Skip to content

Parsing of table content#546

Merged
kermitt2 merged 2 commits intogrobidOrg:masterfrom
Vitaliy-1:master-tables
Feb 29, 2020
Merged

Parsing of table content#546
kermitt2 merged 2 commits intogrobidOrg:masterfrom
Vitaliy-1:master-tables

Conversation

@Vitaliy-1
Copy link
Copy Markdown
Contributor

Algorithms for table parsing. Inspired by: http://ieg.ifs.tuwien.ac.at/projects/pdf2table/yildiz_iicai05.pdf
Particularly, I've implemented the first algorithm from there and the idea in general. I suppose it's called the line-based approach. It works much better than tabula, as a minimum on my test instance.
related to:
#514
#340
Examples: https://vitaliy-1.github.io/tablesExample/

@coveralls
Copy link
Copy Markdown

Coverage Status

Coverage increased (+0.5%) to 37.788% when pulling ca98798 on Vitaliy-1:master-tables into 53044fd on kermitt2:master.

@kermitt2 kermitt2 self-requested a review February 29, 2020 07:57
@kermitt2
Copy link
Copy Markdown
Collaborator

Awesome work @Vitaliy-1 ! I've processed a couple thousand PDF and looked at a few of them. Tables look often very good. The main issue I think, as you noted in your examples, is when the table notes are added in the last line of the table. Possibly we could introduce an additional label in the table model to label those table notes explicitly to get them removed from the table content. What do you think?

Thanks a lot for the contribution!

@kermitt2 kermitt2 merged commit b2bd435 into grobidOrg:master Feb 29, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants