library("DiffBind")
library("DESeq2")
library("GenomicRanges")
library("limma")


########## Functions used 
############ Genomic overlap
to_make_gr_object_chr_base_loc = function (chr_base_loc_vec,column_name) {
  
  chr_base_loc = unique(chr_base_loc_vec)
  chr_base_loc_gr = GRanges(chr_base_loc)
  chr_base_loc_gr$meta_col = chr_base_loc
  
  names(mcols(chr_base_loc_gr)) = column_name
  return (chr_base_loc_gr)  }



##### to_find overlap between gr 
to_overlap_grange_object = function (gr1, gr2, gr1_col, gr2_col) {
  
  query_gr = gr1
  subject_gr = gr2
  
  query_col = gr1_col
  subject_col = gr2_col
  
  overlap = findOverlaps(query_gr,subject_gr)
  
  query_row_ids = queryHits(overlap)
  subject_rowids = subjectHits(overlap)
  
  query_overlap_df = as.data.frame(elementMetadata(query_gr)[query_row_ids,query_col])
  subject_overlap_df= as.data.frame(elementMetadata(subject_gr)[subject_rowids, subject_col])
  
  ### to find the interesection of overlaps
  overlap_width = width(pintersect(query_gr[queryHits(overlap)],subject_gr[subjectHits(overlap )]))
  
  Overlap_df = cbind.data.frame(query_overlap_df,subject_overlap_df)
  Overlap_df["width"] = overlap_width
  
  return(Overlap_df) 
  
}

######## Outfn
grn_evidence_out_fn = function(grn_name,evidence_vector) {
  
  evidence_all = paste(evidence_vector,collapse = "_")
  grn_out_fn = paste(grn_name,evidence_all,sep = "_")
  return(grn_out_fn)
}

### Overlapping T-cell GRN with molecular disease evidence
grn_peak_overalp_with_cols = function(grn, grn_peak_col, disease_freq_df, evidence_col, evidence_name) {
  
  ####### Making col_name
  #evidence_id = paste(evidence_name,"id",sep = "_")
  ###### To make gr object for the overlap
  input_grn_peak_gr = to_make_gr_object_chr_base_loc(chr_base_loc_vec = grn[,grn_peak_col],column_name = "Location")
  input_evidence_gr = to_make_gr_object_chr_base_loc(chr_base_loc_vec = disease_freq_df[,evidence_col],column_name = "Location")
  
  
  grn_evidence_overlap_df = to_overlap_grange_object(gr1 = input_grn_peak_gr,gr2 = input_evidence_gr,
                                                     gr1_col = "Location",gr2_col = "Location")
  
  names(grn_evidence_overlap_df) = c("grn_peak_id","Evidence_id")
  
  ############### Sub-setting grn for the overlapping region 
  col_select_grn = c("peak","TF","ENSEMBL","SYMBOL")
  grn_sub = grn[grn[,grn_peak_col]%in%grn_evidence_overlap_df[,"grn_peak_id"],col_select_grn]
  
  ####### Merging grn_sub with overlap df to include evidence id
  grn_sub_merge = merge(grn_sub,grn_evidence_overlap_df,by.x = "peak",by.y = "grn_peak_id")
  
  ####### To add disease info of evidence (like diffpeak, coloc region) on merge df
  disease_col = c("Disease")
  
  grn_sub_merge[,disease_col] = disease_freq_df[match(grn_sub_merge[,"Evidence_id"],disease_freq_df[,evidence_col]),disease_col]
  
  ###### Adding the evidence col to grn_sub_merge
  grn_sub_merge$Evidence = evidence_name
  
  return(grn_sub_merge)
  
  
}

### to get key freq df
to_get_key_freq_category_df_mod = function (key_value_list,key_name,value_name) {
  
  
  #key_value_list = hQTL_disease_list
  #key_name = "hQTL_id"
  #value_name = "Disease"
  
  ###### Names of the columns in df
  value_freq_name = paste(value_name,"freq",sep = "_")
  value_category = paste(value_name,"category",sep = "_")
  
  ######### To make key- value freq df from key-value list
  key_vector = names(unlist(lapply(X = key_value_list,FUN = length)))
  value_freq_vector = unname(unlist(lapply(X = key_value_list,FUN = length)))
  
  key_value_freq_df = data.frame(key = key_vector,freq = value_freq_vector)
  names(key_value_freq_df) = c(key_name,value_freq_name)
  
  ###### To add cols
  key_value_freq_df$category = NA
  names(key_value_freq_df)[3] = value_category
  
  key_value_freq_df[,value_name] = NA
  
  ####### Adding the disease column for every key 
  
  for (i in seq(1:nrow(key_value_freq_df))) {
    
    key = as.character(key_value_freq_df[i,key_name])
    value = paste(key_value_list[[key]],collapse = "_")
    
    key_value_freq_df[i,value_name] = value
    
  }
  
  ###### To add the category as  "common" and "unique)
  key_value_freq_df[,value_freq_name] = as.numeric(key_value_freq_df[,value_freq_name])
  key_value_freq_df[,value_category] = ifelse(key_value_freq_df[,value_freq_name] == 1,"Unique","Common")
  
  return (key_value_freq_df)
}


to_make_master_freq_all_val_table = function(master_disease_grn_table, key_name, value_vector,merging = TRUE) {
  
  
  #disease_pattern = paste0(".*",disease_name,".*")
  #disease_specific_grn = master_disease_grn_table[grep(pattern = disease_pattern,x = master_disease_grn_table[,"Disease"]),]
  
  key_value_freq_df_list = list()
  #key_value_list_all = list()
  
  for( value_name in value_vector) {
    
    key_value_list = to_key_value_list_from_df(long_df = master_disease_grn_table,key_name = key_name,value_name = value_name)
    
    key_value_freq_df = to_get_key_freq_category_df_mod(key_value_list = key_value_list,key_name = key_name,value_name = value_name)
    key_value_freq_df_list[[value_name]] = key_value_freq_df
    #key_value_list_all[[value_name]] = key_value_list
    
  }
  
  ##### Merge freq dataframe 
  if (merging) {
    key_all_value_freq_df = Reduce(f = function(x,y) {z = merge (x,y,by.x = key_name,by.y = key_name); return (z)}, x = key_value_freq_df_list)
  }
  
  else{
    key_all_value_freq_df  = key_value_freq_df_list
  }
  
  return (key_all_value_freq_df)
}


generate_freq_df_grn_list_from_master_grn_corrected = function (input_master_grn){
  
  ##### master table for gene, TF, peak 
  TF_master_table = input_master_grn[grepl(pattern = "diffTF",x = input_master_grn$Evidence),]
  peak_master_table = input_master_grn[grepl(pattern = "diffpeak|gwas|coloc",x = input_master_grn$Evidence),]
  
  #### Process
  freq_list = list()
  value_vector = c("Disease","Evidence","Evidence_id")
  
  ### To make master targte genes df
  if(nrow(input_master_grn) > 0){
    target_genes_all_values_freq_df = to_make_master_freq_all_val_table(master_disease_grn_table = input_master_grn,key_name = "SYMBOL",value_vector = value_vector)
    freq_list[["target_genes_all_values_freq_df"]] = target_genes_all_values_freq_df}
  
  if(nrow(peak_master_table) > 0){
    peaks_all_values_freq_df = to_make_master_freq_all_val_table(master_disease_grn_table =  peak_master_table ,key_name = "peak",value_vector = value_vector)
    freq_list[["peaks_all_values_freq_df"]] = peaks_all_values_freq_df}
  
  if(nrow(TF_master_table) > 0){
    TF_all_values_freq_df = to_make_master_freq_all_val_table(master_disease_grn_table = TF_master_table,key_name = "TF",value_vector = value_vector)
    freq_list[["TF_all_values_freq_df"]] = TF_all_values_freq_df}
  
  return(freq_list)}


making_grn_from_master_grn = function (input_master_grn) {
  
  input_master_grn$TF_peak_gene = paste(input_master_grn$peak,
                                        input_master_grn$TF,
                                        input_master_grn$SYMBOL,sep = "_")
  
  grn_cols_select = c("peak","TF","ENSEMBL","SYMBOL")
  grn_unique = input_master_grn[!(duplicated(input_master_grn$TF_peak_gene)),grn_cols_select]
  
  return(grn_unique) }





to_find_same_direction_percentage = function (x){
  
  ### if dorection is postive or negative
  if(all(x ==1)|all(x == -1)) {
    
    same_direction_percent = 100
    
  }else{
    same_direction_percent = round((mean(x == 1))*100,2)}
  
  return(same_direction_percent)
  
}