Whole document parsing

Hi,

I'd like to parse a whole document (text) and extract the molecules as either an array or collection.

Looking at the documentation it appears as though MolImporter.importDoc might work - am I on the right track ?

 

Cheers

 

Luke

Hi Luke,

No, the name-to-structure extraction from documents is not released yet. We plan to include it in next version 5.2.3 in about a month. If that schedule is a problem we can discuss an early evaluation version.

Cheers,

Daniel

 

Hi Daniel,

Ok, I'd really like to try the development version on evaluation, or alternately could you provide some guidance on how to approach parsing a body of text. i.e. is it rational to attempt every token etc..

Thanks

 

Luke

It's of course possible to attempt every token, but you would miss names spanning over multiple tokens. You can of course try concatenating several tokens, but that can lead to poor performance, and that will not handle OCR errors, names broken with extra spaces or dashes, ... That's why a document parser is non-trivial.

I'll contact you in a couple of days about an evaluation.

Cheers,

Daniel