New plugin: smart-text-mover

Hi all,
I am happy to announce my first contribution to Obsidian ecosystem as a plugin: smart-text-mover. The problem I tried to solve with it has been personal but I hope you will find it useful too.

Problem statement: A couple of months back I decided to use Obsidian as my default app for bookmarks. So I started saving links I found useful into Obsidian. Obsidian by default adds the link at the end of the file, so I created different heading and used to manually arrange newly added links under the relevant heading. Soon it became tedious I decided to come up with intelligent way to sort the links under the right topic/heading.

Solution: Smart-text-mover uses a light weight machine learning algorithm (Naive Bayes Classifier) to make good default category suggestions, thereby reducing the cognitive load and move links/text within a file without a hassle. It also has optional LLM support to classify links into relevant category.

Hi. Do you have particular reason not to associate links with unique notes? You could then apply properties/tags to these notes to mark subject, topic etc.

On mac you can use Raycast to clip content to Obsidian. Currently there are three such extensions:

  1. Obsidian Smart Capture — August 17, 2023, captures url automatically from browser (Safari, Chrome, Arc, Brave, Firefox, Edge & Opera)
  2. Obsidian Bookmarks — April 23, 2024, captures url automatically from browser (Chrome, Safari, Arc)
  3. Obsidian Clippings — Jun 23, 2023, AI summarised pages using Raycast AI (premium service that costs $10 / month)

Current limitations of these tools:

  • no AI summarisation (except Obsidian Clippings)
  • no automatic translation
  • no automatic transcripts

Hi,
I have a bookmark file where I dump all the links I come across. Categorization happens in that file only in form of headings as well as tags. So it looks something like:

# Information Retrieval
[What AI Engineers Should Know about Search](https://softwaredoug.com/blog/2024/06/25/what-ai-engineers-need-to-know-search) #read  [[2024-06-26]]
[LambdaMART in Depth](https://softwaredoug.com/blog/2022/01/17/lambdamart-in-depth)
# Deep Learning/LLMs
[A Recipe for Training Neural Networks](https://karpathy.github.io/2019/04/25/recipe) #read [2024-07-02]
[Why AI folks need a broad based intro by  Subbarao Kambhampati on LinkedIn] (https://www.linkedin.com/posts/subbarao-kambhampati-3260708_ai-ai-ai-activity-7210796921060216833-iqsI) #to_read 

Regarding associate links.
I dont commit to creating a note/link out of url unless I feel that i would reread or revise it later.
Once commited, then I create a note, and clip text and add tags.

In short I see bookmarking and clipping as 2 different activities(former is more of a dumping ground, later is more of a selective commitment).
Now that I think of it, I would prefer to do the later(separate note per item) for research papers, because need to revisit/refer/quote/cite is more there.