obitools/obitools3
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 66 Wiki Activity

obi import: 'taxid' columns are imported as 'TAXID' to fit view

Browse Source 
predicates, and fixed taxdump import and DMS closing
This commit is contained in:
Celine Mercier
2018-07-28 17:03:00 +02:00
parent c02c15b93f
commit 895d09b133
1 changed files with 33 additions and 27 deletions
Download Patch File Download Diff File Expand all files Collapse all files

54
python/obitools3/commands/import.pyx
View File

@ -70,6 +70,7 @@ def run(config):
cdef Column def_col
cdef Column seq_col
cdef Column qual_col
cdef Column taxid_col
cdef Column old_column
cdef bint rewrite
cdef dict dcols
@ -87,40 +88,37 @@ def run(config):
DMS.obi_atexit()
logger("info","obi import : imports file into an DMS")
logger("info", "obi import : imports file into a DMS")
input = open_uri(config['obi']['inputURI'])
if input is None:
raise Exception("Could not open input URI")
if not config['obi']['taxdump']: # TODO discuss
input = open_uri(config['obi']['inputURI'])
if input is None: # TODO check for bytes instead now?
raise Exception("Could not open input URI")
# TODO read taxdump with URI?
if 'taxdump' in config['obi']:
taxo = Taxonomy.open_taxdump(input[0], config['obi']['taxdump'])
# Read prefix (temporary fix)
uri = config['obi']['outputURI'].split('/')
idx = uri.index('taxonomy') + 1
taxo.write(uri[idx])
taxo.close()
input[0].close()
return
if input[2]==Nuc_Seq:
v = View_NUC_SEQS
# TODO uuuuh
if input[2]==Nuc_Seq:
v = View_NUC_SEQS
else:
v = View
else:
v = View
v = None
output = open_uri(config['obi']['outputURI'],
input=False,
newviewtype=v)
#quality=get_quality) # TODO
if output is None:
raise Exception("Could not create output view")
#print("input:", input)
#print("output:", output)
# Read taxdump
if config['obi']['taxdump']: # The input is a taxdump to import in a DMS
taxo = Taxonomy.open_taxdump(output[0], config['obi']['inputURI'])
taxo.write(output[1])
taxo.close()
output[0].close()
return
pb = ProgressBar(1000000, config, seconde=5) # TODO should be number of records in file
pb = ProgressBar(10000000, config, seconde=5) # TODO should be number of records in file
entries = input[1]
@ -168,6 +166,8 @@ def run(config):
if tag != b"ID" and tag != b"DEFINITION" and tag != b"NUC_SEQ" and tag != b"QUALITY" : # TODO hmmm...
value = entry[tag]
if tag == b"taxid":
tag = b"TAXID"
# Check NA value
if value == NA_value :
@ -262,6 +262,12 @@ def run(config):
print("\n")
print(view.__repr__())
input[0].close() # TODO ?
output[0].close()
try:
input[0].close()
except AttributeError:
pass
try:
output[0].close()
except AttributeError:
pass