WIBA: What Is Being Argued?

Here's a simple example to get started with WIBA in R:

# Import and initialize
wiba <- import("wiba")
analyzer <- wiba$WIBA(api_token="your_api_token_here")

# Example text
text <- "Climate change is real because global temperatures are rising."

# Detect if it's an argument
result <- analyzer$detect(text)
print(paste("Argument detected:", result$argument_prediction))
print(paste("Confidence:", result$confidence_score))

The detect() method identifies whether a text contains an argument:

# Single text
result <- analyzer$detect("Climate change is real because temperatures are rising.")
print(result$argument_prediction)  # "Argument" or "NoArgument"
print(result$confidence_score)     # Confidence score between 0 and 1

# Multiple texts
texts <- c(
    "Climate change is real because temperatures are rising.",
    "This is just a simple statement without any argument."
)
results <- analyzer$detect(texts)
for (r in results) {
    print(paste("Text:", r$text))
    print(paste("Prediction:", r$argument_prediction))
}

# Using data.frame
library(data.frame)
df <- data.frame(text = texts)
results_df <- analyzer$detect(df, text_column='text')

The extract() method identifies the main topic being argued about:

# Single text
result <- analyzer$extract("Climate change is a serious issue because it affects our environment.")
print(result$topics)  # List of extracted topics

# Multiple texts
texts <- c(
    "Climate change is a serious issue because it affects our environment.",
    "We need better healthcare because current systems are inadequate."
)
results <- analyzer$extract(texts)
for (r in results) {
    print(paste("Text:", r$text))
    print(paste("Topics:", paste(r$topics, collapse=", ")))
}

# Using data.frame
df <- data.frame(text = texts)
results_df <- analyzer$extract(df, text_column='text')

The stance() method determines the stance towards a specific topic:

# Single text
text <- "We must take action on climate change because the evidence is overwhelming."
topic <- "climate change"
result <- analyzer$stance(text, topic)
print(paste("Stance:", result$stance))  # "Favor", "Against", or "NoArgument"

# Multiple texts
texts <- c(
    "We must take action on climate change because the evidence is overwhelming.",
    "Climate change policies will harm the economy and cost jobs."
)
topics <- c("climate change", "climate change")
results <- analyzer$stance(texts, topics)
for (r in results) {
    print(paste("Text:", r$text))
    print(paste("Topic:", r$topic))
    print(paste("Stance:", r$stance))
}

# Using data.frame
df <- data.frame(
    text = texts,
    topic = topics
)
results_df <- analyzer$stance(df, text_column='text', topic_column='topic')

The discover_arguments() method finds argumentative segments in longer texts:

# Single text
text <- "Climate change is a serious issue. Global temperatures are rising at an 
unprecedented rate. This is causing extreme weather events. However, some argue 
that natural climate cycles are responsible."

# Parameters are automatically converted to integers internally
results_df <- analyzer$discover_arguments(
    text,
    window_size=2,  # Number of sentences per window
    step_size=1     # Number of sentences to move window
)

# Alternatively, you can explicitly convert to integers if needed
results_df <- analyzer$discover_arguments(
    text,
    window_size=as.integer(2),
    step_size=as.integer(1)
)

print(results_df[c('text_segment', 'argument_prediction', 'argument_confidence')])

# Using data.frame
df <- data.frame(text = c(text1, text2))
results_df <- analyzer$discover_arguments(
    df,
    text_column='text',
    window_size=2,
    step_size=1
)

All methods support batch processing for efficient handling of multiple texts:

# Process large datasets
library(readr)
library(dplyr)

# Read data
df <- read_csv('texts.csv')

# Convert to Python-friendly format (if needed)
df <- as.data.frame(df)  # ensure it's a base R data frame

# Process in batches
results_df <- analyzer$detect(
    df,
    text_column='text',
    batch_size=100,    # Number of texts per batch
    show_progress=TRUE # Show progress bar
)

# Results are automatically converted to R data frame
# Save results
write_csv(results_df, 'results.csv')

The extract() method identifies the main topic being argued about:

# Single text
result <- analyzer$extract("Climate change is a serious issue because it affects our environment.")
topics <- py_to_r(result$topics)  # convert Python list to R vector
print(topics)

# Multiple texts using vectorization
texts <- c(
    "Climate change is a serious issue because it affects our environment.",
    "We need better healthcare because current systems are inadequate."
)
results <- analyzer$extract(texts)

# Process results using R's native functions
results_df <- data.frame(
    text = sapply(results, function(r) r$text),
    topics = sapply(results, function(r) paste(py_to_r(r$topics), collapse=", "))
)

# Using data.frame
df <- data.frame(text = texts)
results_df <- analyzer$extract(df, text_column='text')

Error handling in R follows R's conventions while using Python objects:

# Error handling using R's tryCatch
tryCatch({
    # Missing required column
    bad_df <- data.frame(wrong_column = c("test"))
    analyzer$detect(bad_df)
}, error = function(e) {
    message(sprintf("Validation error: %s", e$message))
})

# Handle Python exceptions
tryCatch({
    result <- analyzer$detect("some text")
}, error = function(e) {
    if (inherits(e, "python.builtin.Exception")) {
        message(sprintf("Python error: %s", py_str(e)))
    } else {
        message(sprintf("R error: %s", e$message))
    }
})