Changes in version 2.1.1 (2026-05-12)                  

This is the first CRAN submission version of easytable.

Compared with the earlier GitHub and Zenodo 2.1.0 release, this version
includes minor bug fixes, documentation updates, and CRAN-requested
formatting changes.

                        Changes in version 2.1.0                        

Minor Updates

  - pkgdown installation instructions now use the R-universe repository
    path.
  - pkgdown reference navigation now focuses on easytable(),
    format_word(), and format_latex().
  - Scientific-notation display, interaction-term rendering, and Word
    layout were refined for small or large estimates.

Breaking Changes

  - Removed markdown output mode from easytable(). Supported outputs are
    now "word" (default) and "latex".

  - Renamed CSV export argument:
    
      - Old: csv
      - New: export.csv (must end with .csv)

New Features

  - Direct Word file export with new export.word argument (must end with
    .docx).
    
      - Example: easytable(m1, m2, export.word = "mytable.docx")

  - Interaction term wrapping in term column: interaction labels now
    wrap after * to reduce table width (e.g., x1 * on one line, x2 on
    the next line).

Formatting and Consistency

  - Word/HTML and LaTeX alignment improved:
      - Shared measure-row detection for divider placement
      - Single divider between coefficient block and model-stat block
      - Zebra striping constrained to coefficient block only

Documentation

  - Rewrote README.md around a clearer user promise: easy to use, easy
    to read.
  - Added a dedicated penguins tutorial vignette: penguins-tutorial.
  - Added a developer roadmap vignette: developer-roadmap.
  - Added architecture and maintenance guides:
      - DESIGN_PHILOSOPHY.md
      - AI_NOTES.md
  - Reorganized _pkgdown.yml navigation and article structure.
  - Added LICENSE.md so the MIT license text is human-readable in the
    repository.

Testing Infrastructure

  - Added tests/README.md with a deterministic, CI-safe testing
    protocol.
  - Added tests/run-tests.R with a committed core profile.
  - Added tests/testthat/helper-testing-profiles.R for environment-aware
    test skipping.
  - Added tests/testthat/test-design-invariants.R to lock key table
    invariants.

API and Validation UX

  - Added validate_output_format() to provide friendly, explicit
    output-format validation.
  - Reworded key validation errors to be more beginner-friendly and
    actionable.

                        Changes in version 2.0.1                        

New Features

  - Direct model interface: New easytable() function accepts model
    objects directly through dots, like easytable(m1, m2, m3). This
    provides a more intuitive API for multiple model input.

  - Default model naming: Model columns are now automatically named
    "Model 1", "Model 2", etc. when using the new easytable() interface.

  - Custom model names: Added model.names parameter to specify custom
    column names for models. Example: easytable(m1, m2, model.names =
    c("Baseline", "Full"))

  - Enhanced term label formatting: Term labels are now automatically
    formatted for improved readability:
    
      - Factor levels separated with colon (e.g.,
        digital_confidence:low)
      - Interactions displayed with asterisk and spaces (e.g., var1 *
        var2)
      - Polynomial contrasts shown as L indices (e.g., var:L1 for .Q,
        var:L2 for .L)
      - Common variable names abbreviated (e.g., fin.prud for
        financial_prudence)

Bug Fixes

  - Fixed markdown output: Resolved issue with stray vertical bars
    appearing inside table cells. Markdown tables now use <br> for line
    breaks within cells, ensuring proper rendering.

  - Improved markdown formatting: Ensured coefficient formatting (stars,
    standard errors) is consistent across markdown and LaTeX output
    formats.

Backward Compatibility

  - The original easy_table() function with list-based interface remains
    fully functional for backward compatibility.

  - All existing code using easy_table(list(Model1 = m1, Model2 = m2))
    continues to work without changes.

Migration from 2.0.0

Most users can adopt the new interface immediately:

# Old way (still works)
models <- list(Model1 = m1, Model2 = m2)
easy_table(models)

# New way (recommended)
easytable(m1, m2)

# With custom names
easytable(m1, m2, model.names = c("Model1", "Model2"))

                        Changes in version 2.0.0                        

Major Changes

This is a complete rewrite of easytable with breaking changes.
Version 2.0.0 modernizes the package architecture and dramatically
expands functionality.

New Features

  - Multi-format output: Now supports Word (via flextable), Markdown
    (for Quarto/RMarkdown), and LaTeX/PDF output
    
      - New output parameter: "word", "markdown", or "latex"
      - Default remains "word" for backward compatibility

  - Modular architecture: Complete refactor into clean, maintainable
    modules
    
      - parse_models.R: Extract coefficients and statistics
      - transform_table.R: Handle control variables and organization
      - format_word.R, format_markdown.R, format_latex.R:
        Format-specific renderers
      - validators.R: Comprehensive input validation

  - Enhanced validation: Informative error messages with clear guidance
    
      - Validates model types and structure
      - Checks for required packages and provides installation
        instructions
      - Warns about control variables not found in models

  - Better dependency management:
    
      - Only core dependencies (broom, dplyr, magrittr) are always
        loaded
      - Format-specific packages loaded conditionally
      - Clear error messages when optional packages are missing

Breaking Changes

⚠️ IMPORTANT: The following changes may affect existing code:

1.  Removed runtime installation (CRITICAL SECURITY FIX)
    
      - OLD: Package automatically installed missing dependencies
      - NEW: Users must install dependencies themselves
      - If you get an error about missing packages, install them with:
        install.packages(c("flextable", "lmtest", "sandwich", "margins"))

2.  Function parameter changes
    
      - OLD: F and T were accepted
      - NEW: Must use FALSE and TRUE (best practice)
      - This shouldn't affect most code, but if you passed variables
        named F or T, they won't work

3.  Improved join operations
    
      - Fixed warnings from dplyr about unspecified join columns
      - No user-facing changes, but cleaner console output

Bug Fixes

  - Fixed control variable regex bug: Previously, searching for control
    variable "hp" would also match "hpq". Now uses word boundaries for
    exact matching.

  - Fixed deduplication issue: Improved handling of duplicate control
    variable rows (previously noted as "slight problem" in code)

  - Explicit namespace calls: All function calls use explicit namespaces
    (e.g., dplyr::filter) to prevent masking issues

  - Better handling of GLM models: Improved extraction of fit statistics
    for generalized linear models

Documentation

  - Comprehensive README: Installation, quick start, and feature
    overview
  - Detailed vignette: vignette("easytable-intro") with working examples
  - Function documentation: Complete roxygen2 documentation with
    examples
  - Developer guide: AI_NOTES.md with architecture and contributor
    guidelines

Testing

  - Full test coverage:
      - Unit tests for all modules
      - Integration tests for full pipeline
      - Tests for all output formats and parameter combinations
      - Using testthat framework

Performance

  - Modular design enables easier maintenance and future enhancements
  - Conditional loading of packages reduces memory footprint
  - Cleaner code with proper separation of concerns

Migration Guide: 0.1.0 → 2.0.0

If You're Using Basic Features

Good news: Most basic usage should work without changes!

# This still works exactly the same
easy_table(model_list)

If You Had Issues with Missing Packages

Old behavior (0.1.0):

# Package would auto-install dependencies (security risk!)
easy_table(models)
# Installing package into '~/R/library'...

New behavior (2.0.0):

# Get clear error with installation instructions
easy_table(models)
# Error: Package 'flextable' is required for Word output.
#   Install it with: install.packages('flextable')

# You install once:
install.packages("flextable")

# Then use normally:
easy_table(models)

If You Want Markdown or LaTeX Output

New in 2.0.0:

# Markdown for Quarto/RMarkdown
easy_table(models, output = "markdown")

# LaTeX for PDF
easy_table(models, output = "latex")

If You're Using Advanced Features

All advanced features work the same way:

# Robust SE - still works
easy_table(models, robust.se = TRUE)

# Marginal effects - still works
easy_table(models, margins = TRUE)

# Control variables - still works (now with fixed regex bug!)
easy_table(models, control.var = c("species", "island"))

# Highlighting - still works
easy_table(models, highlight = TRUE)

# CSV export - still works
easy_table(models, csv = "results")

Package Dependencies

Install what you need:

# For basic usage (Word output)
install.packages(c("broom", "dplyr", "flextable"))

# For robust standard errors
install.packages(c("lmtest", "sandwich"))

# For marginal effects
install.packages("margins")

# For Markdown/LaTeX output
install.packages("knitr")
install.packages("kableExtra")  # optional, for enhanced formatting

What Stays the Same

✅ Function name: easy_table() ✅ Default behavior: Word output via
flextable ✅ Parameter names: robust.se, margins, control.var, highlight,
csv ✅ Model support: lm, glm ✅ Significance stars: *** p < .01, ** p <
.05, * p < .1 ✅ Model fit statistics: N, R sq., Adj. R sq., AIC (glm
only)

What's Different

🔄 Must install dependencies manually (no more auto-install) 🆕 New output
formats: markdown and latex 🐛 Control variable matching now works
correctly 📚 Much better documentation and examples ✅ Comprehensive test
coverage 🏗️ Cleaner, more maintainable code architecture

Detailed Changes by Component

Parse Models (R/parse_models.R)

  - Extracted from monolithic function
  - Added parse_single_model() for individual model parsing
  - Added format_coefficients() for significance star formatting
  - Added extract_model_measures() for fit statistics
  - Added parse_models() for multi-model parsing
  - All functions properly documented with roxygen2

Transform Table (R/transform_table.R)

  - Added collapse_control_vars() with fixed regex patterns
  - Added mark_control_vars() for Y indicators
  - Added deduplicate_control_vars() for cleaning duplicates
  - Added sort_table() for proper ordering
  - Added separate_measures() for bottom placement
  - Added transform_table() as main orchestrator

Format Renderers

  - Word (R/format_word.R): Extracted and cleaned flextable code
  - Markdown (R/format_markdown.R): NEW - knitr/kableExtra support
  - LaTeX (R/format_latex.R): NEW - LaTeX with booktabs

Validators (R/validators.R)

  - Added validate_model_list() with clear error messages
  - Added validate_model_types() for supported model checking
  - Added validate_control_vars() with warnings for missing vars
  - Added validate_parameters() for type checking
  - Added dependency checkers for all optional features

Main Function (R/easytab.R)

  - Complete rewrite using modular components
  - Added output parameter with validation
  - Integrated validators → parser → transformer → formatter pipeline
  - Comprehensive roxygen2 documentation with examples
  - Maintains backward compatibility

Future Roadmap

The modular architecture in 2.0.0 makes it easy to add:

  - New model types: plm (panel data), fixest (fixed effects), survival
    models
  - New output formats: HTML tables, gt tables, custom formats
  - Theme system: Customizable table styling
  - More statistics: Confidence intervals, standardized coefficients
  - Enhanced features: Multi-equation models, nested model tests

Credits

  - Original package: Alfredo Hernandez Sanchez
  - 2.0.0 rewrite: Comprehensive modernization and expansion
  - Testing data: Palmer Penguins dataset (Horst et al.)

Feedback

Found a bug? Have a feature request? Please open an issue on GitHub:
https://github.com/alfredo-hs/easytable/issues

                        Changes in version 0.1.0                        

Initial release (2024)

  - Basic functionality for creating Word regression tables
  - Support for lm and glm models
  - Robust standard errors via lmtest/sandwich
  - Marginal effects via margins package
  - Control variable grouping
  - Highlighting of significant results
  - CSV export

Note: Version 0.1.0 had critical issues:

  - Runtime package installation (security risk)
  - Monolithic code structure (hard to maintain)
  - Control variable regex bug
  - Limited to Word output only
  - No tests or comprehensive documentation

All issues resolved in 2.0.0.