Error loading a json serialized file into a GraphDocument

[gedman4b]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

import json
from langchain_core.load import loads

file = 'graph_documents/g1.json'
with open(file,'r') as content:
    string_representation = json.load(content)
graph_document = loads(string_representation)

Description

I have serialized a LangChain GraphDocument and saved it as a json file.

I now wish to load that file back into a LangChain Graph Document. (see my code above)

When I try to dot his I am getting a NotImplementedError:

NotImplementedError: Trying to load an object that doesn't implement serialization: {'lc': 1, 'type': 'not_implemented', 'id': ['langchain_community', 'graphs', 'graph_document', 'GraphDocument'], ........

I then tried to use your ChatLangChain and it responded with a suggestion the same as the code I came up with. The Chat response:

**To load a serialized JSON file into a LangChain GraphDocument, you can follow these steps:

• Use the loads function from the langchain_core.load module to deserialize the JSON string representation.
• Specify a secrets_map if your serialized object contains sensitive information, such as API keys.

Example code snippet:
from langchain_core.load import loads

with open("/path/to/your/serialized_file.json", "r") as fp:
serialized_data = json.load(fp)

graph_document = loads(serialized_data, secrets_map={"API_KEY": "your-api-key"})
This will instantiate the GraphDocument from the serialized JSON data.**

How can I get this code to work?

System Info

pip freeze | grep langchain
langchain==0.2.12
langchain-chroma==0.1.3
langchain-community==0.2.11
langchain-core==0.2.28
langchain-experimental==0.0.64
langchain-openai==0.1.20
langchain-text-splitters==0.2.2

Platform windows
Python 3.11.2

[dosubot[bot]]

Hey @gedman4b! I'm here to help you with your coding issues. I'm a bot designed to assist with bug fixes, answer questions, and guide you towards becoming a contributor. Let's tackle this challenge together!

To load a serialized JSON file into a LangChain GraphDocument without encountering the NotImplementedError, you should use the LLMGraphTransformer from the langchain_experimental.graph_transformers module. Here is an example of how to do it:

Installation

First, install the necessary packages:

%pip install -qU langchain_community langchain_experimental jq 

Loading JSON File

Now, you can use the LLMGraphTransformer to load your JSON file:

from langchain_experimental.graph_transformers import LLMGraphTransformer
from langchain_core.documents import Document
from langchain_openai import ChatOpenAI

# Initialize the language model
llm = ChatOpenAI(temperature=0)

# Create the transformer
transformer = LLMGraphTransformer(
    llm=llm,
    allowed_nodes=["Person", "Organization"],
    allowed_relationships=["WORKS_FOR", "HAS_AWARD", "PRODUCED_BY", "HAS_CHARACTERISTIC"]
)

# Load the JSON file into a Document
with open("/path/to/your/serialized_file.json", "r") as file:
    json_content = file.read()

doc = Document(page_content=json_content)

# Convert to graph documents
graph_documents = transformer.convert_to_graph_documents([doc])
print(graph_documents[0])

This code snippet reads the JSON file from disk and loads it into a LangChain object using the LLMGraphTransformer. You can customize the allowed_nodes and allowed_relationships parameters as needed to fit the structure of your JSON file ^[1][2][3].

To continue talking to Dosu, mention @dosu.

[gedman4b]

@dosu I want to be able to convert the Document into a GraphDocument without using an LLM.

[dosubot[bot]]

To convert a Document into a GraphDocument without using an LLM, you can use the RelikGraphTransformer class. Here is how you can do it:

1. Ensure you have the relik package installed.
2. Create an instance of RelikGraphTransformer.
3. Use the process_document method to convert a Document into a GraphDocument.

Here is a sample code snippet:

from langchain_core.documents import Document
from langchain_community.graphs.graph_document import GraphDocument
from langchain_experimental.graph_transformers.relik import RelikGraphTransformer

# Ensure you have the relik package installed
# pip install relik

# Create a sample Document
sample_document = Document(page_content="Your document content here")

# Initialize the RelikGraphTransformer
transformer = RelikGraphTransformer()

# Convert the Document to a GraphDocument
graph_document = transformer.process_document(sample_document)

# Now, graph_document contains the nodes and relationships extracted from the sample_document
print(graph_document)

This code will convert the sample_document into a GraphDocument using the RelikGraphTransformer without relying on an LLM ^[1].

To continue talking to Dosu, mention @dosu.