agricolamz

library(tidyverse)
library(phonfieldwork)
library(writexl)

read_from_folder(getwd()) |> 
  mutate(content = str_squish(content)) |> 
  filter(content != "") |> 
  mutate(content = str_squish(content), 
         word = ifelse(tier == 1, str_c(content, "_", time_start, "_", time_end), NA)) |> 
  fill(word) |> 

agricolamz

# based on Stefan Evert's lecture

library(tidyverse)
library(tidytext)
library(stopwords)

# create stopwords list ---------------------------------------------------

map(stopwords_getsources()[-c(3:4, 6, 8)], function(i){
  stopwords(language = "ru", source = i)

agricolamz

library(tidyverse)
library(scholar)

my_id <- "ka_iMFQAAAAJ"

my_pubs <- get_publications(my_id)

map_chr(my_pubs$pubid, function(i){
  res <- get_publication_abstract(id = my_id, pub_id = i)
  ifelse(length(res) > 0, res, "")

agricolamz

library(tidyverse)
library(bib2df)
bib2df("verbal_negation.bib")  |> 
  rowwise() |> 
  mutate(AUTHOR = str_c(AUTHOR, collapse =" and "),
         EDITOR = str_c(EDITOR, collapse =" and ")) |> 
  write_tsv("verbal_negation.tsv")

bib2df("verbal_negation_spec.bib")  |> 
  rowwise() |> 

agricolamz

library(tidyverse)

read_lines("revizor.txt") |> 
  str_squish() |> 
  tibble(text = _) |> 
  filter(!str_detect(text, "^ДЕЙСТВИЕ"),
         !str_detect(text, "^Явление"),
         text != "") |> 
  mutate(id = 1:n()) ->
  revizor

agricolamz

library(tidyverse)
library(phonfieldwork)

files <- list.files(pattern = "TextGrid")
walk(files, function(file){
  textgrid <- textgrid_to_df(file) 
  textgrid |> 
    mutate(content = str_extract(content, "^.*?-"),
           content = str_remove(content, "-"),
           content = str_replace_all(content, "SS", "ss"),

agricolamz

library(tidyverse)
t <- pdftools::pdf_ocr_text("Khan 2008 Jewish Neo-Aramaic Dialect of Urmi-465-497.pdf")

tibble(text = str_split(t, "\n\n") |> unlist()) |> 
  filter(!str_detect(text, "GLOSSARY OF VERBS"),
         nchar(text) > 4) |> 
  slice(-c(1:2)) |> 
  mutate(verb = str_extract(text, "\\S{1,}\\s"),
         verb = str_squish(verb),
         text = str_remove_all(text, "\n")) |> 

agricolamz

library(tidyverse)

read_csv("russian_libraries_subdataset.csv") |> 
  count(dedication, sort = TRUE)

agricolamz

speech_to_text <- function(audio, 
                           output_name = "output",
                           model_path = "ggml-large-v3.bin"){
  library(tidyverse)
  library(audio.whisper)
  
# convert to the format specs ---------------------------------------------
  tmp <- tempdir()
  str_glue("ffmpeg -i {audio} -ar 16000 -ac 1 -c:a pcm_s16le {tmp}/{output_name}.wav") |> 
    system()

agricolamz

library(stringi)

coresp <- "
    ῶι > ῷ;
    ωι > ῳ;
    ὧι > ᾧ;
    ὦι > ᾦ;
    ηι > ῃ;
    ῆι > ῇ;
    ἦι > ᾖ;