Failing to retrieve documents once the

Failing to retrieve documents once the app is reloads storage context

Plain Text

redis_client = redis.Redis(
                    host=self.config.get("REDIS_HOST"),
                    
port=self.config.get("REDIS_PORT"),
                    password=self.config.get("REDIS_PASSWORD"),
                    ssl=True,
                )
docstore = RedisDocumentStore.from_redis_client(
                    redis_client=redis_client,
                    namespace=namespace
                )
storage_context = StorageContext.from_defaults(
                    docstore=self.docstore,
                    index_store=RedisIndexStore.from_redis_client(
                        redis_client=redis_client,
                        namespace=namespace
                    ),
                )

after this I reload the index from the storage context like below:
base_index = load_index_from_storage(storage_context)

when setting this up and loading the index like after I close my app:

Plain Text

try:
 indices = load_indices_from_storage(self.storage_context)
                   for index in indices:
                        print(index.index_id)
                    base_index = load_index_from_storage(storage_context)
                    print("[INFO] Index found at storage")
                except ValueError as e:
                    print("[INFO] No index found at storage")
                    base_index = VectorStoreIndex(
                        nodes=[],               storage_context=storage_context)

and the retrieval is defined as:

Plain Text

base_retriever = base_index.as_retriever(
                similarity_top_k=self.similarity_top_k
            )
            retriever = AutoMergingRetriever(
                base_retriever, self.storage_context, verbose=True
            )
              query_bundle = QueryBundle(query_str=query)
            retrived_nodes = retriever.retrieve(query_bundle)

this outputs empty retrieved nodes for some reason.

49 comments

LLogan M

I replied on github, but you seem to missing configuration for a vector store in your storage context 👀

bbeaverTango

Just saw

bbeaverTango

Sorry must've missed it

bbeaverTango

😅

bbeaverTango

@Logan M Could I use any vector store?

bbeaverTango

I tried using the legacy RedisVectorStore first but the primary disadvantage is my redis instance is on upsert and since the legacy initialization requires a redis url and not the redis client that I've been using for docstore and index store:

Plain Text

self.redis_client = redis.Redis(
                    host=self.config.get("UPSTASH_REDIS_HOST"),
                    port=self.config.get("UPSTASH_REDIS_PORT"),
                    password=self.config.get("UPSTASH_REDIS_PASSWORD"),
                    ssl=True,
                )

so the Redis Vector Store returns

MODULE not supported

bbeaverTango

and I also tried the UpstashVectorStore

LLogan M

you can use any vector store indeed 👀

bbeaverTango

this throws a type error for some reason

bbeaverTango

Plain Text

Can't instantiate abstract class BaseNode with abstract methods get_content, get_metadata_str, get_type, hash, set_content (type=type_error)

bbeaverTango

Here the storage context was:

Plain Text

self.redis_client = redis.Redis(
                    host=self.config.get("UPSTASH_REDIS_HOST"),
                    port=self.config.get("UPSTASH_REDIS_PORT"),
                    password=self.config.get("UPSTASH_REDIS_PASSWORD"),
                    ssl=True,
                )
                self.docstore = RedisDocumentStore.from_redis_client(
                    redis_client=self.redis_client,
                    namespace=self.namespace
                )
                self.vector_store = UpstashVectorStore(
                    url=self.config.get("UPSTASH_VECTOR_URL"),
                    token=self.config.get("UPSTASH_VECTOR_TOKEN"),
                )

                self.storage_context = StorageContext.from_defaults(
                    docstore=self.docstore,
                    index_store=RedisIndexStore.from_redis_client(
                        redis_client=self.redis_client,
                        namespace=self.namespace
                    ),
                    vector_store=self.vector_store
                )

LLogan M

Oooo thats a venv issue 👀 Or at least 90% sure it is. Or you are mixing legacy imports with other newer imports

bbeaverTango

how do I fix it?

bbeaverTango

from llama_index.legacy.vector_stores.upstash import UpstashVectorStore

bbeaverTango

this is the import for Upstash right?

bbeaverTango

Was using a similar one for redis too:

from llama_index.legacy.vector_stores.redis import RedisVectorStore

LLogan M

Plain Text

pip install llama-index-vector-stores-upstash

from llama_index.vector_stores.uptash import UpstashVectorStore

bbeaverTango

so once I do this

LLogan M

it should work 🙂

bbeaverTango

Plain Text

try:
                    indices = load_indices_from_storage(self.storage_context)
                    for index in indices:
                        print(index.index_id)
                    self.base_index = load_index_from_storage(self.storage_context)
                    print(self.base_index)
                    print("[INFO] Index found at storage")
                except ValueError as e:
                    print("[INFO] No index found at storage")
                    self.base_index = VectorStoreIndex(
                        nodes=[],
                        storage_context=self.storage_context,
                    )

bbeaverTango

the loading would look somewhat like this right>

bbeaverTango

the rest would be auto fetched from storage context

LLogan M

Yea pretty much! Although since you are using a vector store integration, normally all the storage is put into the vector store to simplify, so you'd want to override that with this

bbeaverTango

pip install llama-index-vector-stores-upstash

Is there something like this even for the newer redis store?

bbeaverTango

😅

LLogan M

VectorStoreIndex(...., store_nodes_override=True)

bbeaverTango

LLogan M

There is a redis one, but I'm not sure it will solve your issue (it seems like you want to be able to pass the client in right?)

bbeaverTango

yeah

bbeaverTango

I formed the rediss url manually seeing my redis client initialization :

Plain Text

self.redis_client = redis.Redis(
                    host=self.config.get("UPSTASH_REDIS_HOST"),
                    port=self.config.get("UPSTASH_REDIS_PORT"),
                    password=self.config.get("UPSTASH_REDIS_PASSWORD"),
                    ssl=True,
                )

but that seems to throw an error from the upstash end

bbeaverTango

btw is there a list for these newer imports like the upstash one you shared?

bbeaverTango

Would love to have one handy

LLogan M

There is this package registry
https://discord.com/channels/1059199217496772688/1073670729054294197/1207845501660168232

We are also in the process of updating https://llamahub.ai -- most packages are there, each card should show the install and import (we need to flesh this out more over time)

bbeaverTango

did the pip install

pip install llama-index-vector-stores-upstash

bbeaverTango

the import still seems to fail :

from llama_index.vector_stores.uptash import UpstashVectorStore

LLogan M

Typo: from llama_index.vector_stores.upstash import UpstashVectorStore

LLogan M

I think I wrote that earlier and then edited haha

😄

Seems alright now

let me check

works awesome!!! 🙌

couldn't sleep last night because of this vector store integration 😄

bbeaverTango

Thank you @Logan M

bbeaverTango

Is there some visualisation and tacking integration that I can add that better helps visualise the embeddings created and the end to end chunking and retrieval logic?

LLogan M

hmm, not quite at the moment! There are observability integrations (like arize) which will show traces, but not quite the same I think

bbeaverTango

oh, okay

bbeaverTango

Something to look forward to then

Add a reply

Find answers from the community

Failing to retrieve documents once the