Document class initialization difference between langchain-py and langchain-js

[phanitallapudi]

Why the Document from langchain_core.documents has page_content but in javascript it is defined pageContent, why not keep them same?

This is the initialization in javascript:

const docs = [
  new Document({ pageContent: "Harrison went to Harvard." }),
  new Document({ pageContent: "Ankush went to Princeton." }),
];

This is the initialization in python:

new_doc = Document(page_content=chunk, metadata=metadata)

[feijoes]

CamelCase is commonly used in JavaScript while snake_case is used in Python

[phanitallapudi]

is it possible to overwrite in python to use pageContent instead of page_content?

[feijoes]

@phanitallapudi Sure! If you want to modify the creation of the document, you can create your own implementation of the Document class's init method to accept pageContent instead of page_content. However, note that when accessing the document's data, you'll still need to use document.page_content:

from langchain_core.documents import Document

class CustomDocument(Document):
    def __init__(self, pageContent: str, **kwargs) -> None:
        super().__init__(page_content=pageContent, **kwargs) 
        
document = CustomDocument(
    pageContent="Hello, world!",
)

[phanitallapudi]

oh this is nice, thank you for this

[phanitallapudi]

@feijoes So this means wherever langchain using the Document, I need to create my own class which inherits the langchain functions and modify them with my custom document class correct?