search

squaredtechnologies / thread

AI-powered Jupyter Notebook — use local AI to generate and edit code cells, automatically fix errors, and chat with your data
https://www.thread.dev
GNU Affero General Public License v3.0
974 stars 49 forks source link

erdantic's chart can't display #8

Closed HaveF closed 2 weeks ago

HaveF commented 2 weeks ago

Hi, it looks very nice!

Thanks! Just a quick test, find a bug. For the lib below, it can't display erdantic's chart. Usage Example: Pydantic - erdantic

image

alishobeiri commented 2 weeks ago

Thanks a lot, will take a look at this, appreciate the feedback!

alishobeiri commented 2 weeks ago

I have been trying to look at this and every time I am defeated by:

Failed to build pygraphviz
ERROR: Could not build wheels for pygraphviz, which is required to install pyproject.toml-based projects

Will need to dig deeper I think

alishobeiri commented 2 weeks ago
image

Progress has been made

alishobeiri commented 2 weeks ago

Believe this has been fixed with: https://github.com/squaredtechnologies/thread/commit/bda38f1db4acd24a488e3cc1b470dcecfb60978f

Thanks a lot for the issue!

HaveF commented 2 weeks ago

@alishobeiri nice. thanks!

To be honest, I have tried to use AI-related plugins in Jupyter, but the user experience was terrible. I saw your news on HN and came to experiment. I feel that this is the right way to use AI in Jupyter. And you used React, I believe you can make use of a lot of existing ecosystems in React. Awesome.

In addition, I looked at your roadmap, and suddenly thought that if you have a set of methods to use React/Python to make plugins and then put them into your environment in the future, it would be great!

I hope you develop well! I believe you have so many possibilities.

alishobeiri commented 2 weeks ago

Hi HaveF, thanks a lot. Curious to learn, for the React/Python plugin, what is the primary use case you have in mind?

HaveF commented 2 weeks ago

@alishobeiri Dashboard? Similar to your Publish notebooks as shareable webapps in your roadmap. It can leverage numerous React UI libraries. And UI support for tools, like, my most favorite notebook tool -- nbdev.

HaveF commented 2 weeks ago

Also, considering this is a tool under jupyter, I was wondering. Is it necessary to repeatedly implement some existing extensions, such as Table of Contents and so on?(like Add support for Jupyter Widgets in roadmap? do we really need this or want something more? I have no idea.) Although there are many extensions that are no longer required in jupyter lab. Or just leave a plugin workflow for others to create what they need.

I think we may need to carefully consider what makes this tool different from other tools. After all, resources are limited.