An example GeoDeepDive application.
This GeoDeepDive application assumes the following:
- A subset of the GeoDeepDive corpus has already been defined by the intersection of GeoDeepDive's geochron dictionary and geolocation terms included int unlinked.
- The StanfordCoreNLP V3.52 output for this subset exists in a PostgreSQL table
public.nlp_sentences_352
. - That the credentials to connect to this PostgreSQL database are saved in a Credentials.yml file.
- R version 3.4 or later is installed, and that the operating system is Unix-compatible (Linux, OSX).
- That system has internet access or that the R Packages
doParallel
andRPostgreSQL
and their dependencies are already installed. - That there is an empty directory named output to hold the application results.
The application can be run with the following shell command. The default number of cores is 3 if not specified.
RScript ~/Path/usgs_geochron.R numcores
You could presumably try this directly from the git repo without cloning - if you could get around the credentials issues.
curl https://raw.githubusercontent.com/UW-Deepdive-Infrastructure/gdd_byod/master/application/usgs_geochron.R | RScript 4