LSHTM_analysis/mcsm/mcsm_wrapper.py
Tanushree Tunstall b28d866237 handle not ready (refresh) url
Please enter the commit message for your changes. Lines starting
2020-04-21 17:12:18 +01:00

149 lines
5.8 KiB
Python
Executable file

#!/usr/bin/env python3
# mCSM Wrapper
import os,sys
import subprocess
import argparse
import pandas as pd
from mcsm import *
#%% command line args
arg_parser = argparse.ArgumentParser()
arg_parser.add_argument('-d', '--drug',required=True, help='drug name')
arg_parser.add_argument('-g', '--gene', help='gene name (case sensitive)', required=True) # case sensitive
arg_parser.add_argument('-s', '--stage', help='mCSM Pipeline Stage', default = 'get', choices=['submit', 'get', 'format'])
arg_parser.add_argument('-H', '--host', help='mCSM Server', default = 'http://biosig.unimelb.edu.au')
arg_parser.add_argument('-U', '--url', help='mCSM Server URL', default = 'http://biosig.unimelb.edu.au/mcsm_lig/prediction')
args = arg_parser.parse_args()
gene = args.gene
drug = args.drug
stage = args.stage
# Statics. Replace with argparse() later
# Actual Globals :-)
host = args.host
prediction_url = args.url
#host = "http://biosig.unimelb.edu.au"
#prediction_url = f"{host}/mcsm_lig/prediction"
#drug = 'isoniazid'
#gene = 'KatG'
# submit_mcsm globals
homedir = os.path.expanduser('~')
os.chdir(homedir + '/git/LSHTM_analysis/mcsm')
gene_match = gene + '_p.'
datadir = homedir + '/git/Data'
indir = datadir + '/' + drug + '/' + 'input'
outdir = datadir + '/' + drug + '/' + 'output'
in_filename_pdb = gene.lower() + '_complex.pdb'
infile_pdb = indir + '/' + in_filename_pdb
#in_filename_snps = gene.lower() + '_mcsm_snps_test.csv' #(outfile2, from data_extraction.py)
in_filename_snps = gene.lower() + '_mcsm_snps.csv' #(outfile2, from data_extraction.py)
infile_snps = outdir + '/' + in_filename_snps
result_urls_filename = gene.lower() + '_result_urls.txt'
result_urls = outdir + '/' + result_urls_filename
# mcsm_results globals
print('infile:', result_urls)
mcsm_output_filename = gene.lower() + '_mcsm_output.csv'
mcsm_output = outdir + '/' + mcsm_output_filename
# format_results globals
print('infile:', mcsm_output)
out_filename_format = gene.lower() + '_mcsm_processed.csv'
outfile_format = outdir + '/' + out_filename_format
#%%=====================================================================
def submit_mcsm():
my_chain = 'A'
# my_ligand_id = 'DCS' # FIXME
my_ligand_id = 'RMP' # FIXME
my_affinity = 10
print('Result urls and error file (if any) will be written in: ', outdir)
# call function to format data to remove duplicate snps before submitting job
mcsm_muts = format_data(infile_snps)
mut_count = 1 # HURR DURR COUNT STARTEDS AT ONE1`!1
infile_snps_len = os.popen('wc -l < %s' % infile_snps).read() # quicker than using Python :-)
print('Total SNPs for', gene, ':', infile_snps_len)
for mcsm_mut in mcsm_muts:
print('Processing mutation: %s of %s' % (mut_count, infile_snps_len), mcsm_mut)
print('Parameters for mcsm_lig:', in_filename_pdb, mcsm_mut, my_chain, my_ligand_id, my_affinity, prediction_url, outdir, gene)
# function call: to request mcsm prediction
# which writes file containing url for valid submissions and invalid muts to respective files
holding_page = request_calculation(infile_pdb, mcsm_mut, my_chain, my_ligand_id, my_affinity, prediction_url, outdir, gene, host)
time.sleep(1)
mut_count += 1
# result_url = write_result_url(holding_page, result_urls, host)
print('Request submitted'
, '\nCAUTION: Processing will take at least ten'
, 'minutes, but will be longer for more mutations.')
#%%=====================================================================
def get_results():
output_df = pd.DataFrame()
url_counter = 1 # HURR DURR COUNT STARTEDS AT ONE1`!1
success_counter = 1
infile_len = os.popen('wc -l < %s' % result_urls).read() # quicker than using Python :-) #FIXME filenme (infile_urls)
print('Total URLs:', infile_len)
with open(result_urls, 'r') as urlfile:
for line in urlfile:
url_line = line.strip()
# call functions
results_interim = scrape_results(url_line)
if results_interim is not None:
print('Processing URL: %s of %s' % (url_counter, infile_len))
result_dict = build_result_dict(results_interim)
df = pd.DataFrame(result_dict, index=[url_counter])
output_df = output_df.append(df)
success_counter += 1
url_counter += 1
print('Total URLs: %s Successful: %s Failed: %s' % (url_counter-1, success_counter-1, (url_counter - success_counter)))
output_df.to_csv(mcsm_output, index = None, header = True)
#%%=====================================================================
def format_results():
print('Input file:', mcsm_output
, '\n============================================================='
, '\nOutput file:', outfile_format
, '\n=============================================================')
# call function
mcsm_df_formatted = format_mcsm_output(mcsm_output)
# writing file
print('Writing formatted df to csv')
mcsm_df_formatted.to_csv(outfile_format, index = False)
print('Finished writing file:'
, '\nFile:', outfile_format
, '\nExpected no. of rows:', len(mcsm_df_formatted)
, '\nExpected no. of cols:', len(mcsm_df_formatted)
, '\n=============================================================')
#%%=====================================================================
def main():
if stage == 'submit':
print('mCSM stage: submit mutations for mcsm analysis')
submit_mcsm()
elif stage == 'get':
print('mCSM stage: get results')
get_results()
elif stage == 'format':
print('mCSM stage: format results')
format_results()
else:
print('ERROR: invalid stage')
main()