NLP based approach to automatically categorize your bookmarks!

To understand this project in-depth, refer to my technical paper: Bookmark Classification using Multinomial Naive Bayes Model

• Enter your bookmarks in ./links.json file
• To run the code, run categorize.py
• scrape_filter_link.py contains the classes used to scrape information from each URL

It can categorize a variety of bookmarks. Currently it supports all the categories mentioned in the ./corpus/ directory.

To a certain extent! For example: Firefox allows users to backup the bookmarks in a JSON format. You can extract the uri from that JSON file and feed it into ./links.json.
To backup your bookmarks in Firefox, press Ctrl+Shift+O, go to Import and Backup and then to Backup.
Chrome users can check this post on superuser.

No, the mapping of a URL with it's appropriate category is stored in a JSON file: result.json, in a dict format.
The keys are your bookmarks with values being their categories.

Sure, here's one (The highlighted part is the one stored in result.json):

Yes, you can! The code is fairly scalable.
To add your own corpuses:

• Create a directory with a unique category name in ./corpus/
• Inside the ./corpus/your-category-dir add your corpus text in a JSON file with the format: {"text": "_your_corpus_text_here_"}

(NOTE: You can add multiple JSON files in a category directory)
When you run the code, you will find that the categorize.py will take the new/modified corpuses into consideration.

The code is under MIT License