1h 1m 53s logged
Deployed the whole thing onto Render to bypass local installs.
Ship
Changes requested
Introduction
I made a media annotation pipeline for images, videos, and audio datasets, with automated quality control using AI.
Instructions on how to run the app locally are in the source files! :)
How It Works
There are four stages:
-
Admin: Create jobs by uploading media URLs and defining category lists.
-
Labeller: Annotate media tasks with category labels and free-text descriptions.
-
AI QC: Automated quality control using Gemini to sample and verify labels. Gemini reviews a sample of labels per job:
- Small jobs (<10 tasks): one random task per labeller.
- Large jobs: roughly 10% of tasks are randomly sampled.
- If a sampled task fails, the entire labellerās work for that job is flagged for human review.
-
Human QC: Review flagged labels, approve correct ones, or send them back for relabelling by setting the task status back to
open.
Hope this is useful!
1h 45m 14s logged
- Finally fixed AI QC not running (model incompatibility): Human QC page now can receive tasks that failed the AI QC!
- I also wrote end to end tests that start up a uvicorn server and mocks Gemini + uses playwright to drive HTML pages for happy, sad and edge cases to make sure it really works.
Going to ship now :) I hope someone can find this helpful!
2h 30m 21s logged
Finished up unit tests for labeller, ai qc and human qc. Also finished the ai qc and human qc endpoints + added human qc html page :)
The image attached here is also a recent addition, the main index.html page showcasing a brief intro of this project, the pipeline, and links to the three html pages.
Also major development: added a way for others to test my project with/without API keys for Gemini or Turso (database), and updated README with implementation steps; will ship soon yayy
3h 38m 21s logged
Updated the readme, working on a new ānotion-likeā app (utilizes notion api), fixed some stylistic issues, added a noticeboard for important info on the go.
Ship
I made a basic OS interface. I started by following the guide step-by-step so I could have a basic setup to add upon later.
Apps
Ended up adding three apps (photobook, about, video archive): their icons act as window opener / closer buttons to minimise or reopen a window. Each app has a designated resizable and draggable window. ONly video archive can be set to fullscreen (intentional choice that I may change later)
The photobook showcases my art and figurines, and hovering above each image reveals a unique caption for that image!
The video archive is a set of videos from yt that I collected to have a curated video feed. It features embeds.
The about me window is self-explanatory. I tried to use social media icons for the contact section, but ended up using AI to create unique SVGs.
Challenges
I tried to add an MP3 player that I built for my old site as an iframe but there were many broken links and images that it was too much of a hassle. Might add it in afterwards or rebuild it from scratch.
Main challenge was also that for most of the time the interface just looked so ugly ššš took quite some time just getting the visuals right (I still think it can improve but eh).
Also misc challenges, I SPENT 2 HOURS wondering why my windows werenāt draggable. I spent that time tweaking my html and js files, but I DIDNāT EVEN ATTACH THE SCRIPT to my main html. L
Anyway, thatās it for now! Hope this basic OS is acceptable enough :) I intend for it to be a productivity tool for myself so Iāll def add on to it later, this first ship is just for WebOS 1 requirements.
51m 16s logged
-Added fullscreen option
-Added new app :3 video archive! embeds yt videos for a curated feed!
The requirements for webOS 1 are met already so Iāll ship this first and add more features later on!
20m 17s logged
MADE WINDOWS RESIZABLEE
2h 12m 41s logged
Made some stylistic changes!
-Added my programmerās signature to the middle of the top bar
-Changed the design of windows
-Added app icons as if a standard desktop to open the windows
and ofc
-Added to my about me.
My social links are also in :)
2h 36m 46s logged
Added unit tests, restyled both admin and labeller html pages. Started on AI logic
2h 50m 3s logged
Added some windows to test out styles and top bar. Windows are draggable and closable :) I want to make a fairytale themed OS so letās see how this goesā¦
2h 9m 8s logged
Added a text area for labellers to input their names, and updated the āsubmit labelsā button so that it now collects labels from ticked checkboxes and other descriptive labels, and sends them to the database.
So now the labellerās name and the labels can be recorded in the database! Next step is to integrate AI for checking.
2h 0m 55s logged
Admin page (and connected to backend) done, where āadminsā insert urls to be labelled as requested by clients and identify possible categories all urls might fall under, and submit urls and categories to the database. Labeller page partially done, where labellers āget a new taskā that renders a url and lists categories as checkboxes to be ticked off and an additional textbox for any additional identifiable labels. To-do next: connect label submission buttons to endpoint in backend.
