-flash")
GEMMA_MODEL: str = os.getenv("MINDNET_GEMMA_MODEL", "gemma2-9b-it")
# Gemma-Modell für hohen Durchsatz bei der Ingestion
GEMMA_MODEL: str = os.getenv("MINDNET_GEMMA_MODEL", "google/gemma-2-9b-it:free")
# OpenRouter
# OpenRouter Integration
OPENROUTER_API_KEY: str | None = os.getenv("OPENROUTER_API_KEY")
OPENROUTER_MODEL: str = os.getenv("OPENROUTER_MODEL", "google/gemma-2-9b-it:free")
@@ -42,16 +50,20 @@ class Settings:
PROMPTS_PATH: str = os.getenv("MINDNET_PROMPTS_PATH", "config/prompts.yaml")
# --- WP-06 / WP-14 Performance & Last-Steuerung ---
LLM_TIMEOUT: float = float(os.getenv("MINDNET_LLM_TIMEOUT", "120.0"))
LLM_TIMEOUT: float = float(os.getenv("MINDNET_LLM_TIMEOUT", "300.0"))
DECISION_CONFIG_PATH: str = os.getenv("MINDNET_DECISION_CONFIG", "config/decision_engine.yaml")
BACKGROUND_LIMIT: int = int(os.getenv("MINDNET_LLM_BACKGROUND_LIMIT", "2"))
# --- System-Pfade ---
# --- System-Pfade & Ingestion-Logik ---
DEBUG: bool = os.getenv("DEBUG", "false").lower() == "true"
MINDNET_VAULT_ROOT: str = os.getenv("MINDNET_VAULT_ROOT", "./vault")
MINDNET_TYPES_FILE: str = os.getenv("MINDNET_TYPES_FILE", "config/types.yaml")
# WP-22: Dictionary für die Edge-Validierung
MINDNET_VOCAB_PATH: str = os.getenv("MINDNET_VOCAB_PATH", "vault/_system/dictionary/edge_vocabulary.md")
# WP-22: Change Detection (Modus: "full" oder "body")
CHANGE_DETECTION_MODE: str = os.getenv("MINDNET_CHANGE_DETECTION_MODE", "full")
# --- WP-04 Retriever Gewichte ---
# --- WP-04 Retriever Gewichte (Semantik vs. Graph) ---
RETRIEVER_W_SEM: float = float(os.getenv("MINDNET_WP04_W_SEM", "0.70"))
RETRIEVER_W_EDGE: float = float(os.getenv("MINDNET_WP04_W_EDGE", "0.25"))
RETRIEVER_W_CENT: float = float(os.getenv("MINDNET_WP04_W_CENT", "0.05"))