From 5862116debc830a1e724effe7ea43272055f63d2 Mon Sep 17 00:00:00 2001
From: remy <remy.dernat@umontpellier.fr>
Date: Mon, 11 Nov 2019 21:38:49 +0100
Subject: [PATCH] bugs fixed
---
orthofinder/orthofinder.py | 2 +-
orthofinder/scripts/blast_file_processor.py | 29 +++++++++++----------
2 files changed, 16 insertions(+), 15 deletions(-)
diff --git a/orthofinder/orthofinder.py b/orthofinder/orthofinder.py
index 91cd61f..2694d6f 100755
--- a/orthofinder/orthofinder.py
+++ b/orthofinder/orthofinder.py
@@ -1402,7 +1402,7 @@ def DoNormalize(options, speciesInfoObj, seqsInfo):
for iSpecies in range(seqsInfo.nSpecies):
cmd_queue.put((seqsInfo, blastDir_list, Lengths, iSpecies))
# Normalization is done here
- runningProcesses = [mp.Process(target=WaterfallMethod.Worker_ProcessBlastHits, args=(cmd_queue, options.qDoubleBlast, options.qNormalize=True)) for i_ in range(options.nProcessAlg)]
+ runningProcesses = [mp.Process(target=WaterfallMethod.Worker_ProcessBlastHits, args=(cmd_queue, options.qDoubleBlast, options.qNormalize)) for i_ in range(options.nProcessAlg)]
for proc in runningProcesses:
proc.start()
parallel_task_manager.ManageQueue(runningProcesses, cmd_queue)
diff --git a/orthofinder/scripts/blast_file_processor.py b/orthofinder/scripts/blast_file_processor.py
index d928bd8..af0b6aa 100644
--- a/orthofinder/scripts/blast_file_processor.py
+++ b/orthofinder/scripts/blast_file_processor.py
@@ -103,9 +103,8 @@ def GetBLAST6Scores(seqsInfo, blastDir_list, iSpecies, jSpecies, qExcludeSelfHit
raise
return B
-def WriteNormalizedBlastFile(blastDir_list, B, iSpecies, jSpecies, qExcludeSelfHits = True, sep = "_", qDoubleBlast=True):
+def WriteNormalizedBlastFile(blastDir_list, B, iSpecies, jSpecies, sep = "_", qDoubleBlast=True):
qSameSpecies = iSpecies==jSpecies
- qCheckForSelfHits = qExcludeSelfHits and qSameSpecies
if not qDoubleBlast:
qRev = (iSpecies > jSpecies)
else:
@@ -120,15 +119,15 @@ def WriteNormalizedBlastFile(blastDir_list, B, iSpecies, jSpecies, qExcludeSelfH
iH = 1
iSpeciesOpen = iSpecies
jSpeciesOpen = jSpecies
-
tempfile = NamedTemporaryFile(mode='w', delete=False)
#xSeqID, ySeqID, NormVal = ParseMatrix(B)
+ row = ""
for d in blastDir_list:
fn = d + "Blast%d_%d.txt" % (iSpeciesOpen, jSpeciesOpen)
fnNorm = d + "Blast%d_%d.norm.blastout" % (iSpeciesOpen, jSpeciesOpen)
if os.path.exists(fn) or os.path.exists(fn + ".gz"): break
try:
- with (gzip.open(fn + ".gz", file_reader_mode) if os.path.exists(fn + ".gz") else open(fn, file_reader_mode)) as blastfile:
+ with (gzip.open(fn + ".gz", file_read_mode) if os.path.exists(fn + ".gz") else open(fn, file_read_mode)) as blastfile:
blastreader = csv.reader(blastfile, delimiter='\t')
blastwriter = csv.writer(tempfile, delimiter='\t')
for row in blastreader:
@@ -139,14 +138,16 @@ def WriteNormalizedBlastFile(blastDir_list, B, iSpecies, jSpecies, qExcludeSelfH
except (IndexError, ValueError):
sys.stderr.write("\nERROR: Query or hit sequence ID in BLAST results file was missing or incorrectly formatted.\n")
raise
- try:
- score = float(row[11])
- except (IndexError, ValueError):
- sys.stderr.write("\nERROR: 12th field in BLAST results file line should be the bit-score for the hit\n")
- raise
+ #try:
+ # score = float(row[11])
+ #except (IndexError, ValueError):
+ # sys.stderr.write("\nERROR: 12th field in BLAST results file line should be the bit-score for the hit\n")
+ # raise
newscore = B[sequence1ID, sequence2ID]
- newrow = row
- newrow[11] = newscore
- writer.writerow(row)
-
- shutil.move(tempfile.name, fnNorm)
\ No newline at end of file
+ row.append(newscore)
+ blastwriter.writerow(row)
+ move(tempfile.name, fnNorm)
+ except Exception:
+ sys.stderr.write("Malformatted line in %sBlast%d_%d.txt\nOffending line was:\n" % (d, iSpecies, jSpecies))
+ sys.stderr.write("\t".join(map(str, row)) + "\n")
+ raise
--
GitLab