Translation Table Workflow"

In polyglotr: Translate Text

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  eval = FALSE
)

library(polyglotr)
library(dplyr)

Translation Table Workflow

This vignette demonstrates how to use create_translation_table() to translate a vector of phrases across multiple languages, creating a comprehensive multilingual reference table. This is particularly useful for localization projects and multilingual data preparation workflows.

Overview

The create_translation_table() function allows you to:

• Translate multiple phrases simultaneously across several target languages
• Generate organized tabular output for easy reference
• Support multilingual workflows and localization projects
• Create consistent translations for UI elements, documentation, or data labels

Basic Usage

Let's start with a simple example translating common UI elements:

# Define phrases to translate
ui_phrases <- c("Welcome", "Login", "Password", "Submit", "Cancel", "Help")

# Define target languages
languages <- c("es", "fr", "de", "it")

# Create translation table
translation_table <- create_translation_table(ui_phrases, languages)
print(translation_table)

This creates a data frame with the original English phrases and their translations in Spanish (es), French (fr), German (de), and Italian (it).

Multiple API Comparison

While create_translation_table() uses Google Translate by default, you can compare results across different translation APIs for quality assurance. Here's how to create an enhanced function that supports multiple APIs:

# Enhanced function to support multiple translation APIs
create_translation_table_multi_api <- function(words, languages, apis = c("google", "mymemory")) {
  results <- data.frame(original_word = words)

  for (language in languages) {
    for (api in apis) {
      column_name <- paste0(language, "_", api)

      if (api == "google") {
        results[[column_name]] <- sapply(words, function(x) {
          google_translate(x, target_language = language, source_language = "en")
        })
      } else if (api == "mymemory") {
        results[[column_name]] <- sapply(words, function(x) {
          mymemory_translate(x, target_language = language, source_language = "en")
        })
      }
    }
  }

  return(results)
}

# Sample phrases for comparison
phrases <- c("Good morning", "Thank you very much", "How are you today?")

# Create comparison table with multiple APIs
comparison_table <- create_translation_table_multi_api(
  words = phrases, 
  languages = c("fr", "es"), 
  apis = c("google", "mymemory")
)

print(comparison_table)

# You can also use the standard function and add MyMemory translations
google_table <- create_translation_table(phrases, c("fr", "es"))

# Add MyMemory translations for comparison
google_table$fr_mymemory <- sapply(phrases, function(x) {
  mymemory_translate(x, target_language = "fr", source_language = "en")
})

google_table$es_mymemory <- sapply(phrases, function(x) {
  mymemory_translate(x, target_language = "es", source_language = "en")
})

print("Standard function with added MyMemory translations:")
print(google_table)

Advanced Multilingual Workflow

For larger localization projects, you can organize your workflow systematically:

# Define categories of phrases
navigation_phrases <- c("Home", "About", "Contact", "Services", "Portfolio")
form_phrases <- c("First Name", "Last Name", "Email Address", "Phone Number", "Message")
action_phrases <- c("Save", "Delete", "Edit", "Share", "Download", "Print")

# Define comprehensive language set
target_languages <- c("es", "fr", "de", "it", "pt", "nl", "sv", "no")

# Create separate translation tables for each category
navigation_table <- create_translation_table(navigation_phrases, target_languages)
form_table <- create_translation_table(form_phrases, target_languages)
action_table <- create_translation_table(action_phrases, target_languages)

# Add category identifier
navigation_table$category <- "Navigation"
form_table$category <- "Forms"
action_table$category <- "Actions"

# Combine into master translation table
master_table <- rbind(navigation_table, form_table, action_table)

# Display summary
print(paste("Total phrases translated:", nrow(master_table)))
print(paste("Languages covered:", length(target_languages)))

Working with Tibbles and Tidyverse

For more advanced data manipulation, you can use the tidyverse ecosystem:

library(tibble)
library(dplyr)

# Create a more structured approach
content_data <- tibble(
  phrase_id = 1:8,
  category = c("greetings", "greetings", "farewells", "farewells", 
               "questions", "questions", "responses", "responses"),
  priority = c("high", "medium", "high", "medium", 
               "high", "medium", "high", "medium"),
  english = c("Hello", "Good morning", "Goodbye", "See you later",
              "What is your name?", "Where are you from?", 
              "My name is...", "I am from...")
)

# Create translations for priority languages first
priority_languages <- c("es", "fr", "de")
translations <- create_translation_table(content_data$english, priority_languages)

# Combine with original metadata
final_table <- bind_cols(content_data, translations[, -1])  # Remove duplicate original_word column

# Filter and organize as needed
high_priority <- final_table %>% 
  filter(priority == "high") %>%
  select(phrase_id, category, english, es, fr, de)

print(high_priority)

Export Options

Translation tables can be exported in various formats for different use cases:

# For web development - JSON-like structure
phrases <- c("Welcome", "Login", "Logout")
web_translations <- create_translation_table(phrases, c("es", "fr", "de"))

# Create key-value structure for each language
create_json_structure <- function(table, lang) {
  keys <- tolower(gsub(" ", "_", table$original_word))
  values <- table[[lang]]
  structure <- setNames(values, keys)
  return(structure)
}

spanish_json <- create_json_structure(web_translations, "es")
french_json <- create_json_structure(web_translations, "fr")

print("Spanish translations:")
print(spanish_json)

# For documentation purposes - wide format table
doc_table <- web_translations %>%
  rename(English = original_word, Spanish = es, French = fr, German = de)

print("Documentation table:")
print(doc_table)

Best Practices

When working with translation tables:

1. Consistency: Use the same source language throughout your project
2. Context: Group related phrases for better translation context
3. Validation: Compare multiple APIs for critical translations
4. Organization: Use categories and metadata to organize large translation sets
5. Backup: Save translation tables as they represent significant work

Conclusion

The create_translation_table() function provides a powerful foundation for multilingual workflows. By combining it with tidyverse tools and structured approaches, you can efficiently manage localization projects of any size. The tabular format makes it easy to review, validate, and export translations for various applications.

polyglotr documentation built on Aug. 8, 2025, 7:14 p.m.