2020-06-13
- Support for RMarkdown documents (see
rmd_node
). - New option to skip errors when evaluationg jobs. Use either
onError="skip"
argument ofmake()
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.
- 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
2020-06-13
- 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
andsapplyWith()
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 anode_list
). - Improved documentation.
- Added a
NEWS.md
file to track changes to the package.
- eval() is now a universal wrapper function of process() method (where all the node-type specific code has been moved into).
2020-05-28
- New
py_node
andjulia_node
node types allow using Python (via reticulate) and Julia (via JuliaCall).