Help Needed: Extracting Text and Tables from Textbook Using Llama-Parse
Help Needed: Extracting Text and Tables from Textbook Using Llama-Parse
At a glance
The community member is seeking help with using "parsing instructions" in the llama-parse tool to extract only text and tables from a textbook-like image. They have tried various instructions, including "accurate" and "premium-mode", but are having difficulty properly exploiting the functionality. Another community member suggests setting is_formatting_instruction=False in the API/UI when using parsing instructions. The community member thanks them for the suggestion and requests more information on how to structure parsing instructions to extract text from the image, even a draft example. Another community member states they were able to do this successfully, and the original poster thanks them and says they will try it, as they had issues with double columns and transcription errors previously, possibly due to unnecessary details in their instructions.
Hello everyone !!! i would like some help/opinion. what "parsing instructions" would you use with llama-parse to extract only text and tables from a textbook like the one in the image ?
i am trying and trying again with multiple instructions, but it is obvious that i can't exploit this great functionality properly. I mainly use "accurate" and "premium-mode." Any suggestions are welcome ! thank you very much.
Thank you for the suggestion. i would also like to have some more info for the "parsing instructions" for how to make the best use of them. for example how would you structure instructions to extract text from the page in the image. even a draft would be fine, then i work on it. Thank you very much!
thank you very much, i will try. i had used the same page but it had returned text in double column and other transcription errors. maybe i had put too much unnecessary detail that was confusing.
