hi there! how can i convert llama

aalvarojauna

hi there! how can i convert llama_index.schema.TextNode
to str or list?

31 comments

LLogan M

do you want just the node text? Or the entire structure?

aalvarojauna

the entire structure

aalvarojauna

i have been trying with a lot of ways but nothing @Logan M

LLogan M

node.json()

LLogan M

will convert to string

LLogan M

then TextNode.parse_raw(json_str) will recreate it

aalvarojauna

Attachment

aalvarojauna

Attachment

aalvarojauna

it appears that i cant in that way

aalvarojauna

Attachment

LLogan M

right, its a list

nodes_str = [x.json() for x in nodes]

aalvarojauna

but in that case it will be a list and i need in string

aalvarojauna

if now i do str(nodes_str) i will have a lot of \

LLogan M

json.dumps(nodes_str)

aalvarojauna

Attachment

aalvarojauna

what i was talking about

aalvarojauna

i need the same way but in string format

aalvarojauna

or json

aalvarojauna

what i want to do it is save it in a azure container, for that i need to convert it to string

aalvarojauna

it cant be list

LLogan M

This works fine for me, ez pz

Plain Text

>>> from llama_index.schema import TextNode
>>> nodes = [TextNode(text="test1"), TextNode(text="test2")]
>>> import json
>>> nodes_str = [x.dict() for x in nodes]
>>> data = json.dumps(nodes_str)
>>> data
'[{"id_": "d44c4b9c-5803-4d32-95a7-414b5a520138", "embedding": null, "metadata": {}, "excluded_embed_metadata_keys": [], "excluded_llm_metadata_keys": [], "relationships": {}, "text": "test1", "start_char_idx": null, "end_char_idx": null, "text_template": "{metadata_str}\\n\\n{content}", "metadata_template": "{key}: {value}", "metadata_seperator": "\\n", "class_name": "TextNode"}, {"id_": "001fea50-26cf-4a13-9648-b56a206d9a99", "embedding": null, "metadata": {}, "excluded_embed_metadata_keys": [], "excluded_llm_metadata_keys": [], "relationships": {}, "text": "test2", "start_char_idx": null, "end_char_idx": null, "text_template": "{metadata_str}\\n\\n{content}", "metadata_template": "{key}: {value}", "metadata_seperator": "\\n", "class_name": "TextNode"}]'
>>> nodes_str = json.loads(data)
>>> nodes = [TextNode.parse_obj(x) for x in nodes_str]
>>> nodes[0].text
'test1'
>>>

aalvarojauna

yeah this works! here is x.dict instead of x.json

aalvarojauna

thanks

aalvarojauna

Attachment

aalvarojauna

😭

aalvarojauna

i just want to save the nodes using cloud and then recover it as it was

LLogan M

you need to iterate

LLogan M

nodes_str = json.loads(data)
nodes = [TextNode.parse_obj(x) for x in nodes_str]

LLogan M

If you really wanted, we have various storage layers that automate all this

Docstores (mongodb, redis, to disk, postgres, firestore, dynamodb)

Vector Stores (basically any vector store you can think of)

aalvarojauna

thanks but not👍 now it works

aalvarojauna

maybe i will mention you in my Bachelor's Thesis😅

Add a reply

Find answers from the community

hi there! how can i convert llama_index.