I am searching for a table recognition library which can operate on plain text. Ideally the library is not a dirty hack of regular expression, but something inference-based.
All I can find are papers but no implementations of the concepts described. I am willing to implement such a concept, but I can not believe there is nothing out there, yet.
Anyone has the name of an open source project or a good hint for me?
Clarification: I am not interested in OCR libraries.