vectorstore.py — langchain Source File

Source Code

"""Vectorstore stubs for the indexing api."""

from typing import Any

from langchain_core.document_loaders import BaseLoader
from langchain_core.documents import Document
from langchain_core.embeddings import Embeddings
from langchain_core.language_models import BaseLanguageModel
from langchain_core.vectorstores import VectorStore
from langchain_text_splitters import RecursiveCharacterTextSplitter, TextSplitter
from pydantic import BaseModel, ConfigDict, Field

from langchain_classic.chains.qa_with_sources.retrieval import (
    RetrievalQAWithSourcesChain,
)
from langchain_classic.chains.retrieval_qa.base import RetrievalQA


def _get_default_text_splitter() -> TextSplitter:
    """Return the default text splitter used for chunking documents."""
    return RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=0)


class VectorStoreIndexWrapper(BaseModel):
    """Wrapper around a `VectorStore` for easy access."""

    vectorstore: VectorStore

    model_config = ConfigDict(
        arbitrary_types_allowed=True,
        extra="forbid",
    )

    def query(
        self,
        question: str,
        llm: BaseLanguageModel | None = None,
        retriever_kwargs: dict[str, Any] | None = None,
        **kwargs: Any,
    ) -> str:
        """Query the `VectorStore` using the provided LLM.

        Args:
            question: The question or prompt to query.
            llm: The language model to use. Must not be `None`.
            retriever_kwargs: Optional keyword arguments for the retriever.
            **kwargs: Additional keyword arguments forwarded to the chain.

        Returns:
            The result string from the RetrievalQA chain.
        """
        if llm is None:
            msg = (
                "This API has been changed to require an LLM. "
                "Please provide an llm to use for querying the vectorstore.\n"
                "For example,\n"
                "from langchain_openai import OpenAI\n"
                "model = OpenAI(temperature=0)"
            )
            raise NotImplementedError(msg)
// ... (212 more lines)