Class LangChainEmbeddingHelper

embedding/embeddingHelper.py:546–570

@singleton
class LangChainEmbeddingHelper:
    def __init__(self):
        # Map{dbId, db}
        self.dbMap = {}

    def from_documents(self, documents: list[Document], dbId: str, embedding=OpenAIEmbeddings(),
                       persist_directory: str = None):
        if persist_directory is None:
            persist_directory = "./embedding" + dbId
        db = Chroma.from_documents(documents, embedding=embedding, persist_directory=persist_directory)
        self.dbMap[dbId] = db

    def from_texts(self, texts: list[str], dbId: str, embedding=OpenAIEmbeddings(), persist_directory: str = None):
        if persist_directory is None:
            persist_directory = "./embedding" + dbId
        db = Chroma.from_texts(texts, embedding=embedding, persist_directory=persist_directory)
        self.dbMap[dbId] = db

    def similarity_search(self, query: str, fileID: str, topK: int = 4) -> list[str]:
        db = self.dbMap[fileID]
        docs = db.similarity_search(query, k=topK)
        texts = []
        for doc in docs:
            texts.append(doc.page_content)
        return texts


@singleton