Hug data is a Desktop-only Web App dedicated to the world of Computer Vision, inspired by both Kaggle and Medium.
In order to run the project locally you need to execute these step:
-
Install Node.js -> installation guide
-
Install all dependencies via command in your terminal:
npm install -
Run the server via command in your terminal:
node app.js
HugData support 3 types of user:
| Functionality | Anonymous | User | Creator |
|---|---|---|---|
| Search knowledgebase | ✔️ | ✔️ | ✔️ |
| Read Article | ❌ | ✔️ | ✔️ |
| Download Dataset | ✔️ | ✔️ | ✔️ |
| Write Article | ❌ | ❌ | ✔️ |
| Publish Dataset | ❌ | ❌ | ✔️ |
| 6 operations of Data Augmentation | ❌ | ✔️ | ✔️ |
| Leave comment on article | ❌ | ✔️ | ✔️ |
| See creator page | ✔️ | ✔️ | ✔️ |
Several extra libraries had been used for the development of all the functionalities, which are:
- Jimp, image processing.
- Multer, handling files inside HTML form.
- Decompress, decompression of datasets
- Archiver, compression of datasets
- express-session-sqlite, saving sessions on the database.
I choose to use an ORM interface to avoid hardcoded query and have an easier coding experience, inside ./database you can find every operation related to the interaction with the database.
Below you can see an example of how DDL and DML are implemented:
Creation of Dataset table
Searching a dataset by title
Here some credentials to try the website!
| Type | Username | Password |
|---|---|---|
| User | Ford_42 | Artur_Dent01 |
| Creator | Hari_seldon | Foundation11_ |