Refactor combine_data function in transformations.R into smaller, manageable components

[Copilot]

Summary

This PR addresses the issue of the extremely long and complex combine_data function in transformations.R by breaking it down into smaller, more manageable helper functions.

Changes Made

Main Achievement:

• Reduced the main combine_data function from 355+ lines to 32 lines (91% reduction)
• Improved readability by clearly showing the logical flow
• Enhanced maintainability by splitting complex logic into focused helper functions
• Reduced overall complexity while preserving all existing functionality

Helper Functions Created:

1. .combine_data_validate_inputs() - Handles input validation and transformation setup
2. .combine_data_prepare_data() - Manages data frame preparation and column handling
3. .combine_data_validate_and_setup() - Performs data validation and output structure setup
4. .combine_data_handle_unstandardized() - Special handling for unstandardized effect sizes (y)
5. .combine_data_calculations() - Handles OR→logOR transformation, standard errors, sample sizes, and t-statistics calculations
6. .combine_data_transform() - Manages effect size transformations by type (d, z, r, logOR, OR)
7. .combine_data_create_output() - Creates final output with proper attributes

Before and After

Before: One monolithic 355+ line function handling all logic

combine_data <- function(...) {
  # 355+ lines of mixed logic covering:
  # - Input validation
  # - Data preparation
  # - Calculations
  # - Transformations
  # - Output creation
  # ... (extremely long and hard to follow)
}

After: Clean, readable coordinator function

combine_data <- function(...) {
  # Validate inputs and setup transformation
  transformation <- .combine_data_validate_inputs(...)
  
  # Prepare data frame  
  data_info <- .combine_data_prepare_data(...)
  
  # Validate data and prepare output structure
  setup_info <- .combine_data_validate_and_setup(...)
  
  # Handle unstandardized effect sizes
  if(!anyNA(data[,"y"])) {
    return(.combine_data_handle_unstandardized(...))
  }
  
  # Transform OR to logOR and perform calculations
  data <- .combine_data_calculations(data)
  
  # Transform effect sizes and standard errors
  data <- .combine_data_transform(data, transformation)
  
  # Prepare final output
  return(.combine_data_create_output(...))
}

Testing

• ✅ All existing tests continue to pass (128 transformation tests + 958 total package tests)
• ✅ No breaking changes to the external API
• ✅ All functionality preserved exactly as before
• ✅ No changes to function signatures or behavior

This refactoring significantly improves code maintainability and readability while ensuring complete backward compatibility.

Fixes #56.

---

💬 Share your feedback on Copilot coding agent for the chance to win a $200 gift card! Click here to start the survey.