def insert_compounds(
filename,
lib_name,
namekey,
idkey,
lib_ver=0,
):
""" Read SDF file, fetch each compound and process it.
........All compounds will be inserted into a new library,
........If you are updating an existing library,
........please use update_compounds()"""
library = get_or_create_library(lib_name, lib_ver)
f = open(filename)
sdffile = u''
count = 0
for line in f.xreadlines():
line = unicode(line, 'latin1')
sdffile += line
if line.startswith('$$$$'):
moldata = parse_annotation(sdffile, namekey, idkey)
insert_single_compound(moldata, sdffile, library, namekey,
idkey)
sdffile = u''
count += 1
root.warning(' --> count:%s' % count)
f.close()
def gen_joelib_property(sdf):
"""run and parse the property output """
# save the input in FS
t = tempfile.NamedTemporaryFile(suffix='.sdf')
t.write(codecs.encode(sdf, 'utf-8'))
t.flush()
# prepare the output file
(f, out) = tempfile.mkstemp(suffix='.sdf')
os.close(f)
# convert
cmd = \
"""JAVA_HOME=/opt/jre/ JOELIB2=/opt/JOELib2-alpha-20070303/ /opt/JOELib2-alpha-20070303/moleculeConversion.sh +d +h -iSDF -osdf "%s" "%s" > /dev/null""" \
% (t.name, out)
root.warning(' --> running:%s' % cmd)
if os.system(cmd) != 0:
os.unlink(out)
raise 'cannot run JOELib'
# read and parse
f = file(out)
tags = get_sdf_tags(codecs.decode(f.read(), 'utf-8'))
f.close()
# clean
os.unlink(out)
return tags
def parse_annotation(sdf, namekey, idkey):
""" parse annotation from SDF file """
# parse the sdf tags
moldata = get_sdf_tags(sdf)
moldata[idkey] = moldata[idkey].replace(' ', '-')
# --- inchi
inchiconv.SetInAndOutFormats('sdf', 'Inchi')
mol = openbabel.OBMol()
if '(noMol)' not in moldata[idkey]:
res = inchiconv.ReadString(mol, codecs.encode(sdf, 'utf-8'))
if mol.Empty():
root.warning(' --> ERROR on sdf')
raise Exception
# standard data generated
# --- inchi/formula/weight
moldata['inchi'] = inchiconv.WriteString(mol).strip()
moldata['formula'] = mol.GetFormula()
mol.AddHydrogens()
moldata['weight'] = str(mol.GetMolWt())
# if the name is not in sdf:
if namekey == '__NA__':
moldata[namekey] = ''
# smiles
inchiconv.SetInAndOutFormats('sdf', 'smi')
mol = openbabel.OBMol()
if '(noMol)' not in moldata[idkey]:
res = inchiconv.ReadString(mol, codecs.encode(sdf, 'utf-8'))
moldata['smiles'] = inchiconv.WriteString(mol).strip()
else:
moldata['smiles'] = ''
# check if necessary keys are present: the name and key
if not moldata.has_key(idkey):
print 'compound data error: the idkey is not found'
raise Exception
if not moldata.has_key(namekey):
# print "warning: no name key"
moldata[namekey] = ''
# do any necessary fixes for tags
return moldata
def do_process(compounds, filep, tags):
for cmp in compounds:
sdf = SDFFile.objects.get(compound=cmp).sdffile
print ' running compound:%s' % cmp.id
try:
propertydata = gen_joelib_property(sdf)
_insert_property(cmp, propertydata, tags)
except:
root.warning(' --> cannot run property on %s'
% cmp.id)
raise
def insert_single_compound(
moldata,
sdf,
library,
namekey,
idkey,
username='',
):
""" insert single compound into database """
cid = moldata[idkey]
name = moldata[namekey]
if '\n' in name:
name = name.split('\n')[0]
# compound
c = Compound(
cid=cid,
name=name,
formula=moldata['formula'],
weight=moldata['weight'],
inchi=moldata['inchi'],
smiles=moldata['smiles'],
username=username,
)
# sdf_file=s)
c.save()
c.library.add(library)
c_id = c.id
root.warning(' -->new compound inserted: c_id=%s, cid=%s' % (c_id,
cid))
# sdf file
s = SDFFile(sdffile=sdf, compound=c)
s.save()
sdfid = s.id
# annotation
a_ids = []
for (name, value) in moldata.items():
if name in (namekey, idkey):
continue
a = Annotation(name=name, value=value, compound=c)
a.save()
aid = a.id
a_ids.append(aid)
return False
def insert_plate_from_db(
library_name,
library_ver,
format,
plate_key,
plate_base,
well_key,
row_key,
col_key,
):
"""if no file about plate mapping info,
........use plate info from sdf in db
........'format': 24/48/96/384
........'library': library name/id
........'plate_key': required
........'plate_base': '__NA__'/required, e.g. 'LAT', to be stripped from plate
........'well_key': optional, if no, must be '__NA__'
........'row_key' 'col_key': if no well_key, these 2 keys are required,
........ if has well_key, will be ignored
........e.g.:
........insert_plate_from_db('Latca', 96, 'PLATE', 'LAT', '__NA__','ROW', 'COL')"""
library = get_library(library_name, library_ver)
compounds = Compound.objects.filter(library=library)
format = int(format)
count = 0
for cmp in compounds:
plate = Annotation.objects.get(compound=cmp,
name=plate_key).value
if plate_base is not '__NA__':
try:
plate = int(plate.lstrip(plate_base))
except:
root.warning(' --> plate:%s, base:%s' % (plate,
plate_base))
if well_key is not '__NA__':
well = Annotation.objects.get(compound=cmp,
name=well_key).value
else:
row = Annotation.objects.get(compound=cmp,
name=row_key).value
col = Annotation.objects.get(compound=cmp,
name=col_key).value
if len(col) == 1:
col = '0%s' % col
well = '%s%s' % (row, col)
try:
p = Plate(format=format, plate=int(plate), well=well,
compound=cmp)
p.save()
count += 1
root.warning(' --> count:%s, compound:%s' % (count, cmp))
except:
raise
root.warning(' --> count:%s, compound:%s' % (count, cmp))
root.warning(' --> total count:%s' % count)
return
def parse_annotation(sdf, namekeys):
""" parse annotation from SDF file """
# parse the sdf tags
moldata = get_sdf_tags(sdf)
# --- inchi
inchiconv.SetInAndOutFormats('sdf', 'Inchi')
mol = openbabel.OBMol()
res = inchiconv.ReadString(mol, sdf)
if mol.Empty():
root.warning(' --> ERROR on sdf')
raise Exception
# standard data generated
# --- inchi/formula/weight
moldata['inchi'] = inchiconv.WriteString(mol).strip()
moldata['formula'] = mol.GetFormula()
moldata['id'] = mol.GetTitle()
if moldata['id'] == '':
moldata['id'] = 'unspecified_' \
+ ''.join(random.sample(string.digits, 6))
mol.AddHydrogens()
moldata['weight'] = str(mol.GetMolWt())
# Retrieve display name from SDF, or empty string if absent:
for n in namekeys:
if n in moldata:
moldata['name'] = moldata[n]
break
if moldata.get('name') is None:
moldata['name'] = ''
# smiles
inchiconv.SetInAndOutFormats('sdf', 'smi')
mol = openbabel.OBMol()
#res = inchiconv.ReadString(mol, codecs.encode(sdf, 'utf-8'))
res = inchiconv.ReadString(mol, sdf)
if mol.Empty():
root.warning(' --> ERROR on sdf')
raise Exception
moldata['smiles'] = inchiconv.WriteString(mol).strip()
return moldata
def fast_insert_plate_from_file(
library_name,
library_ver,
platefile,
format,
):
"""if we have a file containing plate mapping info,
........use plate info from the file.
........'library': library name/library id
........'platefile' format:
............1st column: compound id
............2nd column: plate
............3rd column: well
........'format': 24/48/96/384 """
format = int(format)
library = get_library(library_name, library_ver)
fp = file(platefile)
count = 0
for i in fp.xreadlines():
cid = ''
plate = ''
well = ''
(cid, plate, well) = i.strip().split(',')[0:3]
cid = str(cid)
if cid == 'NA':
continue
plate = int(plate)
well = str(well)
# sys.stdout.write('\r --> cid:%s' % (cid, ))
compound = Compound.objects.get(library=library, cid=cid)
try:
p = Plate(format=format, plate=int(plate), well=well,
compound=compound)
p.save()
count += 1
except:
print ' --> count:%s, cid:%s' % (count, cid)
raise
root.warning(' --> count:%s' % (count, ))
fp.close()
return
def parse_annotation(sdf, namekey):
""" parse annotation from SDF file """
# parse the sdf tags
moldata = get_sdf_tags(sdf)
# --- inchi
inchiconv.SetInAndOutFormats('sdf', 'Inchi')
mol = openbabel.OBMol()
res = inchiconv.ReadString(mol, sdf)
if mol.Empty():
root.warning(' --> ERROR on sdf')
raise Exception
# standard data generated
# --- inchi/formula/weight
moldata['inchi'] = inchiconv.WriteString(mol).strip()
moldata['formula'] = mol.GetFormula()
moldata['id'] = mol.GetTitle()
if moldata['id'] == '':
moldata['id'] = 'unspecified_' \
+ ''.join(random.sample(string.digits, 6))
mol.AddHydrogens()
moldata['weight'] = str(mol.GetMolWt())
# if the name is not in sdf:
if namekey not in moldata:
moldata[namekey] = ''
# smiles
inchiconv.SetInAndOutFormats('sdf', 'smi')
mol = openbabel.OBMol()
#res = inchiconv.ReadString(mol, codecs.encode(sdf, 'utf-8'))
res = inchiconv.ReadString(mol, sdf)
if mol.Empty():
root.warning(' --> ERROR on sdf')
raise Exception
moldata['smiles'] = inchiconv.WriteString(mol).strip()
return moldata
def insert_single_property_field(
name,
description,
is_integer,
source_tag,
):
if is_integer:
is_integer = True
else:
is_integer = False
field = PropertyField(name=name, description=description,
is_integer=is_integer, source_tag=source_tag)
field.save()
root.warning(' --> id=%s' % field.id)
return False
def parse_annotation(sdf, namekey):
""" parse annotation from SDF file """
# parse the sdf tags
moldata = get_sdf_tags(sdf)
# --- inchi
inchiconv.SetInAndOutFormats('sdf', 'Inchi')
mol = openbabel.OBMol()
res = inchiconv.ReadString(mol, codecs.encode(sdf, 'utf-8'))
if mol.Empty():
root.warning(' --> ERROR on sdf')
raise Exception
# standard data generated
# --- inchi/formula/weight
moldata['inchi'] = inchiconv.WriteString(mol).strip()
moldata['formula'] = mol.GetFormula()
moldata['id'] = mol.GetTitle()
if moldata['id'] == '':
moldata['id'] = 'unspecified_' \
+ ''.join(random.sample(string.digits, 6))
mol.AddHydrogens()
moldata['weight'] = str(mol.GetMolWt())
# if the name is not in sdf:
if not moldata.has_key(namekey):
moldata[namekey] = ''
# smiles
inchiconv.SetInAndOutFormats('sdf', 'smi')
mol = openbabel.OBMol()
res = inchiconv.ReadString(mol, codecs.encode(sdf, 'utf-8'))
if mol.Empty():
root.warning(' --> ERROR on sdf')
raise Exception
moldata['smiles'] = inchiconv.WriteString(mol).strip()
return moldata
def update_compounds(
filename,
library_name,
version,
namekey,
idkey,
):
"""update an existing library:
........ library version += 1 """
library = get_library(library_name, version)
if check_or_update_compounds(filename, library, namekey, idkey,
check_only=True):
root.warning(' --> : compound update found. Will update now.')
check_or_update_compounds(filename, library, namekey, idkey,
check_only=False)
else:
root.warning(' --> : No compound update found.')
def create_library_w_header(header):
""" given header, create library with latest version + 1"""
# library_set w/ given header
libs = header.library_set.all()
# no library yet
if len(libs) == 0:
version = 0
else:
version = libs.latest().version + 1
library = Library(header=header, version=version)
library.save()
root.warning(' --> : New library created: %s' % library)
return library
def insert_property_fields(jfile):
"""file format: NAME, DESCRIPTION, IS_INTEGER?, SOURCE_TAG"""
fp = file(jfile)
counter = 0
for i in fp.readlines():
counter += 1
try:
(name, description, is_integer, source_tag) = \
i.strip().split(',')[0:4]
except:
root.warning(' --> line %s' % counter)
raise
insert_single_property_field(name, description, is_integer,
source_tag)
fp.close()
return False
def gen_joelib_property(sdf):
"""run and parse the property output """
# save the input in FS
t = tempfile.NamedTemporaryFile(suffix='.sdf')
t.write(codecs.encode(sdf, 'utf-8'))
t.flush()
# prepare the output file
(f, out) = tempfile.mkstemp(suffix='.sdf')
os.close(f)
# convert
cmd = \
"""JAVA_HOME=%s/support/jre1.6.0_06/ JOELIB2=%s/support/JOELib2-alpha-20070303/ %s/support/JOELib2-alpha-20070303/moleculeConversion.sh +d +h -iSDF -osdf "%s" "%s" > /dev/null""" \
% (cur_dir, cur_dir, cur_dir, t.name, out)
if os.uname()[1] == 'chemmineweb':
cmd = \
"""JAVA_HOME=/usr/lib/jvm/java-6-sun JOELIB2=%s/support/JOELib2-alpha-20070303/ %s/support/JOELib2-alpha-20070303/moleculeConversion.sh +d +h -iSDF -osdf "%s" "%s" > /dev/null""" \
% (cur_dir, cur_dir, t.name, out)
elif os.uname()[1] == 'biocluster':
cmd = \
"""JAVA_HOME=/usr/lib/jvm/java-1.5.0-sun-1.5.0.14/ JOELIB2=%s/support/JOELib2-alpha-20070303/ %s/support/JOELib2-alpha-20070303/moleculeConversion.sh +d +h -iSDF -osdf "%s" "%s" > /dev/null""" \
% (cur_dir, cur_dir, t.name, out)
root.warning(' --> running:%s' % cmd)
if os.system(cmd) != 0:
os.unlink(out)
raise 'cannot run JOELib'
# read and parse
f = file(out)
tags = get_sdf_tags(codecs.decode(f.read(), 'utf-8'))
f.close()
# clean
os.unlink(out)
return tags
def insert_single_compound(
moldata,
sdf,
namekey,
idkey,
user,
):
""" insert single compound into database """
cid = moldata[idkey]
name = moldata[namekey]
if '\n' in name:
name = name.split('\n')[0]
# compound
c = Compound(
cid=cid,
name=name,
formula=moldata['formula'],
weight=moldata['weight'],
inchi=moldata['inchi'],
smiles=moldata['smiles'],
user=user,
)
# sdf_file=s)
c.save()
c_id = c.id
root.warning(' -->new compound inserted: c_id=%s, cid=%s' % (c_id,
cid))
# sdf file
s = SDFFile(sdffile=sdf, compound=c)
s.save()
sdfid = s.id
return c.id
plates = []
plate = compound.plate_set.all()
for p in plate:
plates.append(dict(format=p.format, plate=p.plate, well=p.well))
# --- get duplicates via fingerprint
try:
fpt = Fingerprint.objects.get(compound=compound)
dup_fpts = \
Fingerprint.objects.filter(fingerprint=fpt.fingerprint)
if len(dup_fpts) == 1:
dup_fpts = []
except:
root.warning('failed to get duplicates for %s' % compound)
dup_fpts = []
if 'addWorkbench' in request.POST:
input_mode = 'view'
addToWorkbench(compound=compound,
username=request.user.username)
matches = None
request.user.message_set.create(message='Compound added to workbench'
)
return render_to_response('compound.html', dict(
libname=library,
compound=compound,
annotations=annotations,
annotations_head=annotations.items()[:11],
def insert_property(
library_name,
library_ver,
offset,
limit,
filename,
):
"""for JOELib property
.... good for small library as it is slow
.... for library with large amount of compounds:
.... 1. use extra scripts to generate Property files
........2. use insert_property_from_file() """
def do_process(compounds, filep, tags):
for cmp in compounds:
sdf = SDFFile.objects.get(compound=cmp).sdffile
print ' running compound:%s' % cmp.id
try:
propertydata = gen_joelib_property(sdf)
_insert_property(cmp, propertydata, tags)
except:
root.warning(' --> cannot run property on %s'
% cmp.id)
raise
library = get_library(library_name, library_ver)
try:
offset = int(offset)
limit = int(limit)
except:
raise
# number of compounds
filep = file(filename, 'w')
compounds = library.compound_set.all()
count = len(compounds)
root.warning(' --> %s compounds to be processed' % count)
tags = []
for i in PropertyField.objects.values_list('source_tag'):
tags.append(i[0])
limit = limit + offset
count = 0
while offset < limit:
cur_limit = 1000
# cur_limit = 7 # for testing only
if offset + cur_limit < limit:
cur_limit = limit - offset
root.warning(' --> cur_limit:%s' % cur_limit)
filep_progress = file(filename + '.progress', 'w')
filep_progress.write('working on %s - %s\n' % (offset, offset
+ cur_limit))
filep_progress.flush()
# process
cmps = compounds[offset:offset + cur_limit]
do_process(cmps, filep, tags)
filep_progress.close()
count += 1
offset += cur_limit
filep.close()
return False
