Skip to content

Latest commit

 

History

History
68 lines (44 loc) · 2.93 KB

example_nesting.md

File metadata and controls

68 lines (44 loc) · 2.93 KB
Raw

Nested headings / branching conversation example

This example shows how nested headings represent branching conversations. Please read through it, noting the notes, and then see the diagram at the end.

  • In short, the plugin will always trace upwards, in a heading / hierarchy sense, starting from the heading containing the cursor up to the highest level containing heading.
  • Ensure that your conversation's starting message is H1 (single #). It can be ## but then there should be nothing above that.
  • Headings starting with "AI" or "Assistant" (case insensitive) are tagged as assistant messages, any other headings as user messages.
  • Embedded images (not links!) will be correctly base64 encoded and submitted together with text to the OpenAI-style API for multi-modal understanding.

First message: joke request

Can you tell me a very short joke about taking notes?

AI: first reply with joke (not so funny...)

Why don't programmers take handwritten notes? They prefer their memory to be RAM.

More serious

can you make it a lot more serious?

AI: more serious

Noted. Excessive reliance on digital note-taking may impair cognitive processing and retention of information.

More funny

can you make it funny?

Note

NOTE: with cursor here, the thread will be: joke request -> AI: first reply -> more funny. In other words, I can add more sub-headings (more funny, more serious, in space, etc.) to the AI message to try create different branches.

AI: attempt 2 at funnier (simply retry)

A student walks into a stationery store and asks for a notebook. The clerk says, "Sorry, we only sell stationary notebooks." The student sighs, "That's okay, I wasn't planning on taking my notes for a walk anyway."

Note

NOTE: with cursor somewhere under More funny I can keep on invoking send current thread to AI and it will keep on adding replies as new children under More funny.

AI: attempt 1 at funnier

I scribbled a joke about note-taking, but I lost the Post-it.

diagram of the heading hierarchy

please create a mermaidjs diagram showing the heading hierarchy above clearly. Count the hashes: all headings with the same number of hashes should be on the same level. "more serious" and "more funny" are children of "AI: first reply". Maintain exact heading labels in the diagram, including any "AI:" prefixes.

Here's a Mermaid.js diagram showing the heading hierarchy as requested:

graph TD
    A["joke request"]
    B["AI: first reply with joke (not so funny...)"]
    C["More serious"]
    D["More funny"]
    E["AI: more serious"]
    F["AI: attempt 2 at funnier (simply retry)"]
    G["AI: attempt 1 at funnier"]

    A --> B
    B --> C
    B --> D
    C --> E
    D --> F
    D --> G
Loading

This diagram accurately represents the heading structure, with "More serious" and "More funny" as children of "first reply", and maintains the exact heading labels including "AI:" prefixes where present.