tabula VS Cuba

Another option would be JRuby. I routinely use an application called Tabula, which is built using JRuby and compiles to a Jar file. This, of course, requires Java on the target machine, but you can ship the Jar file and it will work. It's often easier to rely on a working Java environment than it is a working Ruby environment. Especially on Windows.

Idk if it has been suggested yet, tabulapdf is a self hosted solution to extract tables from PDF

For table extraction from pdfs, have a look at Tabula and Camelot, two open-source projects. They work well with clean tables, both the Tabula Python binding and Camelot allow you to export directly as a pandas dataframe. Otherwise AWS Textract API is very efficient at extracting tables from pdfs, regardless of how clean/messy they are.

No idea if it will work for you, but there is a git project that seems to do what you want https://github.com/tabulapdf/tabula

While trying to find a specific project I recalled, I encountered this list of projects which might be of interest: https://github.com/tstanislawek/awesome-document-understandi...

The project I had in mind was similar to this one but I can't remember the name currently: https://github.com/tabulapdf/tabula

However, if you're looking for a ML-based, invoice-specific project looks like the other comment to your reply might be more useful.

Cuba (cuba.is)

To anyone here who enjoys and values the things Roda is good at: I would recommend you also take a look at Cuba, the project Roda is forked from.

Cuba is a microframework to develop web applications in the Ruby language. Rum inspires Cuba, and the official website defines Cuba as “a tiny but powerful mapper for Rack applications.” making it one of the best ruby frameworks. The GitHub page is a practical guide if you are looking to start development in Cuba.

Ties together a Cuba web app, the tabula-java library and lauch4j to provide a platform executable.