# HG changeset patch # User p.lucas # Date 1644423643 0 # Node ID b5b8f3cb240c66aa3afe9382cbc31fc3af87ae17 # Parent 7b55dec91dd1ab44330e31b19f530defc59d12c5 MAJ python script by FT diff -r 7b55dec91dd1 -r b5b8f3cb240c MEGABLAST_TAB_get_taxid_acc.py --- a/MEGABLAST_TAB_get_taxid_acc.py Fri Feb 04 14:11:46 2022 +0000 +++ b/MEGABLAST_TAB_get_taxid_acc.py Wed Feb 09 16:20:43 2022 +0000 @@ -10,21 +10,14 @@ # NOTE: to update krona tax in conda env, run: # ktUpdateTaxonomy.sh # ktUpdateTaxonomy.sh --accessions (this one NOT PROVIDED IN DOCUMENTATION) -import argparse, os, sys, csv, re, warnings +import argparse, os, sys, warnings # NEEDS to use krona conda environnement if access ktGetTaxIDFromAcc from os import path -# from optparse import OptionParser -from datetime import datetime #, timezone -import pytz -from natsort import natsorted # to be able to report line number in error messages import inspect frame = inspect.currentframe() -# start_time = datetime.now(timezone.utc) -start_time = datetime.now(pytz.utc) - # debug b_test_creates_taxid_acc_f_from_megablast_res = False # ok 2022 01 21 @@ -106,8 +99,9 @@ krona_taxid_acc_f = '' def creates_taxid_acc_f_from_megablast_res(megablast_f, tax_acc_out_f): acc_col_nb_in_megablast_res = str(2) - krona_taxdb_f = os.path.expanduser("/db/krona/") # krona['taxdb'] # "/nfs/data/db/tax_krona/" - if not os.path.isfile(krona_taxdb_f + "all.accession2taxid.sorted"): + # krona_taxdb_f = os.path.expanduser('~/miniconda3/envs/krona/opt/krona/taxonomy/') # krona['taxdb'] # "/nfs/data/db/tax_krona/" + krona_taxdb_f = os.path.expanduser('/db/krona/') # krona['taxdb'] # "/nfs/data/db/tax_krona/" + if not os.path.isfile(krona_taxdb_f + 'all.accession2taxid.sorted'): sys.exit(prog_tag + "[Error] missing "+krona_taxdb_f+" file, please run 'ktUpdateTaxonomy.sh --accessions' in your krona conda environment (and 'ktUpdateTaxonomy.sh' before if you have not done)") # conda: "../envs/krona.yaml" @@ -146,9 +140,6 @@ # creates taxid acc file from megablast result creates_taxid_acc_f_from_megablast_res(megablast_f, tax_acc_out_f) - stop_time = datetime.now(pytz.utc) - duration = stop_time - start_time - print("duration:",duration) #### MAIN END if __name__ == "__main__": __main__()