*disclaimer
870592
udpipe
https://bnosac.github.io/udpipe/en/
Universal Dependencies
https://universaldependencies.org/#ud-treebanks
特徴
インストール
```{r, eval=F} install.packages("udpipe") ``` ```{r} library(udpipe) library(dplyr) library(ggplot2) library(stringr) ```
使用例
library(ggraph) library(ggplot2) library(igraph) library(textplot) bun <- udpipe("What kind of sports do you like?", "english") bun doc_id paragraph_id sentence_id sentence start end term_id token_id token lemma upos xpos feats head_token_id dep_rel deps misc 1 doc1 1 1 What kind of sports do you like? 1 4 1 1 What what DET WDT PronType=Int 2 det NA NA 2 doc1 1 1 What kind of sports do you like? 6 9 2 2 kind kind NOUN NN Number=Sing 7 obj NA NA 3 doc1 1 1 What kind of sports do you like? 11 12 3 3 of of ADP IN NA 4 case NA NA 4 doc1 1 1 What kind of sports do you like? 14 19 4 4 sports sport NOUN NNS Number=Plur 2 nmod NA NA 5 doc1 1 1 What kind of sports do you like? 21 22 5 5 do do AUX VBP Mood=Ind|Tense=Pres|VerbForm=Fin 7 aux NA NA 6 doc1 1 1 What kind of sports do you like? 24 26 6 6 you you PRON PRP Case=Nom|Person=2|PronType=Prs 7 nsubj NA NA 7 doc1 1 1 What kind of sports do you like? 28 31 7 7 like like VERB VB VerbForm=Inf 0 root NA SpaceAfter=No 8 doc1 1 1 What kind of sports do you like? 32 32 8 8 ? ? PUNCT . NA 7 punct NA SpacesAfter=\n textplot_dependencyparser(bun)
日本語対応
日本語モデル
1. ダウンロード
jp.model <- udpipe_download_model(language="japanese")
japanese-gsd-ud-2.5-191206.udpipe がダウンロードされる
2. モデルの読み込み
jp.model <- udpipe_load_model(jp.model$file_model)
3. テキスト解析
parsed <- udpipe_annotate(jp.model, x = "日本語文") bun <- as.data.frame(parsed)
4. プロット
library(textplot) textplot_dependencyparser(bun)
https://sugiura-ken.org/wiki/