R/wikidataESR.R

In cpesr/wikidataESR: From wikidata to R, for french ESR

# This file is part of wikidataESR.
#
# wikidataESR is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# wikidataESR is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# Author: Julien Gossa <gossa@unistra.fr>


#' wikidataESR: A package for retrieving and plotting ESR data from wikidata.
#'
#' Wikidata is a convenient tool to model universities information.
#' This package intends to help using data about the french universities,
#' namely Enseignement Supérieur et Recherche (ESR).
#' Its main purpose is to plot graphs about ESR universities relationship.
#'
#' @docType package
#' @name wikidataESR
#'
#' @references
#' - \url{https://github.com/cpesr/wikidataESR}
#' - \url{https://www.wikidata.org}
#'
#' @author Julien Gossa, \email{gossa@unistra.fr}
NULL


#' Get the ESR status of a wikidata item
#'
#' The status (_statut_) is the _legal status_ of an ESR institution.
#' It is based on the instance_of property (P31) and converted thanks to a local dataset.
#'
#' @param item A wikidata item, as returned by wikidataR.
#'
#' @return The status the item.
#'
#' @examples
#' wdesr_get_item_status(item)
#'
#' @references \url{https://github.com/cpesr/wikidataESR}
#' @seealso \code{\link{wdesr.statuts}}, \code{\link[WikidataR]{WikidataR}}
#' @author Julien Gossa, \email{gossa@unistra.fr}
#' @noRd
wdesr_get_item_status <- function(item) {
  item_id <- item[[1]]$id
  instance_of_id <- wd_get_item_statement_as_list(item,"P31")[[1]][[1]]
  if(is.null(instance_of_id)) {
    warning("The instance of wikidata item ", item_id, " is not set.\n",
            "  Default level (size of the node) is set to 7.\n",
            "  Please check the property P31 at https://www.wikidata.org/wiki/",item_id,"\n",
            "  using the guideline at https://github.com/cpesr")

    instance_of_id <- "NOID"
  }

  if (!instance_of_id %in% wdesr.cache$status$id) {
    it <- WikidataR::get_item(id = instance_of_id)
    label <- wd_get_item_alias(it)

    warning("The instance of wikidata item ", item_id, " is unknown by wikidataESR: ",label,".\n",
            "  Default level is set to 5.\n",
            "  Please check the property P31 at https://www.wikidata.org/wiki/",item_id,"\n",
            "  using the guideline at https://github.com/cpesr")

    wdesr.cache$status <- rbind(
      wdesr.cache$status,
      c(id         = instance_of_id,
        libellé    = label,
        recommandé = "",
        niveau     = 5,
        wikipedia  = "",
        note       = "statut inexistant dans la base wikidataESR"))
  }

  status <- subset(wdesr.cache$status, id == instance_of_id)
  
  if (status$recommandé == "non")
    warning("The instance of wikidata item ", item_id, " is not recommended: ",status$libellé,".\n",
            "  Reason is: ", ifelse(status$note != "",status$note, "Statut pas assez précis"),".\n",
            "  Please check https://www.wikidata.org/wiki/",item_id,"\n",
            "  using the guideline at https://github.com/cpesr")

  return(status)
}

#' Load the data of one university.
#'
#' @param wdid The wikidata id of the university.
#'
#' @return A dataframe with the data of the university.
#'
#' @examples wdesr_load_item("Q3551576")
#'
#' @references
#' - \url{https://github.com/cpesr/wikidataESR}
#' - \url{https://www.wikidata.org}
#' @author Julien Gossa, \email{gossa@unistra.fr}
#' @noRd
wdesr_load_item <- function(wdid) {

  item <- WikidataR::get_item(id = wdid)
  if (!is.null(item[[1]]$redirect)) {
    stop("Redirection ",wdid," -> ",item[[1]]$redirect)
  }
  status <- wdesr_get_item_status(item)
  
  return(
    data.frame(
      id          = wdid,
      label       = wd_get_item_label(item),
      alias       = wd_get_item_alias(item),
      statut      = status$libellé,
      niveau      = as.numeric(status$niveau),
      état        = ifelse(is.na(wd_get_item_statement_as_year(item,"P576")), "actif", "dissout"),

      fondation   = wd_get_item_statement_as_year(item,"P571"),
      dissolution = wd_get_item_statement_as_year(item,"P576"),

      associé      = wd_get_item_statement_as_list(item,"P527"),
      associé_de   = wd_get_item_statement_as_list(item,"P361"),

      composante    = wd_get_item_statement_as_list(item,"P355"),
      composante_de = wd_get_item_statement_as_list(item,"P749"),

      prédécesseur     = wd_get_item_statement_as_list(item,"P1365"),
      prédécesseur_pit = wd_get_item_statement_qualifier_as_list(item,"P1365","P585"),
      successeur       = wd_get_item_statement_as_list(item,"P1366"),
      successeur_pit   = wd_get_item_statement_qualifier_as_list(item,"P1366","P585"),

      séparé_de        = wd_get_item_statement_as_list(item,"P807"),
      séparé_de_pit    = wd_get_item_statement_qualifier_as_list(item,"P807","P585"),
      absorbé_par      = wd_get_item_statement_as_list(item,"P7888"),
      absorbé_par_pit  = wd_get_item_statement_qualifier_as_list(item,"P7888","P585"),
      
            
      membre_de = wd_get_item_statement_as_list(item,"P463"),
      affilié_à = wd_get_item_statement_as_list(item,"P1416")
    )
  )
}

#' Load the data of a set of universities.
#'
#' @param wdids A set of wikidata ids.
#'
#' @return A dataframe with the data of the universities.
#'
#' @examples wdesr_load_items(c("Q3551576","Q2013017"))
#'
#' @references
#' - \url{https://github.com/cpesr/wikidataESR}
#' - \url{https://www.wikidata.org}
#' @author Julien Gossa, \email{gossa@unistra.fr}
#' @noRd
wdesr_load_items <- function(wdids) {
  for(subids in wdids) {
    for(wdid in subids) {
      print(paste("Loading: ",wdid))
      r <- wdesr_load_item(wdid)
      wdesr.cache$items <- rbind(wdesr.cache$items,r)
    }
  }
}

#' Loader of universities data.
#'
#' Get the data of a set of universities.
#'
#' Data are cached: use \code{\link{wdesr_clear_cache}} to refresh data from wikidata.
#'
#' @param wdids A set of wikidata ids.
#'
#' @return A dataframe with the data of the universities.
#' @export
#'
#' @examples items <- wdesr_get_data(c("Q3551576","Q2013017"))
#'
#' @references
#' - \url{https://github.com/cpesr/wikidataESR}
#' - \url{https://www.wikidata.org}
#' @seealso \code{\link{wdesr_clear_cache}}
#' @author Julien Gossa, \email{gossa@unistra.fr}
wdesr_get_data <- function(wdids) {
  if (is.null(wdesr.cache$status)) {
    wdesr_clear_cache()
  }

  wdesr_load_items(wdids[! wdids %in% wdesr.cache$items$id])

  suppressMessages(
    data <- dplyr::left_join(data.frame(id=wdids),wdesr.cache$items)
  )
  return(data)
}


#' Get a graph of universities.
#'
#' From a root wikipedia id, the function follows a given set of properties,
#' building vertice and edges along the way.
#'
#' Data are cached: use \code{\link{wdesr_clear_cache}} to refresh data from wikidata.
#'
#' @param wdid The wikidata id of the root.
#' @param props The set of properties to follow.
#' @param depth The depth of the graph (more or less) (default to 3).
#' @param active_only TRUE to filter dissolved universities (default to FALSE).
#' @param stop_at A list of type of nodes that must not be visited furthermore (default to c("EPST","EPIC","EPCA","Académie","agence publique")).
#'
#' @return A list of edges and vertices.
#' @export
#'
#' @examples
#' g <- wdesr_get_graph("Q61716176",c('composante','associé'), 1)
#' g$edges
#' g$vertice
#'
#' @references
#' - \url{https://github.com/cpesr/wikidataESR}
#' - \url{https://www.wikidata.org}
#' @seealso \code{\link{wdesr_clear_cache}}
#' @author Julien Gossa, \email{gossa@unistra.fr}
wdesr_get_graph <- function(wdid, props, depth = 3, active_only = FALSE, 
                            stop_at = c("EPST","EPIC","EPCA","Académie","agence publique") ) {

  wgge <- new.env()
  wgge$edges <- data.frame(from=character(),to=character(),type=character(),stringsAsFactors = FALSE)
  wgge$vertices <- data.frame()

  wdid <- wd_check_redirection(wdid)
  wdesr_get_subgraph(wgge, wdid, props, depth, active_only, stop_at)

  wgge$vertices <- wgge$vertices %>%
    mutate_all(as.character) %>%
    mutate(
      niveau = as.numeric(niveau)
    ) 
  #wgge$vertices$niveau <- factor(wgge$vertices$niveau, levels = wdesr.niveaux$niveau)

  wgge$edges <- wgge$edges %>% mutate(across(where(is.factor), as.character))
  suppressMessages(
    dupes <- wgge$edges %>% janitor::get_dupes(from,to) %>%
      select(from,to,type) %>% mutate(warning = "duplicated")
    )
  if (nrow(dupes) > 0) {
    warning("Duplicated relations detected:\n", 
            paste0(capture.output(as.data.frame(dupes)), collapse = "\n"))
    #wgge$edges <- wgge$edges %>% group_by(from,to) %>% slice_head() 
  }
  #suppressMessages(
  #  wgge$edges <- left_join(wgge$edges, dupes)
  #)
  
  wgge$vertices <- wgge$vertices %>% unique() %>% mutate(across(where(is.factor), as.character))
  res <- list('edges'=wgge$edges, 'vertices'=wgge$vertices)
  
  return(res)
}

#' Get a sub graph of universities.
#' @return A list of edges and vertices.
#' @noRd
#'
#' @seealso \code{\link{wdesr_get_graph}}
#' @author Julien Gossa, \email{gossa@unistra.fr}
wdesr_get_subgraph <- function(wgge, wdid, props, depth = 3, active_only = FALSE, stop_at = c("EPST") ) {

  df.from <- wdesr_get_data(c(wdid))

  #df.from$depth <- depth
  wgge$vertices <- rbind(wgge$vertices, df.from)

  #print(wdid)
  #print(wgge$vertices$id)
  #print(wgge$vertices[,1:2])

  for(p in props) {
    if(is.null(unlist(df.from[,p])))
      next()

    ppit <- paste(p,'_pit',sep='')

    df.to <- wdesr_get_data(unlist(df.from[,p]))

    # Remove dissolved
    if (active_only) df.to <- filter(df.to, état == "actif")

    # Remove existing to -> from edges
    tmp <- subset(wgge$edges, to == wdid)$from
    df.to <- subset(df.to, !id %in% tmp)

    if (nrow(df.to) == 0) next()

    edges <- data.frame(
      from  = df.from$id,
      to    = df.to$id,
      type  = p,
      date  = ifelse(ppit %in% colnames(df.from),unlist(df.from[,ppit]),NA),
      depth = depth,
      distance = df.from$niveau - df.to$niveau
      )
    wgge$edges <- rbind(wgge$edges,edges)

    #df.to$depth <- depth - 1

    if(depth==1) {
      wgge$vertices <- rbind(wgge$vertices, subset(df.to, !id %in% wgge$vertices$id))
    } else {
      wgge$vertices <- rbind(wgge$vertices, subset(df.to, !id %in% wgge$vertices$id & statut %in% stop_at))

      for(id in subset(df.to, !statut %in% stop_at)$id) {
        if (!id %in% wgge$vertices$id)
          wdesr_get_subgraph(wgge, id,props,depth-1,active_only,stop_at)
      }
    }
  }
}



#' Label builder.
#'
#' Build the label of the nodes.
#'
#' @param node_label Either "alias", "alias_date", "long", or "long_date" (default to "alias").
#' @param alias The alias of the item.
#' @param label The label of the item.
#' @param fondation The foundation date of the item.
#' @param dissolution The dissolution date of the item.
#' @param wrap The number of characters to wrap by line (default to 20)
#'
#' @return a label for a node, as a string
#'
#' @examples node_label_aes("alias", alias, label, fondation, dissolution)
#' @references
#' - \url{https://github.com/cpesr/wikidataESR}
#' - \url{https://www.wikidata.org}
#' @seealso \code{\link{wdesr_clear_cache}}
#' @author Julien Gossa, \email{gossa@unistra.fr}
#' @noRd
wdesr_node_label_aes <- function(node_label = "alias", alias, label, fondation, dissolution, wrap = 20) {
  switch(node_label,
         alias = {
           stringr::str_wrap(alias,wrap)},
         alias_date = {
           paste0(stringr::str_wrap(alias,wrap),'\n',fondation,'-',dissolution) },
         long = {
           stringr::str_wrap(label,wrap)},
         long_date = {
           paste0(stringr::str_wrap(label,wrap),"\n",fondation,'-',dissolution) },
         ""
  )
}

#' Get geom_nodeX function.
#'
#' Get the suitable geom_nodeX function according the node_type
#'
#' @param node_type Either "text", "text_repel", "label", "label_repel", "node" or "none" (default to "text").
#'
#' @return The suitable geom_nodeX function according the node_type.
#'
#' @examples wdesr_node_geom("text_repel")
#'
#' @references
#' - \url{https://github.com/cpesr/wikidataESR}
#' - \url{https://www.wikidata.org}
#' @seealso \code{\link{wdesr_clear_cache}}
#' @author Julien Gossa, \email{gossa@unistra.fr}
#' @noRd
wdesr_node_geom <- function(node_type = "text") {
  switch(node_type,
         text = {geom_nodetext},
         text_repel = {geom_nodetext_repel},
         label = {geom_nodelabel},
         label_repel = {geom_nodelabel_repel},
         node = {geom_nodes},
         geom_blank
  )
}



distance2weight <- function(d) {
  return(10+d)
}

get_status_colors <- function(df.g) {
  unknown_statuts <- setdiff(df.g$vertices$statut, wdesr.statuts$libellé)
  unknown_statuts_colors <- setNames(
    rep("grey",length(unknown_statuts)),
    unknown_statuts)
  
  ws <- wdesr.statuts %>% filter(libellé %in% df.g$vertices$statut)
  statuts_colors <- c(
    setNames(ws$color, ws$libellé),
    unknown_statuts_colors)
}

#' Plot an ESR graph.
#'
#' A wrapper for ggplot2 to plot graph as returned by \code{\link{wdesr_get_graph}}.
#'
#' @param df.g A data representing a graph, as returned by wdesr_get_graph.
#' @param layout The layout to use in the graph computing
#' @param size A multiplier of all of the other sizes (default to 1.0)
#' @param node_sizes The size of the nodes, either a single value or a range c(min,max).
#' @param label_sizes The size of the nodes, either a single value or a range c(min,max).
#' @param node_label Define the label for the nodess. Either "alias", "alias_date", "long", or "long_date" (default to "alias").
#' @param node_type Define the type of drawing for the nodes. Either "text", "text_repel", "label", "label_repel", "node" or "none" (default to "text").
#' @param edge_label TRUE to plot dates on edges (default to "TRUE").
#' @param edge_arrow TRUE or arrow object to plot arrow on edges, FALSE otherwise (default to "FALSE").
#' @param arrow_gap A parameter that will shorten the network edges in order to avoid overplotting edge arrows and nodes see \code{\link[ggnetwork]{fortify.network}}.
#' @param size_guide TRUE to plot the guide for sizes (default to "FALSE").
#' @param legend_position The position of the legend, "none" to remove (default to "right").
#' @param margin_x The horizontal margins, useful for big nodes (defautl to 0.2).
#' @param margin_y The vertical margins, useful for long node texts (default to 0.03).
#' @param label_wrap The number of charaters to wrap labels (default to 20).
#'
#' @return A ggplot2.
#' @export
#'
#' @examples
#' df.aslace <- wdesr_get_graph("Q61716176",c('composante','associé'), 1)
#'
#' wdesr_ggplot_graph(df.alsace,
#'   node_size = c(10,30), label_sizes = c(3,5), arrow_gap = 0.0,
#'   node_label = "alias", node_type = "text",
#'   edge_label = FALSE)
#' @references
#' - \url{https://github.com/cpesr/wikidataESR}
#' - \url{https://www.wikidata.org}
#' @seealso \code{\link{wdesr_clear_cache}}
#' @author Julien Gossa, \email{gossa@unistra.fr}

wdesr_ggplot_graph <- function( df.g,
                                layout = igraph::with_kk,
                                size = 1.0,
                                node_sizes = c(10,30),
                                label_sizes = c(4,6),
                                node_label = "alias",
                                node_type = "text",
                                edge_label = TRUE,
                                edge_arrow = FALSE,
                                arrow_gap = 0.05,
                                size_guide = "none",
                                legend_position = "right",
                                margin_x = 0.2,
                                margin_y = 0.03,
                                label_wrap = 20) {

  if( nrow(df.g$vertices) == 0 | nrow(df.g$edges) == 0 ) {
    stop("Empty ESR graph: something went wrong with the graph production parameters")
  }
  
  node_sizes <- node_sizes * size
  label_sizes <- label_sizes * size
  margin_x <- margin_x * size
  margin_y <- margin_y * size
  
  if (edge_arrow == FALSE) arrow_gap = 0
  
  geom_node_fun <- wdesr_node_geom(node_type)
  if (edge_arrow == TRUE) edge_arrow <- arrow(length = unit(8, "pt"), type = "closed")
  if (edge_arrow == FALSE) edge_arrow <- NULL
   
  df.g$vertices <- arrange(df.g$vertices,id)
  
  statuts_colors <- get_status_colors(df.g)
  
  net <- igraph::graph_from_data_frame(df.g$edges, vertices = df.g$vertices)
  
  ggnet <- ggnetwork(net,
                     layout = layout(weights = distance2weight(df.g$edges$distance)),
                     arrow.gap = arrow_gap)
  
  g <- ggplot(ggnet, aes(x = x, y = y, xend = xend, yend = yend))
  g <- g + geom_edges(aes(linetype = type),#, size = weight),
                      arrow = edge_arrow,
                      alpha=1,
                      color="grey30")

  if(edge_label) g <- g + geom_edgetext(aes(label=date), size = min(label_sizes))

  g <- g + geom_nodes(shape = 21, aes(
    fill = statut,
    color = état,
    #color=statut,
    #alpha = (dissolution != "NA"),
    size = niveau #factor(niveau, levels=wdesr.niveaux$niveau)
  ))
  ggnet.nodes <- ggnet %>% select(name,niveau) %>% unique()
  #g <- g + geom_nodelabel(label=as.character(seq(1,7)), size=10) 
  g <- g + geom_node_fun(aes(
    label = wdesr_node_label_aes(node_label,alias,label,fondation,dissolution, label_wrap),
    fill = statut),
    size = scales::rescale(-ggnet.nodes$niveau,label_sizes,range(-wdesr.niveaux$niveau)),
    lineheight = .8
    )
  g <- g + coord_cartesian(clip="off") 
  g <- g + scale_alpha_manual(labels=c("dissous","actif"), values = (c(0.8,1)), name='statut')
  g <- g + scale_color_manual(values = c("white","black")) #grey30
  g <- g + scale_fill_manual(values = statuts_colors)
  g <- g + scale_size_continuous(breaks=wdesr.niveaux$niveau,
                                 limits=range(wdesr.niveaux$niveau),
                                 range=rev(node_sizes),
                                 labels=wdesr.niveaux$libellé,
                                 name="niveau",
                                 guide=size_guide)
  g <- g + xlim(-margin_x,1+margin_x) + ylim(-margin_y,1+margin_y)
  g <- g + theme_blank() + theme(legend.position=legend_position) +
    guides( fill = guide_legend(override.aes = list(size=5, color="white")) )
           

  return(g)
}

# ggplotly_wdesr_graph <- function(df.g) {
#   net <<- network(df.g$edges,
#                   vertex.attr=df.g$vertices %>% mutate_all(as.character) %>% arrange(id),
#                   matrix.type="edgelist", ignore.eval=FALSE)
#
#   ggnet <<- ggnetwork(net, layout = "kamadakawai",directed=TRUE)
#
#   ggplot(ggnet, aes(x = x, y = y, xend = xend, yend = yend)) +
#     geom_edges(aes(color = type, text=paste('date :',date)),
#                arrow = arrow(length = unit(10, "pt"), type = "closed")) +
#     #geom_edgetext(aes(color = type, label=date)) +
#     geom_nodetext(aes(
#       label = alias,
#       text = paste(label,status,paste('(',fondation,'-',dissolution,')',sep=''),paste("wikidata id:",id),sep='\n'),
#       color=status)) +
#     theme_blank()
#
#   ggplotly(tooltip="text")
# }


#' Wrapper to load and plot ESR graphs.
#'
#' Conveniently call \code{\link{wdesr_get_graph}} and then \code{\link{wdesr_ggplot_graph}}.
#'
#' @param wdid The wikidata id of the root.
#' @param props The properties to follows.
#' @param depth The depth of the following
#' @param plot_type Either "ggplot" or "plotly" (default to ggplot).
#' @param active_only TRUE to filter dissolved universities (default to FALSE).
#' @param ... Additionnal parameters for the plot; see \code{\link{wdesr_ggplot_graph}} for details.
#' @return A ggplot or a plotly.
#' @export
#'
#' @examples
#' wdesr_load_and_plot("Q61716176",c('composante','associé'), 1,
#'   node_size = c(10,30), label_sizes = c(3,5), arrow_gap = 0.0,
#'   node_label = "alias", node_type = "text",
#'   edge_label = FALSE)
#' @references
#' - \url{https://github.com/cpesr/wikidataESR}
#' - \url{https://www.wikidata.org}
#' @seealso \code{\link{wdesr_clear_cache}}
#' @author Julien Gossa, \email{gossa@unistra.fr}

wdesr_load_and_plot <- function( wdid,
                                 props          = c('composante','associé'),
                                 depth          = 3,
                                 active_only    = FALSE,
                                 plot_type      = 'ggplot',
                                 ...) {

  df.g <- wdesr_get_graph(wdid,props,depth,active_only)

  if(plot_type == 'plotly') {
    wdesr_ggplotly_graph(df.g)
  } else {
    wdesr_ggplot_graph(df.g,...)
  }
}


#' Get warnings for the vetices
#'
#' @param df.g A data representing a graph, as returned by wdesr_get_graph.
#'
#' @return A dataframe with vetices warnings
#' @export
#'
#' @examples
wdesr_log_warnings_vertices <- function(df.g) {
  itemslevels <- unique(df.g$vertices$id)
  suppressMessages(
    vw <- left_join(df.g$vertices, wdesr.statuts %>% select(statut = libellé, recommandé, note)) %>%
      mutate(id = factor(id,levels=itemslevels), message=as.character(NA)) %>%
      bind_rows(
        filter(.,recommandé == "non") %>% mutate(message = note),
        filter(.,nchar(alias)>=20) %>% mutate(message = "Alias manquant ou long"),
        filter(.,is.na(fondation)) %>% mutate(message = "Date de fondation manquante")
      ) %>%
      filter(! is.na(message)) %>%
      arrange(id) %>%
      select(entité = id, alias, statut, message)
  )
  
  return(vw)
}

#' Get warnings for the edges
#'
#' @param df.g A data representing a graph, as returned by wdesr_get_graph.
#'
#' @return A dataframe with edges warnings
#' @export
#'
#' @examples
wdesr_log_warnings_edges <- function(df.g) {
  itemslevels <- unique(df.g$vertices$id)
  suppressMessages(
    ew <- df.g$edges %>%
      mutate(from = factor(from,levels=itemslevels), to = factor(to, levels = itemslevels)) %>%
      bind_rows(
        janitor::get_dupes(., from,to) %>% mutate(message = "Relation multiple"),
        filter(.,is.na(date), ! type %in% c('composante','composante_de')) %>% mutate(message = "Date(s) manquante(s)") 
      ) %>%
      select(depuis = from, vers = to, type, message) %>% 
      filter(! is.na(message))
  ) 
  
  return(ew)
}


#' Output warnings in kable format
#'
#' @param df.g A data representing a graph, as returned by wdesr_get_graph.
#' @param kableFormat The format used for kableExtra::kable
#'
#' @return A string describing the warnings
#' @export
#'
#' @examples wdesr_log_warnings(df, "warnings.md")
#' @references
#' - \url{https://github.com/cpesr/wikidataESR}
#' - \url{https://www.wikidata.org}
#' @seealso \code{\link{wdesr_clear_cache}}
#' @author Julien Gossa, \email{gossa@unistra.fr}
wdesr_log_warnings <- function(df.g, kableformat = "pipe") {

  warning <- ""
  vw <- wdesr_log_warnings_vertices(df.g)
  if (nrow(vw) > 0) {
    warning <- paste0(warning,"Problèmes détectés dans les entités :\n\n",
                      paste(vw %>% 
                              mutate(entité = wd_id2url(entité)) %>%
                              kableExtra::kable(format=kableformat), collapse='\n'),
                      "\n\n")
  }

  ew <- wdesr_log_warnings_edges(df.g)
  if (nrow(ew) > 0) {
    warning <- paste0(warning,"Problèmes détectés dans les relations :\n\n",
                      paste(ew %>% 
                              mutate(depuis = wd_id2url(depuis), vers = wd_id2url(vers)) %>%
                              kableExtra::kable(format=kableformat), collapse='\n'), 
                      "\n\nNB : les dates manquantes pour les relations de composante ne sont pas remontées.")
  }
  
  return(warning)
}