docs: how does Langflow run components?

[mendonk]

Feature Request

Documentation focused on more lower-level, under-the-hood Langflow.
Initially raised in Discord, using the Recursivecharactertextsplitter as example.

1. When and how does the get_data_input() method ever get called?

2. When and how does the build_text_splitter() method ever get called?

3. When does the text content found within the data input actually get split? The component's build_text_splitter() method only returns an object instance of the text_splitter. In the Langchain documentation examples, a text splitter object is first instantiated followed by calling that object's .split_text() or .split_document() method.

How is this component supposed to work within Langflow?

Motivation

Better explanation of how Langflow interacts with lower-level Python code and Pydantic models.

Your Contribution

No response