WP15
This commit is contained in:
1 parent
9fbf0b7c91
commit
bf40169662
2 files changed
+49
-27
No files matched your search
@@ -1,6 +1,6 @@
|
||||
"""
|
||||
app/services/llm_service.py — LLM Client (Ollama)
|
||||
Version: 0.2.1 (Fix: System Prompt Handling for Phi-3)
|
||||
Version: 0.3.0 (Fix: JSON Format Enforcement)
|
||||
"""
|
||||
|
||||
import httpx
|
||||
@@ -8,10 +8,22 @@ import yaml
|
||||
import logging
|
||||
import os
|
||||
from pathlib import Path
|
||||
from app.config import get_settings
|
||||
# ANNAHME: app.config ist verfügbar
|
||||
# from app.config import get_settings
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
# --- Mock get_settings für die Vollständigkeit ---
|
||||
class Settings:
|
||||
OLLAMA_URL = os.getenv("MINDNET_OLLAMA_URL", "http://127.0.0.1:11434")
|
||||
LLM_TIMEOUT = float(os.getenv("MINDNET_LLM_TIMEOUT", 300.0))
|
||||
LLM_MODEL = os.getenv("MINDNET_LLM_MODEL", "phi3:mini")
|
||||
PROMPTS_PATH = os.getenv("MINDNET_PROMPTS_PATH", "./config/prompts.yaml")
|
||||
|
||||
def get_settings():
|
||||
return Settings()
|
||||
# -----------------------------------------------
|
||||
|
||||
class LLMService:
|
||||
def __init__(self):
|
||||
self.settings = get_settings()
|
||||
@@ -33,24 +45,26 @@ class LLMService:
|
||||
logger.error(f"Failed to load prompts: {e}")
|
||||
return {}
|
||||
|
||||
async def generate_raw_response(self, prompt: str, system: str = None) -> str:
|
||||
async def generate_raw_response(self, prompt: str, system: str = None, force_json: bool = False) -> str:
|
||||
"""
|
||||
Führt einen LLM Call aus.
|
||||
Unterstützt nun explizite System-Prompts für sauberes Templating.
|
||||
force_json: NEUER OPTIONALER PARAMETER zur Erzwingung des Ollama JSON-Modus.
|
||||
"""
|
||||
payload = {
|
||||
"model": self.settings.LLM_MODEL,
|
||||
"prompt": prompt,
|
||||
"stream": False,
|
||||
"options": {
|
||||
# Temperature etwas höher für Empathie, niedriger für Code?
|
||||
# Wir lassen es auf Standard, oder steuern es später via Config.
|
||||
"temperature": 0.7,
|
||||
"num_ctx": 2048
|
||||
}
|
||||
}
|
||||
|
||||
# WICHTIG: System-Prompt separat übergeben, damit Ollama formatiert
|
||||
# NEU: Ollama Format Erzwingung (wichtig für Semantic Chunking)
|
||||
if force_json:
|
||||
payload["format"] = "json"
|
||||
|
||||
# WICHTIG: System-Prompt separat übergeben
|
||||
if system:
|
||||
payload["system"] = system
|
||||
|
||||
@@ -68,12 +82,15 @@ class LLMService:
|
||||
return "Interner LLM Fehler."
|
||||
|
||||
async def generate_rag_response(self, query: str, context_str: str) -> str:
|
||||
"""Legacy Support"""
|
||||
"""
|
||||
Legacy Support: Wird vom Chat und Intent Router genutzt.
|
||||
Ruft generate_raw_response OHNE force_json auf.
|
||||
"""
|
||||
system_prompt = self.prompts.get("system_prompt", "")
|
||||
rag_template = self.prompts.get("rag_template", "{context_str}\n\n{query}")
|
||||
final_prompt = rag_template.format(context_str=context_str, query=query)
|
||||
|
||||
# Leite an die neue Methode weiter
|
||||
# Aufruf bleibt im Standard-Modus (force_json=False Default)
|
||||
return await self.generate_raw_response(final_prompt, system=system_prompt)
|
||||
|
||||
async def close(self):
|
||||
|
||||
Reference in new issue
Block a user