File size: 1,336 Bytes
95183e4 4ab96ab 95183e4 73e7897 95183e4 4ab96ab 95183e4 4ab96ab 95183e4 4ab96ab |
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 |
from typing import Any, Dict, List
from haystack.schema import Document
from fastrag.retrievers import QuantizedBiEncoderRetriever
class EndpointHandler:
def __init__(self, path=""):
model_id = "Intel/bge-small-en-v1.5-rag-int8-static"
self.retriever = QuantizedBiEncoderRetriever(embedding_model=model_id)
def __call__(self, data: Dict[str, Any]) -> List[Dict[str, Any]]:
queries = data.get("queries", None)
documents = data.get("documents", None)
if queries is not None:
assert isinstance(queries, list), "Expected queries to be a list"
assert all(
isinstance(query, str) for query in queries
), "Expected each query in queries to be a string"
return self.retriever.embed_queries(queries=queries)
elif documents is not None:
assert isinstance(documents, list), "Expected documents to be a list"
assert all(
isinstance(document, dict) for document in documents
), "Expected each document in documents to be a dictionary"
documents = [Document.from_dict(document) for document in documents]
return self.retriever.embed_documents(documents=documents)
else:
raise ValueError("Expected either queries or documents")
|