HI,
Here is a PDF
20240813DEA01_last_times_fr.pdf (177.1 KB)
from which I would like to extract only the highlighted information on PAGE 1
@Brain you can take a look at this example and try to manipulate it. Three approaches with Tabula, Camelot and PDFPlumber extract tables from your PDF file
With the help of PyMuPDF you can extract individual areas of your pdf file so to get the headline and the date. You will have to define the areas so as to capture enough of the text but no unwanted context.
You will have to do some cleaning up and combining and also you could try and use several results to see if they agree on the outcome and only use them or something similar.
The Tika Parser is used to demonstrate how to extract a ll the images that are contained in the PDF - just to see if it works …
More on the logic behind the packages in this article:
What do mean by "the problem starts afterwards "? Do you have model you can share?
If the PDF files are not digital you can use “Tikka node” to extract embedded pictures into a folder (if these pictures are known by Tikka othewise you can use poppler to split PDF files into pictures with the “external Tools node” to poppler) and use “Tess4J node” to do OCR in each pocture and get digital content.
If PDF files are digital you can use Tikka to extract digital content.
With these digital contents, you can use “Regex Extractor node” to extract the columns you want.
Hello,
@Brain you will have to install the necessary packages following the yaml file in the /data/ subfolder and the instructions in the article. I was not yet able to adapt the conda node to windows also as I would normally do. Might take a few days.
py3_knime_pdf.yml (3.2 KB)
