https://twitter.com/gokulnk/status/1524728450879156224

TODO

  1. A basic program that takes a line image as input and converts it into red and green ticks like stock based on the previous value. Candle Stick?
    1. Break the image into vertical strips.
    2. Identify the contiguous black pixels so that you can calculate the bottom height and top height.
    3. Tell me if you want to take a dig at it.
  2. Take a pdf as inputs and extract tables from it. the tables might be spread across the different pages.
    1. http://dspace.kila.ac.in/bitstream/123456789/655/1/LIF%20Book%20on%20Localising%20SDGs%20in%20LSGI%20final.pdf
  3. In the books that i read i use brackets to highlight things. i will share the image with you. from the image you should be able to extract the text that is present in the image. most books have a default font or group of fonts. so a basic font identification and creation of a list should reduce the error percentage. so the only task is to get your brackets right and the english text within that will be easy to get.
    1. once you do this i want to expand this further. so that if i give you video where i capture the pages you should be able to give me the final output.
  4. use AI to scrape the insights from https://ngodarpan.gov.in/index.php/home/statewise
    1. list of all the NGOs
    2. do they have website or not
    3. what is their yearly booking
    4. how many people are affected
    5. generate entities from and find their relationships

Referenced in:

All notes