Spaces:
Sleeping
Sleeping
File size: 3,707 Bytes
af2f8e1 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 |
"""
RAGAS API Endpoints
"""
from fastapi import APIRouter, HTTPException
from pydantic import BaseModel
from typing import List, Optional
router = APIRouter()
# We'll set these from main.py
ragas_pipeline = None
ragas_evaluator = None
class RagasEvalRequest(BaseModel):
"""Direct evaluation request."""
query: str
answer: str
contexts: List[str]
class RagasQueryRequest(BaseModel):
"""Query + evaluate request."""
query: str
top_k: int = 3
def init_ragas_router(pipeline, evaluator):
"""Initialize router with pipeline and evaluator instances."""
global ragas_pipeline, ragas_evaluator
ragas_pipeline = pipeline
ragas_evaluator = evaluator
return router
@router.post("/evaluate")
async def evaluate_direct(request: RagasEvalRequest):
"""
Evaluate a query-answer pair directly.
Use this when you already have the answer and contexts.
"""
if not ragas_evaluator:
raise HTTPException(status_code=503, detail="RAGAS not initialized")
result = await ragas_evaluator.evaluate_single(
query=request.query,
answer=request.answer,
contexts=request.contexts
)
return {
"eval_id": result.eval_id,
"faithfulness": result.faithfulness,
"context_precision": result.context_precision,
"ragas_score": result.ragas_score,
"latency_ms": round(result.latency_ms, 2)
}
@router.post("/query-and-evaluate")
async def query_and_evaluate(request: RagasQueryRequest):
"""
Query the RAG system AND evaluate the response.
Returns both the answer and RAGAS metrics.
"""
if not ragas_pipeline or not ragas_evaluator:
raise HTTPException(status_code=503, detail="RAGAS not initialized")
# Step 1: Query pipeline
response = ragas_pipeline.query_for_evaluation(
query=request.query,
top_k=request.top_k
)
if response.status != "success":
return {
"query": response.query,
"answer": response.answer,
"status": response.status,
"ragas": None
}
# Step 2: Evaluate with RAGAS
eval_result = await ragas_evaluator.evaluate_single(
query=response.query,
answer=response.answer,
contexts=response.contexts
)
return {
"query": response.query,
"answer": response.answer,
"sources": response.sources,
"chunks_used": response.chunks_used,
"response_time_ms": round(response.response_time_ms, 2),
"ragas": {
"eval_id": eval_result.eval_id,
"faithfulness": eval_result.faithfulness,
"context_precision": eval_result.context_precision,
"ragas_score": eval_result.ragas_score,
"eval_time_ms": round(eval_result.latency_ms, 2)
}
}
@router.get("/metrics")
async def get_metrics():
"""Get aggregate RAGAS metrics from all evaluations."""
if not ragas_evaluator:
raise HTTPException(status_code=503, detail="RAGAS not initialized")
results = ragas_evaluator.results
if not results:
return {"total_evaluations": 0, "message": "No evaluations yet"}
# Calculate averages
avg_faithfulness = sum(r.faithfulness for r in results) / len(results)
avg_precision = sum(r.context_precision for r in results) / len(results)
avg_ragas = sum(r.ragas_score for r in results) / len(results)
return {
"total_evaluations": len(results),
"avg_faithfulness": round(avg_faithfulness, 3),
"avg_context_precision": round(avg_precision, 3),
"avg_ragas_score": round(avg_ragas, 3)
} |