networkformat 0.1.2

CRAN resubmission

• Reworded the DESCRIPTION Description field so it no longer opens by restating the package title, and put ‘tidyselect’ in single quotes alongside the other software names, per CRAN review feedback on 0.1.1.
• No user-facing changes to code, documentation or behaviour.

networkformat 0.1.1

Features

• edgelist() generic with methods for:
  • Atomic vectors — sequential edges connecting element i to i + 1
  • data.frame — column-pair edges with tidyselect, na.rm, symmetric_cols, dedupe, and weights
  • randomForest — parent-child splits with treenum filtering
  • tree — parent-child splits with parsed split components
  • rpart — parent-child splits with parsed split components
  • xgb.Booster — parent-child splits with feature/split/quality/cover
  • gbm — parent-child splits with treenum filtering
• nodelist() generic with methods for:
  • Atomic vectors — unique values with frequency counts
  • data.frame — reorder with id_col first
  • randomForest — node attributes per tree
  • tree — node attributes with labels
  • rpart — node attributes with labels
  • xgb.Booster — node attributes with labels
  • gbm — node attributes per tree
• as.igraph() methods for one-step igraph construction from tree, randomForest, rpart, xgb.Booster, and gbm models — registered against igraph’s as.igraph() generic via delayed S3 registration
• as_tbl_graph() methods for one-step tbl_graph construction from the same model types — registered against tidygraph’s as_tbl_graph() generic via delayed S3 registration
• weights parameter for edgelist.data.frame and vector method — collapses duplicate rows and adds a weight count column
• Bundled courses dataset for examples
• 3 vignettes: package introduction, edgelist/nodelist guide, visualization
• Comprehensive test suite (275+ tests)