Changes in version 1.0.1                        

Date: 2025-11-27

This release focuses on improving robustness on Windows systems,
enhancing flexibility for external tools, and standardizing data
structures.

IMPROVEMENTS

  - Flexibility in runPsScan():
    
      - Added optional arguments (database_path, ps_scan_path,
        pfscan_path) to allow users to provide their own paths to
        PROSITE files. The function now prioritizes user-provided paths
        over the internal cache, facilitating usage in offline
        environments or with custom databases.

  - Vignette:
    
      - Added a "Quick Start" section for immediate usage examples.
      - Clarified workflow descriptions to emphasize that visualization
        functions work interchangeably with data imported from GFF, PSA,
        or TXT formats.
      - Simplified data loading examples.

BUG FIXES

  - Windows Path Handling:
    
      - Fixed a critical issue in runPsScan() on Windows where
        backslashes in file paths caused the external Perl script to
        fail. The function now forces forward slashes (/) for all system
        calls.

  - Data Consistency in readPsa():
    
      - Fixed column data types in readPsa() (specifically ensuring
        character types for empty fields instead of logical NA).
      - Aligned the column order of readPsa() output to match the
        standard GRanges/data.frame structure returned by
        rtracklayer::import.gff. This ensures seamless compatibility
        between different input formats.

                       Changes in version 0.99.4                        

Date: 2025-07-17

This version includes a major refactoring to align with Bioconductor
submission standards, addressing all points from the initial review.

SIGNIFICANT CHANGES

  - The runPsScan() function has been completely refactored. It no
    longer downloads files to the working directory. Instead, it now
    uses BiocFileCache to download and cache the ps_scan.pl script,
    prosite.dat, and pfscan executables. This makes the function robust,
    reproducible, and compliant with Bioconductor guidelines for
    handling external data.

  - All system calls have been updated to use system2() instead of
    system() or shell(), providing a safer and more portable way to run
    the external Perl script.

  - The interactive OS confirmation prompt (readline()) has been removed
    from runPsScan() to make the function fully non-interactive and
    suitable for automated scripts.

IMPROVEMENTS

  - Code Style:
    
      - All function names have been standardized to use camelCase for
        consistency (e.g., read.prosite is now readProsite,
        extract_protein_motifs is now extractProteinMotifs, etc.).
    
      - Removed all print() and cat() statements from package functions.
        Status updates now correctly use message().
    
      - Replaced paste() in stop() and warning() calls with sprintf().
    
      - Removed all uses of suppressWarnings(), ensuring that important
        data coercion warnings are visible to the user.

  - Documentation:
    
      - Added a package-level manual page, accessible via ?PMScanR.
    
      - Added examples to all exported functions, including an
        interactive() example for runPMScanRShiny().

  - Vignette:
    
      - The vignette has been updated to include the standard
        Bioconductor installation instructions.
    
      - Removed setwd() and unnecessary if (file.exists(...)) checks to
        make the code cleaner and more focused on workflow.

  - DESCRIPTION file:
    
      - Added URL and BugReports fields pointing to the GitHub
        repository.
    
      - Added Perl to the SystemRequirements field.
    
      - Added BiocFileCache to Imports.

  - NAMESPACE:
    
      - Removed all broad import() statements. The package now
        exclusively uses specific importFrom() tags for all dependencies
        (bslib, bsicons, reshape2, etc.), preventing namespace
        conflicts.

  - Unit Tests:
    
      - Added a suite of unit tests using testthat for core data
        processing functions (gff2matrix, extractProteinMotifs,
        freqPie), increasing package robustness.

BUG FIXES

  - Fixed a bug in gff2matrix() where it would error on empty data frame
    input. It now correctly returns an empty matrix.

  - Fixed a bug in extractProteinMotifs() where the regular expression
    failed to correctly parse motif headers, causing it to return an
    empty list. The parsing logic is now more robust.