e34b1e4c6d
Refactor the `kotaemon/pipelines` module to `kotaemon/indices`. Create the VectorIndex. Note: currently I place `qa` to be inside `kotaemon/indices` since at the moment we only have `qa` in RAG. At the same time, I think `qa` can be an independent module in `kotaemon/qa`. Since this can be changed later, I still go at the 1st option for now to observe if we can change it later.
39 lines
1.2 KiB
Python
39 lines
1.2 KiB
Python
import tempfile
|
|
from typing import List
|
|
|
|
from theflow.utils.modules import ObjectInitDeclaration as _
|
|
|
|
from kotaemon.base import BaseComponent
|
|
from kotaemon.embeddings import AzureOpenAIEmbeddings
|
|
from kotaemon.indices import VectorRetrieval
|
|
from kotaemon.llms.completions.openai import AzureOpenAI
|
|
from kotaemon.storages import ChromaVectorStore
|
|
|
|
|
|
class Pipeline(BaseComponent):
|
|
llm: AzureOpenAI = AzureOpenAI.withx(
|
|
openai_api_base="https://test.openai.azure.com/",
|
|
openai_api_key="some-key",
|
|
openai_api_version="2023-03-15-preview",
|
|
deployment_name="gpt35turbo",
|
|
temperature=0,
|
|
request_timeout=60,
|
|
)
|
|
|
|
retrieving_pipeline: VectorRetrieval = VectorRetrieval.withx(
|
|
vector_store=_(ChromaVectorStore).withx(path=str(tempfile.mkdtemp())),
|
|
embedding=AzureOpenAIEmbeddings.withx(
|
|
model="text-embedding-ada-002",
|
|
deployment="embedding-deployment",
|
|
openai_api_base="https://test.openai.azure.com/",
|
|
openai_api_key="some-key",
|
|
),
|
|
)
|
|
|
|
def run_raw(self, text: str) -> str:
|
|
matched_texts: List[str] = self.retrieving_pipeline(text)
|
|
return self.llm("\n".join(matched_texts)).text
|
|
|
|
def run(self):
|
|
...
|