mem_function_notes.md

In jmw86069/multienrichjam: Analysis and Visualization of Multiple Gene Set Enrichments

Keep track of igraph-related functions that call other functions

subsetCnetIgraph()

• removeIgraphSinglets() - optional, default=TRUE
• removeIgraphBlanks() - optional, default=TRUE
• spread_igraph_labels() - optional, default=TRUE
• reorderIgraphNodes() - optional, default=TRUE

spread_igraph_labels()

• layout_with_qfr() - called if force_relayout=TRUE, OR (layout=NULL and g$layout=NULL)

• layout() - only called if force_relayout=TRUE

• reorderIgraphNodes() - optional, default=TRUE

reorderIgraphNodes()

• returns g which always includes g$layout with layout coordinates
• layout() - called if layout is a function
• relayout_with_qfr() - called if layout=NULL, g$layout=NULL
• isColorBlank() - called with isColorBlank(...)

Re-consider function name style

Re-consider whether renaming functions with some consistent and intuitive naming scheme could be beneficial. I actually complain about my own inconsistency so I think it best to rename functions sooner than later. Old functions will be deprecated and aliased to the new function names.

Function Style Option A: type_verb_subtype()

For me it is sometimes helpful to see the relevant object type first in the function name, however this style may not be idea. For example:

• input_method_subtype()
• igraph_reorder_nodes() instead of reorderIgraphNodes()
• igraph_spread_labels() instead of spread_igraph_labels()
• igraph_subset_cnet() instead of subsetCnetIgraph()
• igraph_remove_singlets() instead of removeIgraphSinglets()
• igraph_remove_blanks() instead of removeIgraphBlanks()
• layout_with_qfr() for consistency with igraph::layout_with_() convention.
• relayout_with_qfr() consistent with layout_with_qfr()
• color_is_blank() possible instead of isColorBlank()

Function Style Option B: verb_type_subtype()

This option is preferred because it matches Tidyverse convention

A counter point, the tidyverse style guide suggests using verbs, all lowercase with underscore '_' between words (not camelCase). For example:

• method_input_typeaffected()
• reorder_igraph_nodes() instead of reorderIgraphNodes()
• spread_igraph_labels()
• subset_cnet_igraph() instead of subsetCnetIgraph()
• remove_igraph_singlets() instead of removeIgraphSinglets()
• remove_igraph_blanks() instead of removeIgraphBlanks()
• layout_with_qfr()
• relayout_with_qfr()
• is_color_blank() instead of isColorBlank()

New alternate function subset_igraph()?

Instead of:

• remove_igraph_singlets()
• remove_igraph_blanks()

One common function:

• subset_igraph()

- or prune_igraph() or filter_igraph() * remove_singlets=FALSE * remove_blanks=FALSE

• subset_cnet_igraph()

• same as above except option to subset by specific entries by V(g)$nodeType.

Tidyverse %>% and R-4.1 > pipe compatibility

Where possible and where relevant the functions should be compatible with tidyverse pipe %>% and the new R-4.1 operator pipe \>

In practice when functions are called for their side-effects, the function should return the same object supplied after applying modifications.

• relayout_with_qfr()
• spread_igraph_labels()
• subset_cnet_igraph()
• remove_igraph_singlets()
• remove_igraph_blanks()

Finding relevant functions

For convenience one can use jamba::grepls("igraph.") to find functions relevant to igraph objects.

jmw86069/multienrichjam documentation built on Feb. 7, 2024, 12:58 a.m.