search

rinkstiekema / table-structures

87 stars 30 forks source link

json not getting created #7

Closed ankur7721 closed 4 years ago

ankur7721 commented 4 years ago

After using pix2pix model, I am unable to generate json. Code runs fine, there is no error but on analyzing I found rulers.py has some problem i guess, couldn't figure out though. P.S I am not using any skip options

Can you please help me point it out.

rinkstiekema commented 4 years ago

I'm a little confused by what you are trying to do. The final result should be a CSV file. The JSON file is generated by PDFFigures2 and is used in the extraction process, but is not a goal on it's own. Can you ellaborate I little more what you are trying to do?

leyuan commented 4 years ago

@rinkstiekema I think the issue is no main manifest attribute, in pdffigures2.jar

and there is no corresponding json files created in the json directory. I have tried couple downloads and went to the source code as you pointed in another PR, is it possible to share the jar file that you use as well?

Thanks

rinkstiekema commented 4 years ago

I am using the one that is in the repository (https://github.com/rinkstiekema/PDF-Table-Structure-Recognition-using-deep-learning/blob/master/pipeline/pdffigures2.jar), I'm not quite sure what is going wrong with your version. The jar does include the manifest in the META-INF folder.