[Question]: Imports no longer working

[RonRastorguev]

Question Validation

• [X] I have searched both the documentation and discord for an answer.

Question

I am on llamaindex version 0.10.30. I have a lot of import statements:

`from llama_index.llms.openai import OpenAI

from llama_index.core.indices.vector_store.base import VectorStoreIndex

from llama_index.core.storage import StorageContext

from llama_index.core.indices import load_index_from_storage

from llama_index.core.service_context import ServiceContext

from llama_index.core.readers import SimpleDirectoryReader, download_loader

from llama_index.core import ChromaVectorStore

from llama_index.embeddings.huggingface import HuggingFaceEmbedding

from llama_index.core.retrievers import VectorIndexRetriever

from llama_index.core.query_engine import RetrieverQueryEngine

from llama_index.core.chat_engine import CondensePlusContextChatEngine

from llama_index.core.node_parser import SentenceSplitter

from llama_index.core.types import ChatMessage

from llama_index.core.vector_stores.types import MetadataFilters

from llama_index.core.vector_stores.types import MetadataFilter

from llama_index.postprocessor.colbert_rerank import ColbertRerank

from llama_index.core import base_query_engine`

The imports are different than those in examples because for some reason when I do follow the examples, for instance importing the VectorStoreIndex by doing from llama_index.core import VectorStoreIndex I get the error: ImportError: cannot import name 'VectorStoreIndex' from 'llama_index.core' (unknown location)

Using these import statements were the only way I could make the text not white, and not cause the flagging by the IDE. When I run the code now it says: from llama_index.core.image_retriever import BaseImageRetriever ModuleNotFoundError: No module named 'llama_index.core.image_retriever'

I never use the BaseImageRetriever though.

How do I go about fixing this?

[logan-markewich]

Seems like a botched update from v0.9.x

Just install with a fresh venv, should be fine

[KanrongYu]

had the same issue with you. Though was able to fix some of these, e.g. changing the import to from llama_index.legacy import VectorStoreIndex, but this seems non-sustainable. In the end, do what @logan-markewich suggests and no issue at all when running sample code in the new venv