Skip to contents

Changelog

pm 0.1.14

Bug Fixes

  • Avoid SLURM result overrides when submitting multiple jobs in a tight loop by adding a 4-character UUID suffix to the auto-generated result_id (while leaving user-specified result_id values unchanged)

pm 0.1.13

New Features

  • Enhance store_image SLURM config option to allow saving only selected objects (instead of the entire workspace or nothing)

Improvements

  • Document and exemplify the store_image SLURM config option in the SLURM integration vignette, including how to save only selected objects when your workspace contains large, unneeded data

pm 0.1.12

Bug Fixes

  • Fix SLURM tests so they pass under R CMD check: use full path to Rscript (R requires a path, not bare "Rscript") in the mock job runner

Improvements

  • Expand test coverage for SLURM, analysis, project, parse_inputs, and file writing: add tests for print status (Running/Completed/Failed), config options (store_image, slurm_flags, modules, result_id), cancel failure, parse_inputs error paths and absolute paths, schema validation, pm_write_file object_names, pm_create_project .gitignore copy failure, and analysis project access when not associated with a project
  • Mock SLURM: when job is not done, squeue now returns "PENDING" so is_done() and print status behave correctly in tests

pm 0.1.11

New Features

  • Support subfolders in PMAnalysis$get_output_path
  • Support subfolders in PMAnalysis$list_outputs
  • Support subfolders in get_artifact (both PMAnalysis and PmProject)

pm 0.1.10

New Features

  • Support multiple names for the code folder name apart from “code” (also “src” and “scripts)
  • Update the template README to include reference to this package :)

pm 0.1.9

New Features

  • Added pm_infer_analysis to easily use package when writing code
  • Added active binding project to PMAnalysis object to easily access containing project
  • Add run_in_slurm() method to PMAnalysis for running R functions in SLURM batch jobs
  • Add PMSlurmRun R6 class for managing SLURM job execution (non-blocking submission, status checking, result retrieval)
  • Support for passing function arguments via ... and SLURM configuration via config parameter
  • Automatic job status checking with is_done() and is_successful() methods
  • Support for timeout-based blocking result retrieval with get_results(timeout = ...)
  • Results saved as RDS files for direct loading (intermediate artifact)
  • Generic SLURM script templates using environment variables for configuration
  • Comprehensive test suite with mock SLURM implementation for testing without actual SLURM

pm 0.1.8

New Features

  • Added pm_infer_analysis to easily use package when writing code
  • Added active binding project to PMAnalysis object to easily access containing project

pm 0.1.7

Major Changes

  • Renamed inputs.yaml to project.yaml for future features generalization.
  • Added TL;DR to the getting-started vignette
  • Added a README.md to the package

Minor Changes

  • Removed PRODUCT.md

pm 0.1.6

New Features

  • Add comprehensive vignettes: “Getting Started”, “Working with File Formats”, and “Defining Project Inputs”
  • Add get_intermediate_artifact() method to PMAnalysis for accessing intermediate files from the current analysis
  • Add support for searching artifacts across all analyses by explicitly passing analysis_name = NULL
  • Improved documentation and examples for artifact discovery and file format handling
  • Add knitr and rmarkdown to Suggests for vignette building

pm 0.1.5

New Features

  • Add get_artifact() method to PMProject for fetching output files from analyses by ID
  • Add get_artifact() method to PMAnalysis for fetching output files from other analyses
  • Add list_outputs() method to PMAnalysis for listing all output files in an analysis
  • Support for searching artifacts across all analyses or within a specific analysis
  • Automatic detection of ambiguous artifacts with helpful error messages

pm 0.1.4

New Features

  • Add get_output_path() method to PMAnalysis for generating output file paths
  • Support for type-based file extension inference (table, object, image, figure, parquet, pqt)
  • Automatic extension validation when both name and type are provided
  • Support for intermediate vs outputs folder selection

pm 0.1.3

New Features

  • Add PMAnalysis R6 class to manage analysis folders within projects
  • Add create_analysis() method to PMProject for creating new analyses from template
  • Add list_analyses() method to PMProject to get all analysis names
  • Add get_analysis() method to PMProject to retrieve analysis objects by name
  • Analysis template structure includes: README.md, code/, outputs/, intermediate/, logs/, and .gitignore
  • Template-based analysis creation using inst/extdata/template_analysis directory
  • Support for dot_ prefix convention in templates (files starting with dot_ are renamed to start with . when copied)
  • PMAnalysis can be constructed from project object + name, or directly from path
  • PMAnalysis$new() automatically validates the analysis structure on initialization

pm 0.1.2

Improvements

  • Enhanced input file validation in PMProject$validate() method
  • Improved error messages for missing input files with detailed information:
    • Shows both original path (from YAML) and resolved absolute path
    • Provides actionable guidance on how to fix issues
    • Includes examples for adding missing entries
    • Better handling of edge cases (empty inputs.local.yaml, missing paths key)
  • Fixed bug where validation could fail when inputs.local.yaml is empty or missing paths key
  • Added comprehensive test coverage for validation edge cases and error message formats

pm 0.1.1

New Features

  • Add pm_read_file() function to read files based on their extension
  • Support reading CSV, TSV, Parquet (.parquet, .pqt), RDS, and RData (.rdata, .rda) files
  • Add read() method to PMData class for reading data files
  • For RData files, objects are loaded into a new environment and returned
  • Add pm_write_file() function to write files based on their extension
  • Support writing CSV, TSV, Parquet (.parquet, .pqt), RDS, and RData (.rdata, .rda) files
  • Add write() method to PMData class for writing data files
  • For RData files, can write multiple objects using ... (named or unnamed)
  • Validate object types: tabular formats (CSV, TSV, Parquet) require data.frame objects
  • Add comprehensive tests for file reading and writing functionality
  • Add round-trip integration tests for read/write operations through PMData
  • Add arrow package as suggested dependency for Parquet file support

pm 0.1.0

New Features

  • Add PMData R6 class to represent input data files with ID and path
  • Add parse_inputs() method to PMProject for parsing inputs.yaml and inputs.local.yaml
  • Implement YAML schema validation for inputs.yaml (supports both array and object formats)
  • Implement YAML schema validation for inputs.local.yaml
  • Support flexible input definitions: inputs can be just IDs (strings) or objects with metadata fields
  • Support extra fields in input definitions (md5, size, description)
  • Handle relative and absolute paths in inputs.local.yaml (automatic conversion to absolute paths)
  • Add comprehensive edge case tests for input parsing
  • Add great error messages for invalid YAML structures
  • Add yaml package as dependency

pm 0.0.1

  • Implement PMProject R6 class for project folder management with validation
  • Add pm_create_project() and pm_project() functions for creating and loading projects
  • Add project template structure with standard files and directories
  • Establish GitHub Actions workflows for CI/CD (R-CMD-check, test coverage, version/NEWS validation)
  • Add comprehensive test suite using testthat
  • Add dependencies: R6, chk (imports); testthat, withr (suggests)