#' @title  build a convenient genomic features object.
#' @description build a convenient \code{\link{genome_features-class}} object needed for find overlaps beetween
#' selected features and genomic coordinates in a single step.
#' @importFrom  AnnotationDbi select keys as.data.frame metadata
#' @importFrom  GenomicFeatures promoters transcripts fiveUTRsByTranscript threeUTRsByTranscript exonsBy intronsByTranscript cdsBy
#' @importFrom  GenomicRanges flank
#' @importFrom rtracklayer import
#' @importFrom  BiocGenerics as.data.frame
#' @importFrom  methods new
#' @import data.table
#' @param txdb a txdb database built with \pkg{GenomicFeatures} (see example).
#' @param features a \code{\link[base]{vector}} of features \code{\link[base]{character}} with by
#' default all the availables features ("promoter", "UTR5", "UTR3", "exon", "intron", "cds", and "downstream").
#' @param parameters a \code{\link[base]{list}} of parameters to use in order to define promoter and
#' downstream ranges distance to TSS (Transcript Start Site) or TES (Transcript End Site).
#' \itemize{
#'  \item{promoter_ranges: a \code{\link[base]{list}} with upstream and dowstream  elements distance (in base) to TSS}.
#'  \item{downstream_range: the dowstream distance (in base) to the TES}.
#' }
#' @details This method build a convenient \code{\link{genomic_features-class}} object, in order to be used for finding overlaps beetween
#' all selected features and genomic coordinates with the \code{\link{genome_features_overlaps}} method.
#' @return a \code{\link{genome_features-class}} object.
#' @include genome_features.R
#' @examples
#'
#' ###################
#  # import the chicken genome GFF from NCBI ftp (galGal5 in this case)
#' utils::download.file("ftp://ftp.ncbi.nlm.nih.gov/genomes/Gallus_gallus/GFF/ref_Gallus_gallus-5.0_top_level.gff3.gz",
#' quiet=T,destfile = "galGal5_gff.gz")
#'
#' ##################
#' #  uncompress the file
#' R.utils::gunzip("galGal5_gff.gz")
#'
#' ###################
#' # build the database
#' toplevel=GenomicFeatures::makeTxDbFromGFF(
#'  file="galGal5_gff",
#'  format="gff",
#'  organism="Gallus gallus",
#'  taxonomyId=9031,
#'  dbxrefTag="ID"
#' )
#'
#' ###################
#' # build genome_features object from txdb
#' features<-GenomeFeatures::build_genome_features(
#'  toplevel,
#'  features=base::c("promoter","UTR5","UTR3","exon","intron","cds","downstream"),
#'  parameters = base::list(
#'    promoter_ranges=base::list(
#'      upstream=3000,
#'      downstream=500
#'    ),
#'    downstream_range=1000
#'  )
#' )
#' @export
setGeneric(name="build_genome_features",def=function(txdb,features=base::c("promoter","UTR5","UTR3","exon","intron","cds","downstream"),
parameters=base::list(promoter_ranges=base::list(upstream=3000,downstream=500),downstream_range=1000)){standardGeneric("build_genome_features")})

setMethod("build_genome_features",definition=function(txdb,features,parameters) {

  ###################
  # check txdb
  ###################
  if(base::class(txdb)!="TxDb") base::stop("txdb must be a TxDb class object")

  ###################
  # txdb ids crossref
  ###################

  ###################
  # build ids cross reference
  crossref<-data.table::data.table(
    AnnotationDbi::select(
      txdb,
      columns=c("TXID","TXNAME"),
      keytype="GENEID",
      keys=AnnotationDbi::keys(txdb)
    )
  )

  ###################
  # convert TXID to character
  crossref[,TXID:=base::as.character(TXID)]

  ###################
  # promoters
  ###################

  promoters=NULL;promoters_DT=NULL

  if("promoter"%in%features){

    ###################
    # get promoters
    promoters <- GenomicFeatures::promoters(
      txdb,
      upstream=parameters$promoter_ranges$upstream,
      downstream=parameters$promoter_ranges$downstream
    )

    ###################
    # convert to table
    promoters_DT<-data.table::data.table(AnnotationDbi::as.data.frame(promoters,row.names = NULL))[,.(tx_id,tx_name)]

    ###################
    # convert to character
    promoters_DT[,tx_id:=base::as.character(tx_id)]

    ###################
    # merge with crossref
    promoters_DT<-merge(promoters_DT,crossref,by.x="tx_id",by.y="TXID",sort=F,all.x=T)

    ###################
    # rename
    base::names(promoters_DT)[1:2]<-base::c("group","feature")

    ###################
    # format feature
    promoters_DT[,  tx_name:=feature]

    ###################
    # format feature
    promoters_DT[,`:=`(
      feature=base::paste(GENEID,feature,"promoter",sep=":"),
      GENEID=NULL,
      TXNAME=NULL
    )]
  }

  ###################
  # downstream
  ###################

  downstream=NULL;downstream_DT=NULL

  if("downstream"%in%features){

    ###################
    # get downstream
    downstream<- GenomicFeatures::transcripts(txdb)
    downstream<-GenomicRanges::flank(downstream,parameters$downstream_range,start=F)

    ###################
    # convert to table
    downstream_DT<-data.table::data.table(AnnotationDbi::as.data.frame(downstream,,row.names = NULL))[,.(tx_id,tx_name)]

    ###################
    # convert to character
    downstream_DT[,tx_id:=base::as.character(tx_id)]

    ###################
    # merge with crossref
    downstream_DT<-merge(downstream_DT,crossref,by.x="tx_id",by.y="TXID",sort=F,all.x=T)

    ###################
    # rename
    base::names(downstream_DT)[1:2]<-base::c("group","feature")

    ###################
    # format feature
    downstream_DT[,tx_name:=feature]

    ###################
    # format feature
    downstream_DT[,`:=`(
      feature=base::paste(GENEID,feature,"downstream",sep=":"),
      GENEID=NULL,
      TXNAME=NULL
    )]
  }

  ###################
  # 5' UTR
  ###################

  UTR5=NULL;UTR5_DT=NULL

  if("UTR5"%in%features){

    ###################
    # get 5' UTR in GRangesList
    UTR5<-GenomicFeatures::fiveUTRsByTranscript(txdb,use.names=F)

    ###################
    # convert to table
    UTR5_DT<-data.table::data.table(BiocGenerics::as.data.frame(UTR5,,row.names = NULL))[,.(group_name,exon_rank)]

    ###################
    # convert to GRanges
    UTR5<-base::unlist(UTR5)

    ###################
    # merge with crossref
    UTR5_DT<-merge(UTR5_DT,crossref,by.x="group_name",by.y="TXID",sort=F,all.x=T)

    ###################
    # format feature
    UTR5_DT[,`:=`(
      feature=base::paste(GENEID,":",TXNAME,":5'UTR(exon",exon_rank,")",sep=""),
      exon_rank=NULL,
      GENEID=NULL,
      tx_name=TXNAME,
      TXNAME=NULL,
      group_name=NULL
    )]
  }

  ###################
  # 3' UTR
  ###################

  UTR3=NULL;UTR3_DT=NULL

  if("UTR3"%in%features){

    ###################
    # get 3' UTR
    UTR3<-GenomicFeatures::threeUTRsByTranscript(txdb,use.names=F)

    ###################
    # convert to table
    UTR3_DT<-data.table::data.table(AnnotationDbi::as.data.frame(UTR3,,row.names = NULL))[,.(group_name,exon_rank)]

    ###################
    # convert to GRanges
    UTR3<-base::unlist(UTR3)

    ###################
    # merge with crossref
    UTR3_DT<-merge(UTR3_DT,crossref,by.x="group_name",by.y="TXID",sort=F,all.x=T)

    ###################
    # format feature
    UTR3_DT[,`:=`(
      feature=base::paste(GENEID,":",TXNAME,":3'UTR(exon",exon_rank,")",sep=""),
      exon_rank=NULL,
      GENEID=NULL,
      tx_name=TXNAME,
      TXNAME=NULL,
      group_name=NULL
    )]
  }

  ###################
  # Exons
  ###################

  exons=NULL;exons_DT=NULL

  if("exon"%in%features){

    ###################
    # get exons
    exons<-GenomicFeatures::exonsBy(txdb,use.names=F)

    ###################
    # convert to table
    exons_DT<-data.table::data.table(AnnotationDbi::as.data.frame(exons,row.names = NULL))[,.(group_name,exon_rank)]

    ###################
    # convert to GRanges
    exons<-base::unlist(exons)

    ###################
    # count exons by transcripts
    exons_nb<-exons_DT[!base::is.na(group_name),.N,by=group_name]

    ###################
    # merge count to exons_DT
    exons_DT<-merge(exons_DT,exons_nb,by="group_name",all.x=T,sort=F)

    ###################
    # merge with crossref
    exons_DT<-merge(exons_DT,crossref,by.x="group_name",by.y="TXID",sort=F,all.x=T)

    ###################
    # format feature
    exons_DT[,`:=`(
      feature=base::paste(GENEID,":",TXNAME,":exon",exon_rank,"/",N,sep=""),
      exon_rank=NULL,
      GENEID=NULL,
      tx_name=TXNAME,
      TXNAME=NULL,
      N=NULL,
      group_name=NULL
    )]
  }

  ###################
  # Introns
  ###################

  introns=NULL;introns_DT=NULL

  if("intron"%in%features){

    ###################
    # get introns
    introns<-GenomicFeatures::intronsByTranscript(txdb,use.names=F)

    ###################
    # convert to table
    introns_DT<-data.table::data.table(AnnotationDbi::as.data.frame(introns,row.names = NULL))[,.(group_name,strand)]

    ###################
    # convert to GRanges
    introns<-base::unlist(introns)

    ###################
    # count introns by transcripts
    introns_nb<-introns_DT[!base::is.na(group_name),.N,by=group_name]

    ###################
    # merge count to introns_DT
    introns_DT<-merge(introns_DT,introns_nb,by="group_name",all.x=T,sort=F)

    ###################
    # add intron_rank
    introns_DT[,intron_rank:=1:base::length(N),by=group_name]

    ###################
    # reverse intron_rank for reverse strand
    introns_DT[strand=="-",intron_rank:=base::rev(intron_rank),by=group_name]

    ###################
    # merge with crossref
    introns_DT<-merge(introns_DT,crossref,by.x="group_name",by.y="TXID",sort=F,all.x=T)

    ###################
    # format feature
    introns_DT[,`:=`(
      feature=base::paste(GENEID,":",TXNAME,":intron",intron_rank,"/",N,sep=""),
      intron_rank=NULL,
      strand=NULL,
      GENEID=NULL,
      tx_name=TXNAME,
      TXNAME=NULL,
      N=NULL,
      group_name=NULL
    )]
  }

  ###################
  # CDS
  ###################

  cds=NULL;cds_DT=NULL

  if("cds"%in%features){

    ###################
    # get exons
    cds<-GenomicFeatures::cdsBy(txdb,use.names=F)

    ###################
    # convert to table
    cds_DT<-data.table::data.table(AnnotationDbi::as.data.frame(cds,row.names = NULL))[,.(group_name,exon_rank)]

    ###################
    # convert to GRanges
    cds<-base::unlist(cds)

    ###################
    # count exons by transcripts
    cds_nb<-cds_DT[!base::is.na(group_name),.N,by=group_name]

    ###################
    # merge count to exons_DT
    cds_DT<-merge(cds_DT,cds_nb,by="group_name",all.x=T,sort=F)

    ###################
    # merge with crossref
    cds_DT<-merge(cds_DT,crossref,by.x="group_name",by.y="TXID",sort=F,all.x=T)

    ###################
    # format feature
    cds_DT[,`:=`(
      feature=base::paste(GENEID,":",TXNAME,":exon",exon_rank,"/",N,sep=""),
      exon_rank=NULL,
      GENEID=NULL,
      tx_name=TXNAME,
      TXNAME=NULL,
      N=NULL,
      group_name=NULL
    )]
  }

  ###################
  # load annotation complement
  ###################

    ###################
    # check file name
    file=AnnotationDbi::metadata(txdb)$value[3]

    if(!base::is.na(file)){

      ###################
      # download
      if(base::length(grep("ftp.+\\.gz$",file))>0){

        ###################
        #  output name
        output=base::sub("^.+/","",file)

        ###################
        # import the GFF
        utils::download.file(file,quiet=T,destfile=output)

        ##################
        #  uncompress
        R.utils::gunzip(output)

        ##################
        # file to download
        file=output
      }

      ###################
      # read the gene file as GRanges
      gene_table<-rtracklayer::import(file)

      ###################
      # convert to data.table
      gene_table<-data.table::data.table(
        BiocGenerics::as.data.frame(gene_table,,row.names = NULL)
      )[ data.table::like(type,"RNA|transcript"),]


      if(base::length(base::grep("gtf$",file))>0){

        ###################
        # select columns from gtf
        gene_table<-gene_table[,
          .(seqnames,start,end,strand,type,gene_name,transcript_id,gene_id)
        ]

      }else{

        ###################
        # select columns from gff
        gene_table<-gene_table[,
          .(seqnames,start,end,strand,type,Dbxref,gene,transcript_id)
        ]

        ###################
        # extract gene_id
        gene_table[,
          gene_id:=base::sub("^.+:","",Dbxref[[1]][1]),
          by=1:base::nrow(gene_table)
        ][,Dbxref:=NULL]

        ###################
        # reneme gene to gene_name
        base::names(gene_table)[6]<-"gene_name"

      }

  }else{

    ###################
    # else NA
    gene_table=data.table::data.table(gene_table=NA)
  }

  ###################
  # create genome_features
  ###################

  ###################
  # create genome_features
  methods::new("genome_features",
    metadata=base::list(data.table::data.table(AnnotationDbi::metadata(txdb))),
    promoter=base::list(GRanges=promoters,DT=promoters_DT),
    UTR5=base::list(GRanges=UTR5,DT=UTR5_DT),
    UTR3=base::list(GRanges=UTR3,DT=UTR3_DT),
    exon=base::list(GRanges=exons,DT=exons_DT),
    intron=base::list(GRanges=introns,DT=introns_DT),
    cds=base::list(GRanges=cds,DT=cds_DT),
    downstream=base::list(GRanges=downstream,DT=downstream_DT),
    gene_table=gene_table
  )
})