50 lines
1.2 KiB
Python
50 lines
1.2 KiB
Python
"""
|
|
BQAS - Breakpilot Quality Assurance System
|
|
|
|
LLM-based quality assurance framework for voice service with:
|
|
- LLM Judge (Qwen2.5-32B based evaluation)
|
|
- RAG Judge (Specialized RAG/Correction evaluation)
|
|
- Synthetic Test Generation
|
|
- Golden Test Suite
|
|
- Regression Tracking
|
|
- Automated Backlog Generation
|
|
- Local Scheduler (Alternative zu GitHub Actions)
|
|
"""
|
|
|
|
from bqas.judge import LLMJudge, JudgeResult
|
|
from bqas.rag_judge import (
|
|
RAGJudge,
|
|
RAGRetrievalResult,
|
|
RAGOperatorResult,
|
|
RAGHallucinationResult,
|
|
RAGPrivacyResult,
|
|
RAGNamespaceResult,
|
|
)
|
|
from bqas.metrics import BQASMetrics, TestResult
|
|
from bqas.config import BQASConfig
|
|
from bqas.runner import BQASRunner, get_runner, TestRun
|
|
|
|
# Notifier wird separat importiert (keine externen Abhaengigkeiten)
|
|
# Nutzung: from bqas.notifier import BQASNotifier, Notification, NotificationConfig
|
|
|
|
__all__ = [
|
|
# Intent Judge
|
|
"LLMJudge",
|
|
"JudgeResult",
|
|
# RAG Judge
|
|
"RAGJudge",
|
|
"RAGRetrievalResult",
|
|
"RAGOperatorResult",
|
|
"RAGHallucinationResult",
|
|
"RAGPrivacyResult",
|
|
"RAGNamespaceResult",
|
|
# Metrics & Config
|
|
"BQASMetrics",
|
|
"TestResult",
|
|
"BQASConfig",
|
|
# Runner
|
|
"BQASRunner",
|
|
"get_runner",
|
|
"TestRun",
|
|
]
|