Spaces:
Sleeping
Sleeping
File size: 18,902 Bytes
8145f0e 5f02390 8145f0e 5f02390 8145f0e 5f02390 8145f0e 5f02390 8145f0e 3647a13 5f02390 8145f0e 5f02390 8145f0e 5f02390 8145f0e 5f02390 8145f0e 5f02390 8145f0e 5f02390 8145f0e 5f02390 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 |
library(shiny)
library(shinyalert)
library(shinythemes)
library(shinydisconnect)
library(shinycssloaders)
library(shinyWidgets)
library(spotifyr)
library(SpotifyNetwork)
library(spsComps)
library(searcher)
library(reactable)
library(visNetwork)
library(plotly)
library(bslib)
library(dplyr)
library(igraph)
# set global options
options(
shiny.browser = T,
spinner.color = "#16F529",
spinner.color.background = "#FFFFFF",
spinner.size = 2
)
# create theme
my_theme <- bs_theme(
bg = "#fdfefe",
fg = "black",
primary = "red",
base_font = font_google("Michroma"),
"font-size-base" = "0.75rem",
version = 5,
"navbar-bg" = "#16F529"
)
# define UI for application that gets Spotify network data
ui <- navbarPage(
title = strong("Spotify Data Importer"), id = "navbar",
# useShinyalert(),
windowTitle = "Spotify Data Importer",
# add telemetry javascript elements
header = use_telemetry(),
footer = h5(strong(tagList(h5(span("Spotify Data Importer for",style="color:green"), a("NodeXL", href = "https://www.nodexl.com/"))))),
theme = my_theme, collapsible = T, setBackgroundImage(src = "music.jpg"),
# add marquee visual element
tags$html(HTML("<marquee direction = right scrollamount = '12' style = 'color:green; font-size:17px;'><strong>Spotify Data Importer For <span style='color:red'>NodeXL</span></strong></marquee>"),
),
# Add social media icons to navbar
tags$style(".navbar-nav.socialmedia { display: flex; justify-content: right; flex-direction:row; padding: 5px; font-size: 20px; }"),
tags$div(class = "navbar-nav socialmedia",
tags$a(href = "https://www.github.com/Ifeanyi55", target = "_blank", icon("github", lib = "font-awesome"))),
tabPanel(
id = "tabOne", value = "oneTab", title = strong("Home"), icon = icon("home"),
sidebarLayout(
sidebarPanel(
id = "side", width = 3, h4(strong("Credentials")), hr(),
textInput("client_id", strong("Enter Client ID")),
textInput("client_secret", strong("Enter Client SECRET")), br(),
actionButton("validate", strong("Authenticate"), icon = icon("caret-right")), br(),hr(),
textOutput("valout")
),
mainPanel(
id = "main", width = 8,
span(style = "color:blue;",h5(strong(p(id = "dateclock")))),br(),
h2(strong("Welcome to the NodeXL Spotify Data Importer!")),
p(h5(strong("Before you begin scraping data, you will need to complete the steps below:"))), hr(),
p(h5(strong(tagList("STEP 1: Go to", a("https://developer.spotify.com/dashboard/", href = "https://developer.spotify.com/dashboard/"), "and login with your credentials")))), br(),
p(h5(strong("STEP 2: Create an app, and give it a name and a description"))), br(),
p(h5(strong("STEP 3: Get the generated client ID and client Secret, and return here"))), br(),
p(h5(strong("STEP 4: Enter the client ID and client Secret in the 'Credentials' box"))), br(),
p(h5(strong("STEP 5: Click the 'Authenticate' button"))), br(),
p(h5(strong("STEP 6: Wait for the system to authenticate your credentials and print out a reference ID"))), br(),
p(h5(strong("Great! You can now proceed to scraping network data via the Spotify API."))),
actionButton("nextTab", strong("Proceed")), hr()
)
)
),
tabPanel(
id = "tabA", value = "Atab", title = strong("Network Data"), icon = icon("table"),
# load and run the CSS script
includeCSS("style.css"),
# load and run the javascript script
includeScript("JSCode.js"),
sidebarLayout(
sidebarPanel(
width = 2, id = "sidebar",
actionButton("info", strong("Info"), icon = icon("info")), br(), br(), br(), tags$a(img(src = "spotify.png"), href = "https://open.spotify.com/"), br(), hr(),
textInput("id", strong("Enter Artist's Spotify Id")),
actionButton("run", strong("Related Data"), icon = icon("caret-right")),br(),hr(),br(),
h5(strong("Collaborators Data")),
actionButton("fetch", strong("Collab Data"), icon = icon("caret-right")),hr()
),
mainPanel(
# go to top button
spsGoTop(id = "up",right = "3%",bottom = "10%",icon = icon("arrow-up",color = "green")),
textInput("search", span(strong("Search Box"),style = "color:white;"), placeholder = "Search Google", width = "150px"),actionButton("search_bttn", strong("Search")),hr(),
fluidRow(column(12, h3(strong(span(style = "color:white;text-align:center;",h4("Related Artists Network Data")))))),
fluidRow(
column(12, withSpinner(reactableOutput("network_data",width = 1000,height = 400), type = 1)),
),
fluidRow(
column(6, downloadButton("down_csv", strong("Download CSV"), icon = icon("download"))),
column(6, downloadButton("down_graphml", strong("Download GraphML"), icon = icon("download")))
),br(),br(),br(),
fluidRow(column(12, h3(strong(span(style = "color:white;text-align:center;",h4("Artists Collaboration Network Data")))))),
fluidRow(
column(12,withSpinner(reactableOutput("collabs_data", width = 1000, height = 400),type = 1))
),
downloadButton("down_flat",strong("Download CSV"),icon = icon("download")),
br(), hr(), uiOutput("out")
)
)
),
tabPanel(
id = "tabB", value = "Btab", title = strong("80s Hits"), icon = icon("music"),
sidebarLayout(sidebarPanel = "", mainPanel(tags$iframe(
style = "border-radius:12px",
src = "https://open.spotify.com/embed/playlist/37i9dQZF1DXb57FjYWz00c?utm_source=generator",
width = "1350px",
height = "550px",
frameBorder = "0",
allowfullscreen = "",
allow = "autoplay; clipboard-write; encrypted-media; fullscreen; picture-in-picture",
loading = "lazy"
)))
),
tabPanel(
id = "tabD", strong("NodeXL YouTube"), icon = icon("youtube"),
sidebarLayout(sidebarPanel(id = ""), mainPanel(
tags$iframe(
width = "620",
height = "350",
src = "https://www.youtube.com/embed/xKhYGRpbwOc",
title = "YouTube video player",
frameborder = "0",
allow = "accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture",
allowfullscreen = T
), hr(),
tags$iframe(
width = "620",
height = "350",
src = "https://www.youtube.com/embed/Gs4NPuKIXdo",
title = "YouTube video player",
frameborder = "0",
allow = "accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture",
allowfullscreen = T
), hr(),
tags$iframe(
width = "620",
height = "350",
src = "https://www.youtube.com/embed/J1W5uqAyHTg",
title = "YouTube video player",
frameborder = "0",
allow = "accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture",
allowfullscreen = T
), hr(),
tags$iframe(
width = "620",
height = "350",
src = "https://www.youtube.com/embed/zEgrruOITHw",
title = "YouTube video player",
frameborder = "0",
allow = "accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture",
allowfullscreen = T
), hr(),
tags$iframe(
width = "620",
height = "350",
src = "https://www.youtube.com/embed/pwsImFyc0lE",
title = "YouTube video player",
frameborder = "0",
allow = "accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture",
allowfullscreen = T
), hr(),
tags$iframe(
width = "620",
height = "350",
src = "https://www.youtube.com/embed/mjAq8eA7uOM",
title = "YouTube video player",
frameborder = "0",
allow = "accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture",
allowfullscreen = T
)
))
)
)
# define server logic required to get Spotify's network data
server <- function(input, output, session) {
# set up Spotify API credentials environment
authentication <- function(id, secret) {
client_ID <- id
client_secret <- secret
# authenticate the spotify client side
Sys.setenv(SPOTIFY_CLIENT_ID = client_ID)
Sys.setenv(SPOTIFY_CLIENT_SECRET = client_secret)
access_token <- get_spotify_access_token()
}
client_id <- reactive(as.character(input$client_id))
client_secret <- reactive(as.character(input$client_secret))
authenticate <- reactive(authentication(
client_id(),
client_secret()
))
validation <- eventReactive(input$validate, {
authenticate()
})
output$valout <- renderText({
validation()
})
# update nav bar page using the value of the tabPanel
# (i.e. value = "Atab")
observeEvent(input$nextTab, {
updateNavbarPage(session, inputId = "navbar", selected = "Atab")
})
# activate the search box
search_input <- reactive(input$search)
searchGoogle <- reactive(search_google(search_input(), rlang = F))
search_result <- eventReactive(input$search_bttn, {
searchGoogle()
})
output$out <- renderUI({
search_result()
})
# render spotify
# renderUI({
# tags$iframe(src = "https://open.spotify.com/")
# })
# turn text input into a reactive object
id_input <- reactive({
as.character(input$id)
})
# wrap SpotifyNetwork functions in reactive wrappers
# related_network <- reactive({
# related_artists_network(id_input())
# })
# artist_plot <- reactive({
# artists_popularity(id_input())
# })
nodes_table <- reactive({
related_artists_nodes(id_input())
})
edges_table <- reactive({
related_artists_edges(id_input())
})
# create event reactive element for each output
# network_react <- eventReactive(input$run, {
# related_network()
# })
# artist_react <- eventReactive(input$run, {
# artist_plot()
# })
nodes_react <- eventReactive(input$run, {
nodes_table()
})
edges_react <- eventReactive(input$run, {
edges_table()
})
# function to wrangle nodes & edges data into a NodeXL flat file
create_flatTabler <- function(dfN,dfE){
# scrape data from Vertex1
# as.vector(data,mode) converts the returned list into a vector
popularity <- apply(dfE,1,function(df) subset(dfN,dfN$name == df[["Vertex1"]])[[3]])
popularity <- as.vector(popularity,"numeric")
followers <- apply(dfE,1,function(df) subset(dfN,dfN$name == df[["Vertex1"]])[[4]])
followers <- as.vector(followers,"numeric")
profile <- apply(dfE,1,function(df) subset(dfN,dfN$name == df[["Vertex1"]])[[5]])
profile <- as.vector(profile,"character")
images <- apply(dfE,1,function(df) subset(dfN,dfN$name == df[["Vertex1"]])[[6]])
images <- as.vector(images,"character")
genre <- apply(dfE,1,function(df) subset(dfN,dfN$name == df[["Vertex1"]])[[7]])
genre <- as.vector(genre,"character")
# scrape data from Vertex2
# as.vector(data,mode) converts the returned list into a vector
popularityB <- apply(dfE,1,function(df) subset(dfN,dfN$name == df[["Vertex2"]])[[3]])
popularityB <- as.vector(popularityB,"numeric")
followersB <- apply(dfE,1,function(df) subset(dfN,dfN$name == df[["Vertex2"]])[[4]])
followersB <- as.vector(followersB,"numeric")
profileB <- apply(dfE,1,function(df) subset(dfN,dfN$name == df[["Vertex2"]])[[5]])
profileB <- as.vector(profileB,"character")
imagesB <- apply(dfE,1,function(df) subset(dfN,dfN$name == df[["Vertex2"]])[[6]])
imagesB <- as.vector(imagesB,"character")
genreB <- apply(dfE,1,function(df) subset(dfN,dfN$name == df[["Vertex2"]])[[7]])
genreB <- as.vector(genreB,"character")
# assign scraped data for Vertex1 to new columns
dfE$`Vertex1 popularity` <- popularity
dfE$`Vertex1 followers` <- followers
dfE$`Vertex1 profile` <- profile
dfE$`Vertex1 images` <- images
dfE$`Vertex1 genre` <- genre
# assign scraped data for Vertex2 to new columns
dfE$`Vertex2 popularity` <- popularityB
dfE$`Vertex2 followers` <- followersB
dfE$`Vertex2 profile` <- profileB
dfE$`Vertex2 images` <- imagesB
dfE$`Vertex2 genre` <- genreB
dfE$`Edge Weight` <- round(dfE$`Vertex1 popularity`/dfE$`Vertex2 popularity`,2)
dfE <- dfE |> relocate(`Edge Weight`,.after = Vertex2)
return(dfE)
}
# parse edges_react to function
flat_file <- reactive({create_flatTabler(nodes_react(),edges_react())})
# add edge metadata
# flat_file()["Edge Weight"] <- reactive({round(flat_file()$`Vertex1 popularity`/flat_file()$`Vertex2 popularity`,2)})
#
# flat_file() <- flat_file() |>
# reactive({relocate(`Edge Weight`,.after = Vertex2)})
# create flat file event reactive object
flat_react <- eventReactive(input$run,{
flat_file()
})
# # render outputs
# output$network <- renderVisNetwork({
# network_react()
# })
#
# output$plot <- renderPlotly({
# artist_react()
# })
# function to download nodes data file
output$down_csv <- downloadHandler(
filename = function() {
paste("Related_artists", ".csv", sep = "")
},
content = function(file) {
write.csv(flat_react(), file)
}
)
output$network_data <- renderReactable({
tryCatch(
{
reactable(flat_react(),
theme = reactableTheme(
highlightColor = "#00e600",
borderColor = "#00e600",
borderWidth = 3
),
outlined = T,
bordered = T,
filterable = T,
striped = T,
compact = T,
highlight = T,
defaultColDef = colDef(
align = "center",
headerStyle = list(background = "#00e600")
),
paginationType = "simple"
)
},
error = function(e){
message("There was an error!")
print(e)
}
)
})
# function to generate GraphML file
create_graphml <- function(nodes,edges){
# create new graph object
graph <- graph_from_data_frame(edges)
# add attributes to graph nodes
{
V(graph)$Name <- nodes$name
V(graph)$Popularity <- nodes$popularity
V(graph)$Followers <- nodes$followers
V(graph)$Profile <- nodes$profile
V(graph)$Images <- nodes$images
V(graph)$Genre <- nodes$genre
}
return(graph)
}
# make function reactive
graphml_react <- reactive({
create_graphml(nodes_react(),
edges_react())})
# make function event reactive so that it is triggered
# by run action button
graphmlReact <- eventReactive(input$run,{
graphml_react()
})
# write GraphML download function
output$down_graphml <- downloadHandler(
filename = function(){
paste("Related_artists",".graphml",sep = "")
},
content = function(file){
write_graph(graphmlReact(),file,format = "graphml")
}
)
# import get_artists_collaborators() function
source("Collaborators.R")
# wrap in reactive wrappers
artists_collaborations <- reactive({get_artists_collaborators(id_input())})
# make event reactive
collabs_react <- eventReactive(input$fetch,{artists_collaborations()})
output$collabs_data <- renderReactable({
tryCatch(
{
reactable(collabs_react(),
theme = reactableTheme(
highlightColor = "#3498DA",
borderColor = "#3498DA",
borderWidth = 3
),
outlined = T,
bordered = T,
filterable = T,
striped = T,
compact = T,
highlight = T,
defaultColDef = colDef(
align = "center",
headerStyle = list(background = "#3498DA")
),
paginationType = "simple"
)
},
error = function(e){
message("There was an error!")
print(e)
}
)
})
# activate download button
output$down_flat <- downloadHandler(
filename = function(){
paste("CollabData",".csv",sep = "")
},
content = function(file){
write.csv(collabs_react(),file)
}
)
# function to download edges data file
# output$down_edges <- downloadHandler(
# filename = function() {
# paste("Edges", ".csv", sep = "")
# },
# content = function(file) {
# write.csv(edges_react(), file)
# }
# )
# output$edges <- renderReactable({
# reactable(edges_react(),
# theme = reactableTheme(
# highlightColor = "#00FFAB",
# borderColor = "#00FFAB"
# ),
# outlined = T,
# bordered = T,
# filterable = T,
# striped = T,
# compact = T,
# highlight = T,
# defaultColDef = colDef(
# align = "center",
# headerStyle = list(background = "#00FFAB")
# ),
# paginationType = "simple"
# )
# })
observeEvent(input$info, {
shinyalert(
title = "About Software", closeOnEsc = T, confirmButtonCol = "#006400",
imageUrl = "spotify.png",
closeOnClickOutside = T,
confirmButtonText = "Got It", showConfirmButton = T,
animation = "pop", timer = 20000,
text = "The Spotify Data Importer allows a user to
query the Spotify API and get network data
of artists that are related to a particular artist. You can also get data of artists that have collaborated together.
It typically takes between 16 and 20
seconds to get a response from the Spotify server for related
artists network data and more than that for artists collaboration
network data."
)
})
observeEvent(input$fetch,{
shinyalert(
closeOnEsc = T,
confirmButtonText = "Got It!",
confirmButtonCol = "#3498DA",
closeOnClickOutside = T,
showConfirmButton = T,
animation = "slide-from-top",
cancelButtonText = T,
timer = 10000,
text = "This may take a while to run, sorry!"
)
})
}
# Run the application
shinyApp(ui = ui, server = server)
|