Skip to content

Latest commit

 

History

History
45 lines (29 loc) · 1.69 KB

NEWS.md

File metadata and controls

45 lines (29 loc) · 1.69 KB
Raw

Rflow 0.3.3

2020-06-13

Changes to the API

  • Support for RMarkdown documents (see rmd_node).
  • New option to skip errors when evaluationg jobs. Use either onError="skip" argument of make() or RFLOW_ON_ERRORS global option.
  • new_rflow() is deprecated for rflow()
  • New .DATA() and .NODES() for referencing values of other nodes or node objects inside R jobs.
  • Improved automatic checking of dependencies between nodes. .RFLOW[[depends[n]]] is now recognized and does not throw warnings.
  • igraph package is no longer a dependency.

Bug fixes

  • Argument read_args in excel_sheet nodes are now functional.
  • Changes in read_args are now persistent.
  • Updating node definitions to NULL (or just leaving the default value) is now handled correctly.
  • csv_node nodes' definitions can be updated

Rflow 0.3.2

2020-06-13

Visible new features

  • Names of cache files now contain hash values of the data stored inside them to speed up its loading (true lazy/delayed is now possible).
  • New FilterWith, lapplyWith and sapplyWith() family of functions to make it easier to apply vectorized functions on nodes.
  • Rflow now records time needed to evaluate a node obtainable via node$time_to_eval().
  • Indexing a node_list object now preserves class attribute (result is still a node_list).
  • Improved documentation.
  • Added a NEWS.md file to track changes to the package.

Internal changes

  • eval() is now a universal wrapper function of process() method (where all the node-type specific code has been moved into).

Rflow 0.2.23

2020-05-28

Visible new features

  • New py_node and julia_node node types allow using Python (via reticulate) and Julia (via JuliaCall).