From 9b0ac775afa206fd9ccb3c3a586be796ed9d6bd7 Mon Sep 17 00:00:00 2001 From: Erik de Vries Date: Wed, 16 Jan 2019 19:36:03 +0100 Subject: [PATCH] class_update: add ver variable to set version for class updated articles --- R/class_update.R | 5 +++-- man/class_update.Rd | 5 ++++- 2 files changed, 7 insertions(+), 3 deletions(-) diff --git a/R/class_update.R b/R/class_update.R index d02204f..a33914b 100644 --- a/R/class_update.R +++ b/R/class_update.R @@ -9,6 +9,7 @@ #' @param words String indicating the number of words to keep from each document (maximum document length), 999 indicates the whole document #' @param text String indicating whether the "merged" field will contain the "full" text, old-style "lemmas" (will be deprecated), new-style "ud", or ud_upos combining lemmas with upos tags #' @param clean Boolean indicating whether the results should be cleaned by removing words matching regex (see code). +#' @param ver Short string (preferably a single word/sequence) indicating the version of the updated document (i.e. for a udpipe update this string might be 'udV2') #' @param es_super Password for write access to ElasticSearch #' @return As this is a nested function used within elasticizer, there is no return output #' @export @@ -17,11 +18,11 @@ ################################################################################################# #################################### Update any kind of classification ########################## ################################################################################################# -class_update <- function(out, localhost = T, model_final, dfm_words, varname, text, words, clean, es_super = .rs.askForPassword('ElasticSearch WRITE')) { +class_update <- function(out, localhost = T, model_final, dfm_words, varname, text, words, clean, ver, es_super = .rs.askForPassword('ElasticSearch WRITE')) { print('updating') dfm <- dfm_gen(out, text = text, words = words, clean = clean) %>% dfm_keep(dfm_words, valuetype="fixed", verbose=T) pred <- data.frame(id = out$`_id`, pred = predict(model_final, newdata = dfm)) - bulk <- apply(pred, 1, bulk_writer, varname = varname, type = 'set') + bulk <- apply(pred, 1, bulk_writer, varname = varname, type = 'set', ver = ver) res <- elastic_update(bulk, es_super = es_super, localhost = localhost) } diff --git a/man/class_update.Rd b/man/class_update.Rd index d91a80e..5d7b26b 100644 --- a/man/class_update.Rd +++ b/man/class_update.Rd @@ -5,7 +5,8 @@ \title{Classifier function for use in combination with the elasticizer function as 'update' parameter (without brackets), see elasticizer documentation for more information} \usage{ class_update(out, localhost = T, model_final, dfm_words, varname, text, - words, clean, es_super = .rs.askForPassword("ElasticSearch WRITE")) + words, clean, ver, + es_super = .rs.askForPassword("ElasticSearch WRITE")) } \arguments{ \item{out}{Does not need to be defined explicitly! (is already parsed in the elasticizer function)} @@ -24,6 +25,8 @@ class_update(out, localhost = T, model_final, dfm_words, varname, text, \item{clean}{Boolean indicating whether the results should be cleaned by removing words matching regex (see code).} +\item{ver}{Short string (preferably a single word/sequence) indicating the version of the updated document (i.e. for a udpipe update this string might be 'udV2')} + \item{es_super}{Password for write access to ElasticSearch} } \value{