YuuMJ1997/EukPhylo
1
0
Fork 0
mirror of http://43.156.76.180:8026/YuuMJ/EukPhylo.git synced 2025-12-27 10:50:25 +08:00
Code Issues Packages Projects Releases Wiki Activity

Update preguidance.py

Browse Source
This commit is contained in:
Auden Cote-L'Heureux 2023-11-24 12:35:21 -05:00 committed by GitHub
parent 8fa40adf9c
commit 8872d1724d
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 6 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
PTL2/Scripts/preguidance.py
View File

@ -15,6 +15,11 @@ def run(params):
except (FileNotFoundError, TypeError) as e: except (FileNotFoundError, TypeError) as e:
Logger.Error('Unable to read taxon list file. Please make sure that the path is correct and that the file is formatted correctly.\n\n' + str(e)) Logger.Error('Unable to read taxon list file. Please make sure that the path is correct and that the file is formatted correctly.\n\n' + str(e))
try:
blacklist_seqs = list(dict.fromkeys([line.strip() for line in open(params.blacklist)]))
except (FileNotFoundError, TypeError) as e:
print('\nUnable to read blacklist file. Please make sure that the path is correct and that the file is formatted correctly.\n\n' + str(e))
if not os.path.isdir(params.data): if not os.path.isdir(params.data):
Logger.Error(Logger.Error('Input amino-acid data files not found. Please make sure that the given path (--data) is correct.')) Logger.Error(Logger.Error('Input amino-acid data files not found. Please make sure that the given path (--data) is correct.'))
@ -31,7 +36,7 @@ def run(params):
with open(params.output + '/Output/Pre-Guidance/' + og + '_preguidance.faa', 'w') as preguidance_file: with open(params.output + '/Output/Pre-Guidance/' + og + '_preguidance.faa', 'w') as preguidance_file:
for taxon_file in aa_files: for taxon_file in aa_files:
recs = [] recs = []
for rec in sorted([rec for rec in SeqIO.parse(params.data + '/' + taxon_file, 'fasta') if rec.id[-10:] == og], key=lambda x: -len(x.seq)): for rec in sorted([rec for rec in SeqIO.parse(params.data + '/' + taxon_file, 'fasta') if rec.id[-10:] == og and rec.id not in blacklist_seqs], key=lambda x: -len(x.seq)):
if(rec.id == rec.description): if(rec.id == rec.description):
recs.append(rec) recs.append(rec)
else: else: