Thread

So that entire _note_content field is what gets used to re-create the node

I'd be really surprsied if the retrieved nodes don't include all that info

metadata_dict_to_node() should be converting the whole thing

unless it's silently hitting the try/except for some reason

oh..

there's no metadata in there in the actual db..

i wonder how i'm making my text nodes..

Hmmm your screenshot definitely has metadata (file_name and answer fields) -- unless thats from a different db lol

in the relationship it has the metadata

oh right!

i think

You can still access that on the source nodes, just less intuitive

I agree it should be on the actual nodes

yeah, but the actual node itself should have it

i don't think that's a llama issue though

it's strange

because both the node in question and its related node were made in the exact same way

solved @Logan M

using a helper function called build_nodes_from_splits to generate my nodes

unless i'm missing something, it does not add the metadata from the document

ah yea, that's an option in the node parser class (for some reason)

rip

ty

(as always)

:dotsCATJAM: