In sailuh/kaiaulu: Kaiaulu

Introduction

Within a project, we might want to see the in-between file or in-between class dependencies. This notebook showcases how to obtain either dependencies using Scitool's Understand). Note you will need a license from Scitools to use this Notebook.

rm(list = ls())
require(kaiaulu)
require(visNetwork)
require(XML)
require(stringi)
require(igraph)
require(data.table)
require(gt)

Project Configuration File

For this notebook, we will use Apache Helix as an example. Refer to Helix project configuration file under the conf/ folder, in particular the understand section. For supported languages, see scitools documentation.

tool <- parse_config("../tools.yml")
scitools_path <- get_tool_project("scitools", tool)

conf <- parse_config("../conf/helix.yml")
keep_dependencies_type <- get_understand_keep_dependencies_type(conf)
project_path <- get_understand_project_path(conf)

# Scitools
understand_folder <- get_understand_output_path(conf)
code_language <- get_understand_code_language(conf)

db_path <- stringi::stri_c(understand_folder,"Understand.und")

file_dependencies_path <- stringi::stri_c(understand_folder,"file_dependencies.xml")
class_dependencies_path <- stringi::stri_c(understand_folder,"class_dependencies.xml")

Build Understand DB

To export dependencies, scitools require a und database to be created:

db_path <- build_understand_project(scitools_path = scitools_path, project_path = project_path, language = code_language, output_dir = understand_folder)

Export Dependencies

Subsequently, we can make our dependencies export out of the database. We can either export XML files or class dependencies.

File Dependencies

file_dependencies_path <- export_understand_dependencies(scitools_path = scitools_path, db_filepath = db_path, parse_type = "file", output_filepath = file_dependencies_path)

The XML files can then be parsed into tables for manipulation in Kaiaulu:

file_dependencies <- parse_understand_dependencies(dependencies_path = file_dependencies_path)

head(file_dependencies[["node_list"]])  %>%
  gt(auto_align = FALSE)

head(file_dependencies[["edge_list"]])  %>%
  gt(auto_align = FALSE)

Class Dependencies

The process to export and parse class dependencies is identical, save for specifying the parse_type to "class":

class_dependencies_path <- export_understand_dependencies(scitools_path = scitools_path, db_filepath = db_path, parse_type = "class", output_filepath = class_dependencies_path)

class_dependencies <- parse_understand_dependencies(dependencies_path = class_dependencies_path)

head(class_dependencies[["node_list"]])  %>%
  gt(auto_align = FALSE)

head(class_dependencies[["edge_list"]])  %>%
  gt(auto_align = FALSE)

Network Visualization

We can display the parsed dependencies as a network. You can use the dependency_kind parameter to subset the dependencies used. Refer to Kaiaulu wiki for the types of dependencies Scitools support.

File Network

file_graph <- transform_understand_dependencies_to_network(parsed = file_dependencies, weight_types = c(keep_dependencies_type[2], keep_dependencies_type[5]))


project_function_network <- igraph::graph_from_data_frame(d=file_graph[["edge_list"]], 
                      directed = TRUE,
                      vertices = file_graph[["node_list"]])
visIgraph(project_function_network,randomSeed = 1)

Class Network

The same applies to outputting our class_dependencies, but we can visualize what our class dependency data is using the same parameters sans the parsed data table.

class_graph <- transform_understand_dependencies_to_network(parsed = class_dependencies, weight_types = c(keep_dependencies_type[2], keep_dependencies_type[5]))

project_function_network <- igraph::graph_from_data_frame(d=class_graph[["edge_list"]], 
                      directed = TRUE,
                      vertices = class_graph[["node_list"]])
visIgraph(project_function_network,randomSeed = 1)