In greaterlouisvilleproject/glptools: Tools for analysis of Greater Louisville Project data

knitr::opts_chunk$set(echo = TRUE)
library(stringr)
library(magrittr)
library(dplyr)
library(tidyr)
library(visNetwork)
library(RColorBrewer)

Read in dependency data and format in terms of nodes and edges

dependencies <- readxl::read_excel("glpverse.xlsx")

dependencies %<>%
  pivot_longer(Calls:`...23`, values_to = "dependencies") %>%
  select(-name) 

nodes <- dependencies %>%
  select(-dependencies) %>%
  unique() %>%
  mutate(
    id = row_number(),
    label = Item,
    color = if_else(Type == "function", "blue", "green"))

nodes_minimal <- nodes %>% transmute(id, Item)

dependencies %<>%
  select(Item, dependencies) %>%
  filter(!is.na(dependencies)) %>%
  left_join(nodes, by = "Item") %>%
  rename(from=id) %>%
  right_join(nodes, by = c("dependencies" = "Item")) %>%
  rename(to=id) %>%
  filter(!is.na(from))

Create shiny document using vizNetwork

ui <- fluidPage(
  visNetworkOutput("network", height = "800px")
)

server <- function(input, output) {
  output$network <- renderVisNetwork({
    # minimal example
    visNetwork(nodes, dependencies) %>%
      visEdges(arrows = list(from = list(enabled = TRUE))) %>%
      visNodes(font = list(size = 40)) %>%
      visOptions(selectedBy = list(variable = "group", selected = "general_utils"),
                 nodesIdSelection = TRUE) %>%
      visOptions(#selectedBy = "selected",
                 highlightNearest = list(enabled = TRUE,
                                         algorithm = "hierarchical",
                                         degree = list(from = 0, to = 1),
                                         labelOnly = FALSE),
                 nodesIdSelection = TRUE)
  })
}

shinyApp(ui = ui, server = server)

Attempt to compile data programmatically...questionable

library(stringr)
library(magrittr)
library(dplyr)
library(tidyr)

# Create list of all files
all_files <- list.files("R")

# Don't include functions that are not top-level
dont_include <- c("fxn", # census_api
                  "bind_fxn", "line_1_2_fxn", "label_fxn", #make_map
                  "calc_output", "add_MOE") # microdata

for(f in all_files) {

  # read R file as string
  text <- readChar("R/" %p% f, file.info("R/" %p% f)$size)

  # Create list of function names and locations. If no functions, skip to next file.
  fxn_names <- str_extract_all(text, ".{1,100}(?= <- function)")[[1]] %>%
    as.character()

  n_fxns = length(fxn_names)

  if (n_fxns == 0) next

  fxn_locations <- str_locate_all(text, ".{1,100}(?= <- function)")[[1]]

  # Convert output to data frame
  fxn_locations <- data.frame(
    file = f,
    function_name = fxn_names,
    search_start = fxn_locations[,"end"],
    search_end = c(fxn_locations[2:n_fxns,"start"], nchar(text)))

  output <- assign_row_join(output, fxn_locations)
}

for(f in all_files) {

  # read R file as string
  text <- readChar("R/" %p% f, file.info("R/" %p% f)$size)

  # Create list of function names and locations. If no functions, skip to next file.
  fxn_names <- str_extract_all(text, ".{1,100}(?= <- function)")[[1]] %>%
    as.character()

  n_fxns = length(fxn_names)

  if (n_fxns == 0) next

  fxn_locations <- str_locate_all(text, ".{1,100}(?= <- function)")[[1]]

  # Convert output to data frame
  fxn_locations <- data.frame(
    file = f,
    function_name = fxn_names,
    search_start = fxn_locations[,"end"],
    search_end = c(fxn_locations[2:n_fxns,"start"], nchar(text)))

  output <- assign_row_join(output, fxn_locations)
}

output %<>%
  mutate(function_name = str_trim(function_name) %>%
           str_remove("`")) %>%
  filter(function_name %not_in% dont_include)

for(f in all_files) {

  # read R file as string
  text <- readChar("R/" %p% f, file.info("R/" %p% f)$size)

  lookup_vars <- output %>%
    filter(file == f)

  for (v in 1:nrow(lookup_vars)) {
    this_var = lookup_vars[v,"function_name"]
    this_text = str_sub(text, lookup_vars[v,"search_start"], lookup_vars[v,"search_end"])

    matching_vars <- output$function_name[str_detect(this_text, output$function_name)]

    if (length(matching_vars) == 0) next
    if (is.na(this_var)) next


    output_df <- data.frame(
      ancestor = this_var,
      descendents = matching_vars)

    final_output = assign_row_join(final_output, output_df)

  }

}

# Run it back to search for function text inside of other functions

# Create list of all data frames
all_folders <- list.files("data-raw")
all_folders <- all_folders[!str_detect(all_folders, "\\.R")]

for(f in all_folders) {
  R_file <- list.files("data-raw/" %p% f)
  R_file <- all_R_files[str_detect(all_R_files, "\\.R")]

  r_file_name <- paste0("data-raw/", f, "/", R_file)

  # read R file as string
  text <- readChar(r_file_name, file.info(r_file_name)$size)

  # Create list of function names and locations. If no functions, skip to next file.
  fxn_names <- str_extract_all(text, ".{1,100}(?= <- function)")[[1]] %>%
    as.character()

  n_fxns = length(fxn_names)

  if (n_fxns == 0) next

  fxn_locations <- str_locate_all(text, ".{1,100}(?= <- function)")[[1]]

  # Convert output to data frame
  fxn_locations <- data.frame(
    file = f,
    function_name = fxn_names,
    search_start = fxn_locations[,"end"],
    search_end = c(fxn_locations[2:n_fxns,"start"], nchar(text)))

  file_names <- data.frame(
    file = f,
    data_name =
  )

}

greaterlouisvilleproject/glptools documentation built on Feb. 9, 2025, 11:21 a.m.