From 386ac42aee78d3b73cd5b675cdf10f5a39ed0cd4 Mon Sep 17 00:00:00 2001
From: Erik de Vries <erik@devries.pm>
Date: Tue, 15 Jan 2019 11:36:51 +0100
Subject: [PATCH] lemma_writer: new function to write raw lemma's (without
 interpunction) to text file. Is structured as elasticizer update function
 (despite not updating anything on the server)

---
 R/lemma_writer.R    | 21 +++++++++++++++++++++
 man/lemma_writer.Rd | 24 ++++++++++++++++++++++++
 2 files changed, 45 insertions(+)
 create mode 100644 R/lemma_writer.R
 create mode 100644 man/lemma_writer.Rd

diff --git a/R/lemma_writer.R b/R/lemma_writer.R
new file mode 100644
index 0000000..d4daf45
--- /dev/null
+++ b/R/lemma_writer.R
@@ -0,0 +1,21 @@
+#' Generates text output files (without punctuation) for external applications, such as GloVe embeddings
+#'
+#' Generates text output files (without punctuation) for external applications, such as GloVe embeddings
+#' @param out The elasticizer-generated data frame
+#' @param file The file to write the output to (including path, when required)
+#' @param localhost Unused, but defaults to FALSE
+#' @return A Quanteda dfm
+#' @export
+#' @examples
+#' dfm_gen(out, words = '999')
+
+
+#################################################################################################
+#################################### Lemma text file generator #############################
+#################################################################################################
+
+lemma_writer <- function(out, file, localhost = F) {
+  out <- unnest(out,`_source.ud`)
+  lemma <- str_c(unlist(out$lemma)[-which(unlist(out$upos) == 'PUNCT')], unlist(out$upos)[-which(unlist(out$upos) == 'PUNCT')], sep = '_')
+  cat(lemma, file = file, append = T)
+}
diff --git a/man/lemma_writer.Rd b/man/lemma_writer.Rd
new file mode 100644
index 0000000..bc2a099
--- /dev/null
+++ b/man/lemma_writer.Rd
@@ -0,0 +1,24 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/lemma_writer.R
+\name{lemma_writer}
+\alias{lemma_writer}
+\title{Generates text output files (without punctuation) for external applications, such as GloVe embeddings}
+\usage{
+lemma_writer(out, file, localhost = F)
+}
+\arguments{
+\item{out}{The elasticizer-generated data frame}
+
+\item{file}{The file to write the output to (including path, when required)}
+
+\item{localhost}{Unused, but defaults to FALSE}
+}
+\value{
+A Quanteda dfm
+}
+\description{
+Generates text output files (without punctuation) for external applications, such as GloVe embeddings
+}
+\examples{
+dfm_gen(out, words = '999')
+}