natural language processing for lithological descriptions

[jmp75]

I saw in the video from scipy2015 that there were ideas for using NLP for parsing lithological descriptions and would be interested in the feature. I did not see any issue about it hence opening one.

To give a bit of context I am currently developping this package for exploratory lithology analysis. I am using striplog for some of it (abbreviations), and doing "my own" regex based analysis for lithology classification. I reckon this classification capability has more its place in the striplog package.

If using striplog e.g. with

striplog.Component.from_text(u'black clay with layers of fine sands', lex)

I obtain:

attr	val
colour	black
lithology	clay
grainsize	fine

More sophisticated regular expressions may alleviate such issues but NLP is probably something to explore these days (Disclaimer: I am far from an NLP guru). I am trying to find project resources (e.g. intern) to work on this. I'm doing a bit of lit. research but I'd value guidance or feedback (best case, something already implemented)