Base code for exercises of the session on "AI for coding".

# Base code for exercises

This is the code from which you can start working on the exercises of the session on "AI for coding".

debug-function.R

# This function is supposed to calculate the median of a numeric vector
# after removing NA values, but it has an error
calculate_median <- function(x) {
  # Remove NA values
  clean_data <- na.omit(x)
  
  # Sort the data
  sorted_data <- sorted(clean_data) 
  
  # Calculate the median
  n <- length(sorted_data)
  if (n %% 2 is 1) {
    # Odd number of elements
    median_value <- sorted_data[(n + 1 / 2]
  } else {
    # Even number of elements
    median_value <- (sorted_data[n / 2] + sorted_data[(n / 2) - 1]) / 2
  }
  
  return(median_value)
}

# Test data
test_vector <- c(5, 2, NA, 9, 1, NA, 3)
result <- calculate_median(test_vector)
print(result)

document-function.R

# This function needs proper documentation
analyze_sales_data <- function(sales, dates, categories, min_value = 0) {
  valid_idx <- which(sales > min_value)
  
  sales <- sales[valid_idx]
  dates <- dates[valid_idx]
  categories <- categories[valid_idx]
  
  total_sales <- sum(sales)
  avg_sales <- mean(sales)
  
  sales_by_category <- tapply(sales, categories, sum)
  
  dates <- as.Date(dates)
  monthly_sales <- tapply(sales, format(dates, "%Y-%m"), sum)
  
  yearly_trend <- tapply(sales, format(dates, "%Y"), mean)
  
  result <- list(
    total = total_sales,
    average = avg_sales,
    by_category = sales_by_category,
    monthly = monthly_sales,
    yearly_trend = yearly_trend
  )
  
  return(result)
}

# Example usage
sales <- c(120, 250, 30, 45, 190, 320, 15, 80)
dates <- c("2023-01-15", "2023-01-28", "2023-02-10", "2023-02-22", 
           "2023-03-05", "2023-03-18", "2023-04-02", "2023-04-15")
categories <- c("Electronics", "Furniture", "Clothing", "Electronics", 
                "Furniture", "Electronics", "Clothing", "Clothing")

results <- analyze_sales_data(sales, dates, categories)

optimize-function.R

# This function calculates the cumulative sum of squares for values in a vector
# It works correctly but is very inefficient for large vectors
slow_cumulative_sum_squares <- function(vector) {
  result <- numeric(length(vector))
  
  for (i in 1:length(vector)) {
    sum_squares <- 0
    for (j in 1:i) {
      sum_squares <- sum_squares + vector[j]^2
    }
    result[i] <- sum_squares
  }
  
  return(result)
}

# Test with a small vector
small_test <- c(1, 2, 3, 4, 5)
result_small <- slow_cumulative_sum_squares(small_test)
print(result_small)

# This would be very slow with a large vector
# large_test <- rnorm(10000)
# system.time(result_large <- slow_cumulative_sum_squares(large_test))