nomic-ai / gpt4all

GPT4All: Run Local LLMs on Any Device. Open-source and available for commercial use.
https://nomic.ai/gpt4all
MIT License
69.62k stars 7.62k forks source link

v3.3.0: LLM reply: The selected LocalDocs collections to be listed vertically, as to fit in the chat window #2987

Open SINAPSA-IC opened 3 days ago

SINAPSA-IC commented 3 days ago

Bug Report

When the model is "searching localdocs:", the selected LocalDocs collections are all listed on 1 single line, one after another, single file, the length of the resulted line surpassing the width of the region holding the text of the conversation.

Suggestion: when building (in code) the list of N strings, for the first N-1 strings Replace the comma with (comma+Environment.NewLine).

Steps to Reproduce

  1. select more than 10 LocalDocs collections - it may be 3 or 15, depending on the length of their names and the Font Size of the application
  2. ask the model something
  3. while "searching localdocs:" the comma-separated names of those collections are listed in a continuous sequence, on a single line/row

Expected Behavior

The user should be able to see all the names of the selected collections, instead of an incomplete text which defeats the intention itself to list the collections: either you show them all, or you don't.

Your Environment

So in v3.3.0 we the users (like, the users) do not have neither this^, nor a logical Tab order in the textboxes on Settings->Model, nor an alphabetical ordering of models, nor the selection of CPU or CUDA as the Embedding Device is remembered from one session to the next, among other basic functionality for the UX, so neither elementary things, nor complex ones. Does anyone, u know, test this program? Besides the users, that is? Does anyone even read, understand, replicate these Issues, to correct them, or is it Everything Already Known (well, newsflash, it doesn't show)? Is there any form of quality control over implementing even something elementary? The coders themselves have to test their own code, what does it do, how does it look. Is this^ even the stuff for an Issue? Really? Or a Feature Request? Is this^ a Feature? Really? I'm trying to use this program which is not that complex to work with, which makes it attractive, but I keep bumping into inexcusable, elementary things like this. So I advertise this program everywhere I travel, talk about it in papers, and the Developers give me this^ that can be resolved in 30 seconds? Really? Feel offended? You should, for truth is ugly, and I the user feel even more offended.

SINAPSA-IC commented 3 days ago

ui4 ui5