Skip to content

Latest commit

 

History

History
64 lines (50 loc) · 2.42 KB

Todos.md

File metadata and controls

64 lines (50 loc) · 2.42 KB

TODOs

A mini task management system for the project.

Prio 1

  • use the signin.

  • use the signup to add a user.

  • Authenticate the api-calls via token.

  • run the backend

  • move the backend into a folder

  • There's gonna be a separate react frontend. The components are not there, so it needs to be designed. Initiate a react frontend first.

  • Setup postgres

  • Connect the backend to an actual sentiment-analysis-tool

  • Make a usecase scenario list: Scenarios are:

    - user sign up
- user login
- user logout
- see list of user's texts with their evaluation.
- give feedback to the evaluation of a text.
- Admin can see the list of evaluations that have a feedback, with the feedback.

  • Create table for texts: id:number, content: char(1000), userid: number, optional evaluation?: short number

  • Create table of feedback: FK, text_id, content: char(140)

  • Create a Post route for sending feedback

  • Create a post route for an admin to get all feedbacks

  • Create a post route for a user to see all their texts

The frontend:

  • Make a rough list of Components: A logout button, A login page, A sign up page, A list of texts with eval, a list of feedbacks, a button to add feedback, and finally a navigation bar (includes logout/login/signup, About page, feedback/text list pages.)
  • import bootstrap
  • Navbar
  • About page
  • Login page
  • Sign up page
  • List of user's texts
  • Upload new text

The Sentiment Analysis Tool

  • I don't have a proper GPU; Is it possible to use a free online service? YES, but local installation is working alright.
  • Probe the hugging face. Does it offer free llm APIs for sent. analysis? Their transformers library is Ok-ish.
  • Apply the Good/Bad/Neutral categories.

Prio 2 Important

  • remove hardcoded url
  • Apply unified bootstrap.

Prio 3 Nice to have

  • Hash or salt the password before uploading over the network.
  • Https with certs (certbot...)
  • Integrate with blockchain as connecting with metamask using web3.js.
  • Deploy your solution to the cloud.
  • Implement engineering best practices (source control, CI/CD, infrastructure-as-a-code)
  • Commit to the master branch
  • Benchmark and choose CI/CD tool set (Docker, Kubernetes, cloud platforms) on (Jenkins, GitHub Actions, gitlab,..)
  • Levelled logging based on environment?