OpenFn / OpenFn/apollo

global chat: provide a better way to provide a better promt

Open
#500 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

global assistant
Dominant language
Jupyter Notebook
Stars
5
Forks
10
Avg merge
2d 20h
Merged PRs (30d)
17

Description

As a consultant using the global assistant to generate a complex workflow with multiple steps... the current user experience is a big let down.

My workflow has several steps, many requirements and details, and needs careful planning and consideration.

The naive solution would be: upload a full 200 page (or whatever) design brief to the assistant and it'll magically generate the workflow for you. Then all subsequent edits would re-use that context.

But thats not possible in today's technology - and even if it is a solution requires a bit more nuance.

Maybe, rather than structuring chats as a conversation, try a single prompt, like a spec, which the user edits and re-generates the workflow from. This is a high level overview of the workflow. It's not for edits - it's a constant one-shot guideline. It would be big, and would have some UI for it. We may even have a service which can take your design doc and convert it into the prompt

This closely relates to the doc assistant idea - upload your docs and the assistant will just use them to answer questions.

This splits the UX into two modes: chat (for edit, debug, learning) and design (re-generate a workflow some a single spec).

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

The issue does not name files, tests, or implementation entry points. Start by reviewing the existing global assistant flow and the related doc assistant idea, then clarify the proposed chat and design modes, their requirements, and success criteria before implementation; done should be a decided, scoped workflow and UI plan.

Written by the indexing model from the issue text.

Assessment

Domain
ai, design
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Quiet
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.