Add support for TableField and TableQuestion in dataset creation

[JonnyTran]

Summary

As part of our dataset creation workflow in the frontend, we need to support two new item types:

• TableField: a table that represent fields in the dataset schema. It may have 1 column (i.e. a list of values), or multiple column
• TableQuestion: Questions that allow creating and editing
  table annotation

This issue tracks adding UI and logic support for creating TableField and TableQuestion within the extralit-frontend/components/features/dataset-creation/ module.

Requirements

1. Update dataset creation composables (e.g. useDatasetConfiguration) to:
   • for Dataset fields, allow mapping json types to TableField
   • Allow creating TableQuestion
2. Extend DatasetConfiguration UI:
   • Add a new TableFieldSelector component under configuration/questions/
   • Add a new DatasetConfigurationTableQuestion.vue to render and configure table questions
3. Ensure record preview and question mapping correctly handle table contexts
4. Add unit tests covering:
   • Initialization of TableField and TableQuestion in the store/composable
   • Rendering of Table questions in the form

Acceptance Criteria

• Dropdowns and forms allow users to add and configure table fields/questions
• Configuration persists in the dataset object and flows through to preview
• No regressions in existing question types

Feel free to iterate on the UI/UX in follow-up PRs. SchemaService and full SDK integration are out of scope for this sprint.

[sahaniaditya]

@JonnyTran can i start working on this issue?

[JonnyTran]

Hi @sahaniaditya , would love to have your help!

Currently there is already TableField and TableQuestion types in both extralit-server/ and extralit/ (somewhat in rough shape), but it's not supported in the dataset-creation/ module in the UI, so I think you might have to touch the Vue.js (v2) code in extralit-frontend/.

[sahaniaditya]

thank you for the info. i will start looking into it.

[JonnyTran]

Also, it's not straightforward on at the dataset-creation on which data to populate in the TableField - so we'd need some discussion/thoughts here. I suggest adding TableQuestion in the dataset-creation first, since it won't have a lot of dependencies

[sahaniaditya]

okay. I will try understanding the issue first and let you know if anything comes up.

[sahaniaditya]

hi @JonnyTran, can you share any guide to setup the repository to my local device? I could not find any on the repository.

[JonnyTran]

Hi @sahaniaditya, if you are setting up the whole backend and frontend in your local machine, you can follow this guide

Otherwise if you're only working on the frontend, it might be easier to spin up a Huggingface instance with a Supabase postgres backend with our tool https://hub.extralit.ai, then would be able to just connect the frontend your HF instance with

# in extralit-frontend/
# `https://{hf-instance}.hf.space` is your instance's URL
API_BASE_URL=https://extralit-public-demo.hf.space/ npm run dev

You can join our slack https://extralit.slack.com/join/shared_invite/zt-3gw1ah8bl-AiVNrkIVYOL4yVGOxN8WFw#/shared-invite/email if you want more tech support - would be happy to answer any question there too!