Hey I have a problem of saving

It's a tad janky.

All sub-indexes should share the same storage context object.

Then you can just call persist on the root

https://gpt-index.readthedocs.io/en/latest/core_modules/data_modules/index/composability.html#optional-persisting-the-graph

Thank you for helping me out. The problem is that I have been doing the same, but still got the problem

Are you able to share the code for this?

I can share a partial of it

We have a unit-test in the code for saving-loading a graph, it may also be helpful
https://github.com/jerryjliu/llama_index/blob/main/tests/indices/test_loading_graph.py

Thank you. I appreciate that!

rather than setting the ID attribute directly, maybe try using the setter instead.

graph.root_index.set_index_id("my_id")

Also, just double checking that all subindexes are sharing the same storage context? 🤞

I realize this is pretty janky. It's actually a task for this week to make this a little smoother 🙂

Yes, they all share the same storage_context

Haha. thank you for your effort in developing this for a better

I wonder if there is a remote storage for this, so that i don't need to rely on file system

Oops I still got the same issue even though i already use this graph.root_index.set_index_id

here is the index.json in the storage

You can change the backend using fsspec, to use amazon s3 for example
https://gpt-index.readthedocs.io/en/latest/core_modules/data_modules/storage/save_load.html#using-a-remote-backend

I'm very confused why this isn't working 🤔

Writing a full example to confirm it works haha

would it be alright if i could dm you and share the code ? it maybe the case that my logic has the problem

yea go for it 👍

oh I see. thank you.