Merge and process tables (special)

clusters.tsv

data.tsv

transform.py

import sys
import os
from statistics import mean, median

if len(sys.argv) < 4:
    sys.exit("run: python transform.py <data TSV> <clusters TSV> <func:max,min,mean,median>")

data_file = sys.argv[1]
clusters_file = sys.argv[2]
func_name = sys.argv[3]

func = None

if func_name == "max":
    func = max
elif func_name == "min":
    func = min
elif func_name in ["avg", "mean"]:
    func = mean
elif func_name == "median":
    func = median

# This save all the information
all_clusters = list()

number_of_runs = 0
with open(clusters_file) as clustersFileReader:
    for line in clustersFileReader:
        number_of_runs += 1
        tmp_column_to_cluster = dict()
        line = line.strip().split('\t')
        for colID, clusterID in enumerate(line, start = 1):
            tmp_column_to_cluster[colID] = clusterID
        
        collapsed_tmp_column_to_cluster = dict()
        for colID, clusterID in tmp_column_to_cluster.items():
            if clusterID in collapsed_tmp_column_to_cluster:
                collapsed_tmp_column_to_cluster[clusterID].append(colID)
            else:
                collapsed_tmp_column_to_cluster[clusterID] = [colID]

        all_clusters.append(collapsed_tmp_column_to_cluster)


run_id_to_file = dict()
for i in range(1, number_of_runs + 1, 1):
    run_id_to_file[i] = open(f"run_{i}_{func_name}.tsv", 'w')

with open(data_file) as dataFileReader:
    next(dataFileReader) # Skip header (delete if no header)
    for line in dataFileReader:
        line = list(map(float, line.strip().split('\t')))
        current_cluster_to_columns = []

        current_run_id = 0
        for cluster_to_cols in all_clusters:
            current_run_id += 1
            cluster_to_final_value = dict()
            for clusterID, cols in cluster_to_cols.items():
                tmp_selected_rows = list()
                for col in cols:
                    tmp_selected_rows.append(line[col-1])
                final_value = func(tmp_selected_rows)
                cluster_to_final_value[clusterID] = final_value

            transformed_final_values = list()
            for clusterID, finalValue in cluster_to_final_value.items():
                transformed_final_values.append(finalValue)

            transformed_row = '\t'.join(list(map(str, transformed_final_values))) + '\n'
            run_id_to_file[current_run_id].write(transformed_row)



for i, fileWriter in run_id_to_file.items():
    fileWriter.close()