def smiles_extraction(cas_keys, main):
if len(cas_keys) != 0:
try:
top_ck = mode(cas_keys) # take most frequent cas key
top_cmpds = pcp.get_compounds(top_ck, 'name')
if len(top_cmpds) != 0:
top_smiles = top_cmpds[
0].isomeric_smiles # pick top result (best match) for cas key
new_smiles[main] = top_smiles
except statistics.StatisticsError: # two are equally frequent
top_ck = []
smiles1 = []
lst_count = [x for x in set(cas_keys) if cas_keys.count(x) > 1]
for i in range(len(lst_count)):
top_ck.append(lst_count[i])
for j in top_ck:
top_cmpds = pcp.get_compounds(j, 'name')
if len(top_cmpds) != 0:
top_smiles = top_cmpds[0].isomeric_smiles
smiles1.append(top_smiles)
new_smiles[main] = [smiles1]
def save(self, *args, additional_data=None, cid2=False, **kwargs):
"""
Sets data for various fields. Assumes that if the object does not have inchikey data that it has a SMILES string
"""
if not all([self.smiles, self.cid_number, self.chemical_properties]):
try:
pcp_data = pcp.get_compounds(self.inchikey, 'inchikey')[0] if hasattr(self, 'inchikey') else \
pcp.get_compounds(self.smiles, 'smiles')[0]
except (IndexError, pcp.BadRequestError):
raise ValidationError('Something went wrong')
if not self.iupac_name:
self.iupac_name = pcp_data.iupac_name or 'n/a'
self.smiles = pcp_data.isomeric_smiles or pcp_data.canonical_smiles or ''
self.set_chemical_data(pcp_query=pcp_data)
if not self.chemical_name:
self.chemical_name = self.scrape_compound_name(self.cid_number) or \
self.synonyms.split(',')[0] if self.synonyms != 'n/a' else ''
if cid2 and len(self.smiles.split('.')) > 1:
try:
self.cid_number_2 = pcp.get_compounds(
self.smiles.split('.')[0], 'smiles')[0].cid
except (IndexError, pcp.BadRequestError):
pass
if len(self.smiles) > 200:
self.smiles = ''
if self.iupac_name and len(self.iupac_name) > 250:
self.iupac_name = ''
if hasattr(self, 'activity') and not self.activity and hasattr(
self, 'category') and self.category == 1:
act_find = FindActivity(self.chemical_name)
self.activity = act_find.activity
super(CompoundMixin, self).save(*args, **kwargs)
def FingerprintCOMP(a1, a2, key):
p = pcp.get_compounds(a1, "name")
d = pcp.get_compounds(a2, "name")
for c in p:
if key == "isomeric_smiles":
a = c.isomeric_smiles
elif key == "molecular_formula":
a = c.molecular_formula
elif key == "fingerprint":
a = c.fingerprint
else:
a = c.fingerprint
print "Input 1 = " + a1
print a
print len(a)
for i in d:
if key == "isomeric_smiles":
b = i.isomeric_smiles
elif key == "molecular_formula":
b = i.molecular_formula
elif key == "fingerprint":
b = i.fingerprint
else:
b = i.fingerprint
print "Input 2 = " + a2
print b
print len(b)
s = difflib.SequenceMatcher(None, a, b)
print a1 + " and " + a2 + " Have a similarity score of " + str(
s.ratio() * 100)
for block in s.get_matching_blocks():
print "Input 1 1[%d] and Input 2[%d] match for %d elements" % block
def get_molecule_cid(file_name, test=False):
"""
Take a file_name, remove the ending and prefix and return the pubchem
molecule object. If the file_name does not have the cas number, try the
name of the system, else return None so as to disreguard the molecule
from training.
"""
my_file = file_name.split("_")
#try to get the cas # if not then try name.
cas = my_file[1]
if test:
print("Get CID, CAS: ", cas)
else:
pass
try:
cid = pcp.get_compounds(cas, 'name')[0]
except:
#file doesn't have a cas number with it, try name
cas = my_file[-1]
cas.replace("_", ' ')
my_name = cas.split(".")
try:
cid = pcp.get_compounds(my_name[0], 'name')[0]
except:
cid = None
finally:
if test:
print("Get CID, CID: ", cid)
return cid
def mol_prop_gen(dataframe,outputpickle):
#loading initial data
# dataframe = pd.read_csv(filename)
# newdf = pd.DataFrame()
finaldf = pd.DataFrame()
for index, row in dataframe.iterrows():
if index == 0:
name = row['Name']
inchi = row['InChI']
cmpd = pcp.get_compounds(inchi,'inchi')
props = cmpd[0].to_dict(properties=['cactvs_fingerprint',
'isomeric_smiles', 'xlogp', 'rotatable_bond_count','charge','complexity',
'exact_mass','fingerprint'])
smiles=props['isomeric_smiles']
props['mol']=Chem.MolFromSmiles(smiles)
props['RT'] = row['RT']
props['Name'] = name
props['System'] = row['System']
desc = np.array(fps_plus_mw(props['mol']))
descdf = pd.DataFrame(desc)
descdf = descdf.T
descdf.reindex([index])
newdf=pd.DataFrame(props,index=[index])
finaldf=pd.concat([descdf,newdf],axis=1)
print('test')
else:
inchi = row['InChI']
try:
cmpd = pcp.get_compounds(inchi,'inchi')
except:
print('line bypassed')
pass
try:
props = cmpd[0].to_dict(properties=['cactvs_fingerprint','isomeric_smiles', 'xlogp', 'rotatable_bond_count','charge','complexity','exact_mass','fingerprint'])
except:
print('line bypassed')
pass
name = row['Name']
smiles=props['isomeric_smiles']
props['mol']=Chem.MolFromSmiles(smiles)
props['RT'] = row['RT']
props['Name'] = name
props['System'] = row['System']
newdf=pd.DataFrame(props,index=[index])
desc = np.array(fps_plus_mw(props['mol']))
cols=range(len(desc))
descdf=pd.DataFrame(desc)
descdf = descdf.T
descdf.index = [index]
# descdf = descdf.T
# descdf = pd.DataFrame(descdf, index=[index])
interdf = pd.concat([descdf,newdf],axis=1)
finaldf = finaldf.append(interdf)
print('on index ' + str(index+1) + ' of ' + str(len(dataframe)))
finaldf.to_pickle(outputpickle)
#mol_prop_gen('PredRetDB.csv','test.pickle')
def set_chem_data(self):
for d in self.drugs_data:
try:
pcp_query = pcp.get_compounds(d['cid_number'], 'cid')[0]
smiles = pcp_query.canonical_smiles
d.update({
'smiles':
smiles,
'inchikey':
pcp_query.inchikey,
'iupac_name':
pcp_query.iupac_name
or cirpy.resolve(smiles, 'iupac_name', ['smiles']),
'chemical_properties':
dict_from_query_object(smiles, pcp_query, additional=True),
})
if len(smiles.split('.')) > 1:
d.update({
'cid_number_2':
pcp.get_compounds(smiles.split('.')[0],
'smiles')[0].cid
})
except (IndexError, TypeError, pcp.BadRequestError):
self.drugs_data.remove(d)
return self.drugs_data
def get_iupac_name_from_smiles(smiles):
try:
cpd = pcp.get_compounds(smiles, 'smiles')
except pcp.PubChemHTTPError:
cpd = pcp.get_compounds(smiles, 'smiles')
if len(cpd) != 1: return
else: return cpd[0].iupac_name
def convert_feature_to_standard_pubchem(df):
# create an empty df for later use
dummyarray = np.empty((len(df), 6))
dummyarray[:] = np.nan
df0 = pd.DataFrame(
dummyarray,
columns=['iupac', 'synonyms', 'compounds', 'formulae', 'cid', 'flag'])
# deal with second column with CAS number
print('2nd')
df2 = df.iloc[:, 1].copy()
df2 = pd.Series(df2).replace('nan', np.nan)
# df2[mask1] = None
df2 = [
str(string).lstrip("'") if string is not np.nan else np.nan
for string in df2
] # remove ' from the beginning of the cas number
idx2 = [
pcp.get_compounds(component, 'name')
if component is not np.nan else [] for component in df2
]
df0 = _fill_df_pubchem(df0, idx2)
mask2 = [not not elem for elem in idx2]
df0.loc[mask2, 'flag'] = 'from column B'
# deal with the third column with mix number
print('3rd')
df3 = df.iloc[:, 4].copy()
df3[mask2] = np.nan
idx3 = [
pcp.get_compounds(component, 'name')
if component is not np.nan else [] for component in df3
]
df0 = _fill_df_pubchem(df0, idx3)
mask3 = [not not elem for elem in idx3]
df0.loc[mask3, 'flag'] = 'from column E'
# deal with first column with given names
print('1st')
df1 = df.iloc[:, 0].copy()
df1[list(pd.Series(mask3) | pd.Series(mask2))] = np.nan
name2trans = list(df1)
name_transed_cls = [
Translator().translate(name, src='nl', dest='en')
if name is not np.nan else [] for name in name2trans
]
# name_transed_cls = name2trans
idx1 = [
pcp.get_compounds(component.text, 'name') if not not component else []
for component in name_transed_cls
]
df0 = _fill_df_pubchem(df0, idx1)
mask1 = [not not elem for elem in idx1]
df0.loc[mask1, 'flag'] = 'from column A'
return df0
def geometry_from_pubchem(name, structure=None):
"""Function to extract geometry using the molecule's name from the PubChem
database. The 'structure' argument can be used to specify which structure
info to use to extract the geometry. If structure=None, the geometry will
be constructed based on 3D info, if available, otherwise on 2D (to keep
backwards compatibility with the times when the argument 'structure'
was not implemented).
Args:
name: a string giving the molecule's name as required by the PubChem
database.
structure: a string '2d' or '3d', to specify a specific structure
information to be retrieved from pubchem. The default is None.
Recommended value is '3d'.
Returns:
geometry: a list of tuples giving the coordinates of each atom with
distances in Angstrom.
"""
import pubchempy
if structure in ['2d', '3d']:
pubchempy_molecule = pubchempy.get_compounds(name,
'name',
record_type=structure)
elif structure is None:
# Ideally get the 3-D geometry if available.
pubchempy_molecule = pubchempy.get_compounds(name,
'name',
record_type='3d')
# If the 3-D geometry isn't available, get the 2-D geometry instead.
if not pubchempy_molecule:
pubchempy_molecule = pubchempy.get_compounds(name,
'name',
record_type='2d')
else:
raise ValueError('Incorrect value for the argument structure=%s' %
structure)
# Check if pubchempy_molecule is an empty list or None
if not pubchempy_molecule:
print(
'Unable to find structure info in the PubChem database for the specified molecule "%s".'
% name)
return None
pubchempy_geometry = \
pubchempy_molecule[0].to_dict(properties=['atoms'])['atoms']
geometry = [(atom['element'], (atom['x'], atom['y'], atom.get('z', 0)))
for atom in pubchempy_geometry]
return geometry
def get_iupac_name_from_smiles(smiles):
if not pcp:
raise RuntimeError('No Pubchempy')
try:
cpd = pcp.get_compounds(smiles, 'smiles')
except pcp.PubChemHTTPError:
cpd = pcp.get_compounds(smiles, 'smiles')
if len(cpd) != 1:
return
else:
return cpd[0].iupac_name
def process_bioactive_identifier(request):
cas_no = request.GET.get('cas_number')
inchikey = request.GET.get('inchikey', '').strip()
obj = None
if cas_no:
obj = Bioactive.objects.filter(
chemical_properties__synonyms__icontains=cas_no).first()
elif inchikey:
obj = Bioactive.objects.filter(inchikey__exact=inchikey).first()
if obj:
data = {
'object_exists': obj.get_absolute_url(),
'object_exists_name': str(obj),
}
return JsonResponse(data)
try:
iupac_name = None
if cas_no:
smiles = cirpy.query(cas_no, 'smiles')[0].value
if '.' in smiles:
smiles = [i for i in smiles.split('.') if len(i) > 5][0]
pcp_query = pcp.get_compounds(smiles, 'smiles')[0]
if not pcp_query.iupac_name:
iupac_name = cirpy.resolve(smiles, 'iupac_name', ['smiles'])
else:
pcp_query = pcp.get_compounds(inchikey, 'inchikey')[0]
if not pcp_query.iupac_name:
iupac_name = cirpy.resolve(inchikey, 'iupac_name',
['stdinchikey'])
if not pcp_query.cid:
raise IndexError
except (IndexError, pcp.BadRequestError):
return JsonResponse({'error': 'No compound found for this CAS number'})
data = {
'chemical_name':
Bioactive.scrape_compound_name(pcp_query.cid),
'iupac_name':
pcp_query.iupac_name or iupac_name or 'n/a',
'inchikey':
pcp_query.inchikey,
'structure_url':
'https://pubchem.ncbi.nlm.nih.gov/image/imgsrv.fcgi?cid={}&t=l'.
format(pcp_query.cid),
'hidden_cid':
pcp_query.cid,
'smiles':
pcp_query.isomeric_smiles or pcp_query.canonical_smiles or '',
}
return JsonResponse(data)
def _load_from_pubchem(self, name):
'''
Method used to load data from pubchem
name - name of compound to search
returns nothing
'''
import pubchempy as pcp
try:
name = int(name)
except:
pass
record_type = '3d'
mol = pcp.get_compounds(name, ('name', 'cid')[type(name) is int],
record_type=record_type)
if len(mol) == 0:
utils.message(
'Error: Could not find 3d structure of {name}... Attempting to find 2d structure...',
'red')
record_type = '2d'
mol = pcp.get_compounds(name, ('name', 'cid')[type(name) is int],
record_type=record_type)
if len(mol) == 0:
utils.message('Error: No structural data found for {name}.', 'red')
else:
mol = mol[0]
coords = np.asarray([[a.x, a.y, a.z] for a in mol.atoms])
coords = np.where(coords == None, 0, coords).astype(float)
elements = np.asarray([a.element for a in mol.atoms])
self.name = name.capitalize()
self.atoms = []
[
self.atoms.append(Atom(elements[i], coords[i]))
for i in range(len(coords))
]
if record_type == '3d':
self.save_to_xyz(os.getcwd() +
rf'\Molecules\{name.lower()}.xyz')
self._mol_load_finish()
def getChemicalName(smiles):
try:
name = pcp.get_compounds(smiles, "smiles")[0].iupac_name
return name
except:
print("Pubchempy could not convert SMILES to a IUPAC name")
return smiles
def sid_to_smiles(sid):
"""Takes an SID and prints the associated SMILES string."""
substance = pc.Substance.from_sid(sid)
cid = substance.standardized_cid
compound = pc.get_compounds(cid)[0]
return compound.isomeric_smiles
def import_from_pubchem():
compounds = read_csv()
# cria uma planilha no mesmo local do arquivo .py
workbook = xlsxwriter.Workbook(filename='to_database.xlsx')
# cria uma aba
worksheet = workbook.add_worksheet(name='results')
row = 1
print('\nEstabelecendo conexão com o PubChem...')
# para cada composto na tabela
for comp in compounds:
# pega dados no pubChem
results = pcp.get_compounds(comp[0], 'smiles')
# baixa a imagem de composto
pcp.download('PNG',
os.path.join(CURR_PATH, 'images', comp[0] + '.png'),
comp[0],
'smiles',
overwrite=True)
# para cada resultado, escreve na planilha nova o SMILES, o aroma, o nome IUPAC e a fórmula molecular
for c in results:
print('\nComposto ' + c.iupac_name)
worksheet.write(row, 0, comp[0])
worksheet.write(row, 1, comp[1])
worksheet.write(row, 2, c.iupac_name)
worksheet.write(row, 3, c.molecular_formula)
row += 1
workbook.close()
print('Pronto! Compostos Atualizados')
def get_compound_from_smiles(smiles):
attempts = 5
time_delay = 1 # in seconds
while attempts >= 1:
try:
compounds = pcp.get_compounds(smiles, namespace='smiles')
cid = compounds[0].cid
if cid == None:
print(
'No PubChem record'
) # https://pubchempy.readthedocs.io/en/latest/guide/gettingstarted.html
return None
compound = pcp.Compound.from_cid(cid)
except:
attempts -= 1
print('Could not get compound. ' + str(attempts) +
' attempts remaining.')
time.sleep(time_delay)
else:
return compound
if attempts <= 0:
print(
'Failed to get compound from smiles after exhausting all attempts')
return None
def get_smiles(idx, names=None, cids=None, sids=None, binding_db=None):
try:
cid = int(cids[idx])
smiles_string = binding_db.loc[binding_db['PubChem CID'] ==
cid]["Ligand SMILES"].values[0]
except:
try:
sid = int(sids[idx])
smiles_string = binding_db.loc[binding_db['PubChem SID'] ==
sid]["Ligand SMILES"].values[0]
except:
try:
cid = int(cids[idx])
smiles_string = str(
Compound.from_cid(int(cid)).isomeric_smiles)
except:
try:
chembl_id = Substance.from_sid(sids[idx]).source_id
print(chembl_id)
try:
compounds = CompoundResource()
c = compounds.get(chembl_id)
smiles_string = c["smiles"]
except:
smiles_string = binding_db.loc[
binding_db["ChEMBL ID of Ligand"] ==
chembl_id]["Ligand SMILES"].values[0]
except:
try:
name = names[idx]
cs = get_compounds(name, 'name')
smiles_string = cs[0].isomeric_smiles
except:
smiles_string = np.nan
return (smiles_string)
def get_name_from_pubchem(self, smiles: str) -> Optional[str]:
"""Tries to get the name of a molecule from the Pubchem website.
Args:
smiles: A SMILES string.
Returns:
The molecule name if a match is found else ``None``.
"""
try:
comp = get_compounds(smiles, namespace="smiles")[0]
except (BadRequestError, IndexError):
return None
traditional = comp.synonyms[0] if comp.synonyms else None
names = {"traditional": traditional, "iupac": comp.iupac_name}
match = None
for source in self.name_preference:
if source in names and names[source]:
match = names[source]
break
if isinstance(match, str):
match = match.lower()
return self._process_match(smiles, match)
def process_cas(request):
cas_no = request.GET.get('cas_number')
try:
obj = Odorant.objects.get(cas_number__exact=cas_no)
data = {
'object_exists': obj.get_absolute_url(),
'object_exists_name': str(obj),
}
return JsonResponse(data)
except ObjectDoesNotExist:
pass
try:
smiles = cirpy.query(cas_no, 'smiles')[0].value
pcp_query = pcp.get_compounds(smiles, 'smiles')[0]
cid_no = pcp_query.cid
except IndexError:
return JsonResponse({
'error': 'No compound found for this CAS number'
})
if smiles and cid_no:
data = {
'chemical_name': Odorant.scrape_compound_name(cid_no),
'iupac_name': pcp_query.iupac_name,
'structure_url': 'https://pubchem.ncbi.nlm.nih.gov/image/imgsrv.fcgi?cid={}&t=l'.format(cid_no),
'hidden_cid': cid_no,
'smiles': smiles,
}
return JsonResponse(data)
def cid_df_to_smiles(df, cid_colname):
"""
Args:
df : pandas dataframe with SID numbers
column_name (str) : name of column that contains PubChem SID numbers
Returns:
df : modified with columns containing CID and SMILES
CID becomes first column, SMILES second
unsuccessful_list : list of SIDs for which no CID or SMILES were found
"""
res = []
unsuccessful_list = []
for index, row in df.iterrows():
cid = row[cid_colname]
try:
compound = pc.get_compounds(cid)[0]
smiles = compound.canonical_smiles
res.append(smiles)
except BaseException:
res.append('none')
unsuccessful_list.append(cid)
pass
df['SMILES'] = res
#df.to_csv(r'../datasets/df_cleaned_kegg_with_smiles.csv')
return df, unsuccessful_list
def _set_inchi_pcc(self, in_str, pcp_type, elem):
"""Check pubchem compounds via API for both an inchikey and any available compound details
"""
if not in_str:
return 0
try:
pccs = pcp.get_compounds(in_str, pcp_type)
except pcp.BadRequestError as e:
print(e)
return 0
except pcp.TimeoutError as e:
print(e)
return 0
except pcp.ServerError as e:
print(e)
return 0
except URLError as e:
print(e)
return 0
except BadStatusLine as e:
print(e)
return 0
if pccs:
pcc = pccs[elem]
self.compound_info['inchikey_id'] = pcc.inchikey
self.compound_info['pubchem_id'] = pcc.cid
self.compound_info['molecular_formula'] = pcc.molecular_formula
self.compound_info['molecular_weight'] = pcc.molecular_weight
self.compound_info['exact_mass'] = pcc.exact_mass
self.compound_info['smiles'] = pcc.canonical_smiles
if len(pccs) > 1:
print('WARNING, multiple compounds for ', self.compound_info)
def download_from_list(self, name_list, type="name"):
"""Given a list of compounds, download them
Parameters
----------
name_list
type
Returns
-------
"""
results = []
for n in name_list:
try:
r = get_compounds(n, namespace=type)
except:
print("Ligand %s not found" % n)
r = []
if len(r):
results.append(r[0])
else:
results.append(None)
return results
def getCompoundsInfo(ChemIDList, MoleculeNameList):
compoundList = []
molecular_formulasList = []
isomeric_smilesList = []
CidsList = []
chemIDList = []
i = 0
for chemID, molecule in list(zip(ChemIDList, MoleculeNameList)):
#print('ChemID:', chemID)
compoundsResults = pubchempy.get_compounds(molecule, namespace='name')
#print('compoundsResults:\n', compoundsResults)
#compoundsResults.to_excel(outputPath+"result.xlsx")
for compound in compoundsResults:
#print('CID: {}\tMass: {}\tName: {}\tMolfor: {}\tSmi: {}\tSyn: {}'.format(compound.cid,compound.exact_mass,compound.iupac_name,compound.molecular_formula, compound.isomeric_smiles, compound.synonyms))
molecular_formulas=compound.molecular_formula
#MWs=compound.molecular_weight
isomeric_smiles=compound.isomeric_smiles
#synonyms=compound.synonyms
Cids=compound.cid
chemIDList.append(chemID)
compoundList.append(molecule)
molecular_formulasList.append(molecular_formulas)
isomeric_smilesList.append(isomeric_smiles)
CidsList.append(Cids)
dataframe=pd.DataFrame({'ChemID':chemIDList,
'CompoundName':compoundList,
'molecular_formula':molecular_formulasList,
'smiles':isomeric_smilesList,
'cid':CidsList})
dataframe.to_excel(outputPath+"SmilesResult.xlsx",index=False)
print(f'{i}-{chemID}:{molecule} write to excel file successfully!')
i += 1
time.sleep(3)
def apply_get_compounds(mol_name):
try:
iso_smiles = pcp.get_compounds(mol_name, 'name')[0].isomeric_smiles
except:
iso_smiles = ''
return iso_smiles
def get_from_pubchem(name, path=structures_folder):
'''
Function that downloads a molecule from pubchem
Save the molecule to path
name - name of molecule
path - path to save molecule to
'''
import pubchempy as pcp
mols = pcp.get_compounds(name, 'name', record_type='3d')
#check if there was a match
if len(mols) == 0:
raise Exception(f'Molecule {mol} not found on PubChem or on disk.')
#save xyz file to disk
else:
coords = np.asarray([[a.x, a.y, a.z] for a in mols[0].atoms])
coords = np.where(coords == None, 0, coords).astype(float)
elements = np.asarray([a.element for a in mols[0].atoms])
mol_path = path + '\\' + name + '.xyz'
with open(mol_path, 'w+') as f:
f.write(f'{len(elements)}\n')
f.write('Downloaded from PubChem\n')
for i, e in enumerate(elements):
f.write(
f'{e: <2} \t {coords[i][0]: >8.5f} \t {coords[i][1]: >8.5f} \t {coords[i][2]: >8.5f}\n'
)
return mol_path
def get_molecular_weight(name):
try:
cpd = pcp.get_compounds(name, "name")
return cpd[0].to_dict(
properties=["molecular_weight"])["molecular_weight"]
except:
return 300 #not real mw, but it's in the middle - therefore won't be misued in min/max calculations
def name2inchikey(odor_names):
"""
Args:
odor_names (iterable of strs): the odors to be translated to
inchikeys
Returns:
odor_inchikeys (list of strs): the inchikeys for the input odors
inchikey2name (dict: str -> str): for quickly getting the name for an
inchikey
"""
# TODO support single str inputs too
odor_inchikeys = []
inchikey2name = dict()
for o in odor_names:
matches = pcp.get_compounds(o, 'name')
if len(matches) > 1:
print('WARNING: more than one pubchem match for {}. ' + \
'ambiguous!'.format(o))
continue
elif len(matches) == 0:
print('WARNING: no pubchem matches found for {}!'.format(o))
continue
match = matches[0]
odor_inchikeys.append(match.inchikey)
inchikey2name[match.inchikey] = o
return odor_inchikeys, inchikey2name
def _default(self):
if self.app.pargs.by_name:
compounds = pubchempy.get_compounds(self.app.pargs.name_or_id, 'name')
results = [[compound.synonyms[0], 'pubchem.compound', compound.cid, compound.inchi] for compound in compounds]
else:
unichem = bioservices.UniChem()
structure = unichem.get_structure(int(float(self.app.pargs.name_or_id)), self.app.pargs.namespace)
if structure:
results = [['', self.app.pargs.namespace, self.app.pargs.name_or_id, structure['standardinchi']]]
else:
results = []
if not results:
print('Unable to find structure', file=sys.stderr)
return
lens = [
max(4, max(len(str(r[0])) for r in results)),
max(9, max(len(str(r[1])) for r in results)),
max(2, max(len(str(r[2])) for r in results)),
max(9, max(len(str(r[3])) for r in results)),
]
format = '{{:<{}}} {{:<{}}} {{:<{}}} {{:<{}}}'.format(*lens)
print(format.format('Name', 'Namespace', 'ID', 'Structure'))
print(format.format('=' * lens[0], '=' * lens[1], '=' * lens[2], '=' * lens[3]))
for result in results:
print(format.format(*result))
def get_pubchem_by_name(dir):
files = [
f for f in os.listdir(dir) if os.path.isfile(os.path.join(dir, f))
]
for in_file in files:
#print in_file
file_name, extension = os.path.splitext(str(in_file))
if "smiles" not in file_name:
out_file = file_name + "_smiles" + extension
print "processing:", in_file, "writing to:", out_file
with open(dir + in_file, 'r') as input, open(dir + out_file,
'w') as output:
mol_info = [re.split(',|:', line)[-2:] for line in input]
#print mol_info
smiles_string = ""
for molecule in mol_info:
query_results = pubchempy.get_compounds(
molecule[0], "name")
if len(query_results) != 1:
print "Query for:", molecule[0], "yielded", len(
query_results), "results"
if len(query_results) > 1:
print "Using 1st result: ", query_results[
0].canonical_smiles
if len(query_results) > 0:
smiles_string += query_results[
0].canonical_smiles + '\n'
output.write(smiles_string)
def convertsmiles():
t_smiles.set('')
t_sol.set('')
t_lip.set('')
t_sasc.set('')
molecule = pcp.get_compounds(t_name.get(), 'name')
print('molecule')
print(molecule[0])
#print('canocical_smile', molecule[0].canonical_smiles)
print('isomeric_smile', molecule[0].isomeric_smiles)
mol_canonical_smiles = molecule[0].canonical_smiles
mol_isomeric_smiles = molecule[0].isomeric_smiles
t_smiles.set(mol_isomeric_smiles)
mol_ = Chem.MolFromSmiles(mol_isomeric_smiles)
Draw.MolToFile(mol_, 'tmp.png')
global image_
image_open = Image.open('tmp.png')
image_ = ImageTk.PhotoImage(image_open, master=frame1)
canvas.create_image(150, 75, image=image_)
def product_prediction_algorithm(target_smiles_list, df, row):
start = timeit.default_timer()
targets = []
substrates = []
products = []
results = []
enzymes = []
for x in xrange(len(df)):
try:
results.append( explore_substrate(target_smiles_list[x], row, df) )
substrates.append( df['Substrates'].irow(row) )
products.append( df['Products'].irow(row) )
targets.append( target_smiles_list[x] )
enzymes.append( df['Enzymes'].irow(row) )
except:
pass
stop = timeit.default_timer()
#unique_prods, unique_substrates, unique_products, unique_enzymes = filter_results(results, substrates, products, TestUni, enzymes)
unique_prods, unique_substrates, unique_products, unique_enzymes = results, substrates, products, enzymes
print "Finished reactions... searching hits on PubChem... "
#search pubchem for unique hits
novel_compounds = []
for i in unique_prods:
try:
searches = pcp.get_compounds('CanonicalSMILES', str(Chem.MolToSmiles(i)), 'smiles')
if str(searches) == '[Compound()]':
novel_compounds.append( i )
except:
pass
#put things in a df
unique_prods_smiles = []
novel_smiles = []
for i in unique_prods:
unique_prods_smiles.append( Chem.MolToSmiles(i) )
for i in novel_compounds:
novel_smiles.append( Chem.MolToSmiles(i) )
novels =[ ]
for i in unique_prods_smiles:
if i in novel_smiles:
novels.append( 'Novel' )
else:
novels.append( 'Found in Pubchem' )
Results = pd.DataFrame({'Native Substrate': unique_substrates, 'Native Product': unique_products, 'Products': unique_prods_smiles, 'Novel Compound?': novels, 'Enzymes': unique_enzymes})
Results = Results.drop_duplicates(cols=['Products'])
print 'Novel Compounds found...'
print len(novel_compounds)
print "Runtime..."
print stop - start
product_pictures = Draw.MolsToGridImage(unique_prods,molsPerRow=8, includeAtomNumbers=False)
return Results, product_pictures
def chunks_get_compounds(l,n):
""" Yield successive n-sized chunks from l. (From Stack Overflow)"""
for i in xrange(0,len(l),n):
try:
yield pcb.get_compounds(l[i:i+n])
except Exception, e:
print e
pass
def pubchem_search(comp_name, search_type='name'):
iupac = ''
inchi = ''
inchi_key = ''
smiles = ''
cid = ''
formula = ''
synonyms = ''
structure = ''
try:
compound = None
# For this to work on Mac, run: cd "/Applications/Python 3.6/"; sudo "./Install Certificates.command
# or comment out the line below:
# ssl._create_default_https_context = ssl._create_unverified_context # If no root certificates installed
pubchem_compound = get_compounds(comp_name, namespace=search_type)
try:
compound = pubchem_compound[0] # Only read the first record from PubChem = preferred entry
except IndexError:
logger.info('Could not find PubChem compound for ' + comp_name) # Nothing was found
if compound:
inchi = compound.inchi
inchi_key = compound.inchikey
smiles = compound.canonical_smiles
iupac = compound.iupac_name
iupac = iupac.replace('~', '').replace('{', '').replace('}', '')
cid = compound.cid
formula = compound.molecular_formula
for synonym in compound.synonyms:
if get_relevant_synonym(synonym):
synonyms = synonyms + ';' + synonym
if synonyms:
synonyms = synonyms.replace(";", "", 1) # Remove the leading ";"
logger.debug('Searching PubChem for "' + comp_name + '", got cid "' + str(cid) +
'" and iupac name "' + iupac + '"')
except Exception as error:
logger.error("Unable to search PubChem for compound " + comp_name)
logger.error(error)
return iupac, inchi, inchi_key, smiles, cid, formula, synonyms, structure
import pandas as pd
import pubchempy as pcp
df=pd.read_csv('SAMPL4.csv')
for oname in df.NAME:
cs = pcp.get_compounds(oname, 'name')
print oname,cs[0].canonical_smiles
df['NSMILES'] = [pcp.get_compounds(oname, 'name')[0].canonical_smiles for oname in df.NAME]
print df.to_csv(index=False)
smiles = []
for i in positive_smiles:
smiles.append( i )
for i in negative_smiles:
smiles.append( i )
for i in range(len(positive_smiles)):
activity.append( 1 )
for i in range(len(negative_smiles)):
activity.append( 0 )
for i in negative_smiles:
try:
results = pcp.get_compounds(str(i), 'smiles')
cpn = pcp.Compound.from_cid( int(str(results[0])[9:-1]) )
compound_names.append( cpn.iupac_name )
except:
compound_names.append( 'error parsing' )
# <codecell>
training_set = pd.DataFrame({'Smiles': smiles, 'Activity': activity, 'Compound Name': compound_names})
# <codecell>
import numpy as np
import pandas as pd
from random import sample
def get_name(name):
chemical_name = pcp.get_compounds(name, 'name')
try:
return chemical_name[0].isomeric_smiles
except IndexError:
return "\n"
def rf_find_alternative_substrates(enzyme, dataset, threshold):
###grabs the substrates of the given enzyme and their smiles
compound_names, positive_smiles = retrieve_enzyme_substrates(enzyme)
###generate 3D coordinates for each molecule
positive_structures = []
for i in positive_smiles:
positive_structures.append( Chem.MolFromSmiles(i) )
endogenous_structures = []
for i in endogenous_steroids:
endogenous_structures.append( Chem.MolFromSmiles(i) )
for m in positive_structures: AllChem.Compute2DCoords(m)
for m in endogenous_structures: AllChem.Compute2DCoords(m)
positive_fps=[AllChem.GetMorganFingerprintAsBitVect(x,2) for x in positive_structures]
endogenous_fps=[AllChem.GetMorganFingerprintAsBitVect(x,2) for x in endogenous_structures]
#recombine endogenous structures with their scores
sims = DataStructs.BulkTanimotoSimilarity(positive_fps[0],endogenous_fps)
nbrs = sorted(zip(sims,endogenous_structures),reverse=False)
#grab bottom 10% of matches
negative_structures = [x[1] for x in nbrs[:12]]
negative_smiles = []
for i in negative_structures:
negative_smiles.append( Chem.MolToSmiles(i) )
#Draw.MolsToGridImage([x[1] for x in nbrs[:12]],legends=['%.4f'%x[0] for x in nbrs])
###gets the information for each molecule that we've grabbed thus far
activity = []
smiles = []
for i in positive_smiles:
smiles.append( i )
for i in negative_smiles:
smiles.append( i )
for i in range(len(positive_smiles)):
activity.append( 1 )
for i in range(len(negative_smiles)):
activity.append( 0 )
for i in negative_smiles:
try:
results = pcp.get_compounds(str(i), 'smiles')
cpn = pcp.Compound.from_cid( int(str(results[0])[9:-1]) )
compound_names.append( cpn.iupac_name )
except:
compound_names.append( 'error parsing' )
###our model
training_set = pd.DataFrame({'Smiles': smiles, 'Activity': activity, 'Compound Name': compound_names})
###identity is in the RF_functions.py file
values = []
mols = training_set['Smiles'].tolist()
activities = training_set['Activity'].tolist()
for i in range(len(dataset)):
if str( rf_classifier(mols, activities, dataset[i], threshold) ) == '[1]':
if dataset[i] not in mols:
values.append( dataset[i] )
values = canonicalize_smiles(values)
return values, training_set
def cross_validate_sliding_threshold(ec_number):
'''Cross validates the input enzyme across a sliding identity threshold. Output will be the average accuracy over 100 simulations per each threshold.'''
#Compute the coordinates of both the positive hits and all the endogenous steroids
positive_structures = []
for i in positive_smiles:
positive_structures.append( Chem.MolFromSmiles(i) )
endogenous_structures = []
for i in endogenous_steroids:
endogenous_structures.append( Chem.MolFromSmiles(i) )
for m in positive_structures: AllChem.Compute2DCoords(m)
for m in endogenous_structures: AllChem.Compute2DCoords(m)
positive_fps=[AllChem.GetMorganFingerprintAsBitVect(x,2) for x in positive_structures]
endogenous_fps=[AllChem.GetMorganFingerprintAsBitVect(x,2) for x in endogenous_structures]
#recombine endogenous structures with their scores
sims = DataStructs.BulkTanimotoSimilarity(positive_fps[0],endogenous_fps)
nbrs = sorted(zip(sims,endogenous_structures),reverse=False)
#grab most disimilar matches, we're arbitrarily picking 12
negative_structures = [x[1] for x in nbrs[:12]]
negative_smiles = []
for i in negative_structures:
negative_smiles.append( Chem.MolToSmiles(i) )
#Let's code the cross validation section
activity = []
smiles = []
compound_names = []
for i in positive_smiles:
smiles.append( i )
for i in negative_smiles:
smiles.append( i )
for i in range(len(positive_smiles)):
activity.append( 1 )
for i in range(len(negative_smiles)):
activity.append( 0 )
for i in negative_smiles:
try:
results = pcp.get_compounds(str(i), 'smiles')
cpn = pcp.Compound.from_cid( int(str(results[0])[9:-1]) )
compound_names.append( cpn.iupac_name )
except:
compound_names.append( 'error parsing' )
for i in positive_smiles:
try:
results = pcp.get_compounds(str(i), 'smiles')
cpn = pcp.Compound.from_cid( int(str(results[0])[9:-1]) )
compound_names.append( cpn.iupac_name )
except:
compound_names.append( 'error parsing' )
training_set = pd.DataFrame({'Smiles': smiles, 'Activity': activity, 'Compound Name': compound_names})
#Scrambles up the training_set dataframe so I can test the rest of it
percentages = list(np.arange(0, 1.05, 0.10))
validation_per_percent = []
for iteration in percentages:
cross_validation_scores = []
for x in xrange(100):
# given data frame df
# create random index
rindex = np.array(sample(xrange(len(training_set)), int( len(training_set)*.25 )))
#get some random rows from df
pulled_data = training_set.ix[rindex]
pulled_smiles = pulled_data['Smiles'].tolist()
training_set_culled = training_set.drop(rindex)
predicted_values = []
actual_values = pulled_data['Activity'].tolist()
mols = training_set_culled['Smiles'].tolist()
activities = training_set_culled['Activity'].tolist()
try:
for j in range(len(pulled_smiles)):
if rf_validate(mols, activities, pulled_smiles[j]) > iteration:
if pulled_smiles[j] not in mols:
predicted_values.append( 1 )
else:
predicted_values.append( 0 )
#cross_validation_scores.append( predicted_values == actual_values )
sm = difflib.SequenceMatcher(None, predicted_values, actual_values) ###uses the Ratcliff and Obershelp algorithm for matching these lists
cross_validation_scores.append( sm.ratio() )
except:
pass
#validation_per_percent.append( float(cross_validation_scores.count(True))/len(cross_validation_scores) )
validation_per_percent.append( mean(cross_validation_scores) )
return validation_per_percent
names=[]
formula=""
inchi=""
for line in open(sys.argv[1]):
if "CH$NAME" in line:
names.append(''.join(line.split(" ")[1:]).strip().lower())
if "CH$FORMULA" in line:
formula=''.join(line.split(" ")[1:]).strip()
if "CH$IUPAC" in line:
inchi=''.join(line.split(" ")[1:]).strip()
if "CH$SMILES" in line:
smiles=''.join(line.split(" ")[1:]).strip()
results = pcp.get_compounds(inchi, namespace=u'inchi')
if len(results) != 1:
print "#results != 1; exiting"
sys.exit(1)
if results[0].molecular_formula == formula:
print 'Formula matches ' + '\033[92m[OK]\033[0m'
else:
print results[0].molecular_formula
print formula
print 'Formulas different ' + '\033[93m[OK]\033[0m'
synonyms=[x.encode('utf-8').lower() for x in results[0].synonyms]
common_names=[]
def get_pubchem_cids(com, user=None):
cid_array = []
for iden in gnomics.objects.auxiliary_files.identifier.filter_identifiers(com.identifiers, ["cid", "pubchem cid", "pubchem compound", "pubchem compound id", "pubchem compound identifier"]):
if iden["identifier"] not in cid_array:
cid_array.append(iden["identifier"])
if cid_array:
return cid_array
ids_completed = []
for iden in gnomics.objects.auxiliary_files.identifier.filter_identifiers(com.identifiers, ["chemspider", "chemspider id", "chemspider identifier", "cs id", "csid"]):
if iden["identifier"] not in ids_completed:
ids_completed.append(iden["identifier"])
results = pubchem.get_compounds(gnomics.objects.compound.Compound.inchi(com, user = user), 'inchi')
for x in results:
gnomics.objects.compound.Compound.add_identifier(com, identifier=x.cid, identifier_type="PubChem CID", language=None, source="PubChem")
cid_array.append(x.cid)
elif user is None:
print("Cannot use ChemSpider conversion when user is None. Please create and pass a valid user with a ChemSpider security token to this method.")
for iden in gnomics.objects.auxiliary_files.identifier.filter_identifiers(com.identifiers, ["kegg compound", "kegg compound id", "kegg compound identifier", "kegg", "kegg compound accession", "kegg id", "kegg identifier", "kegg accession"]):
if iden["identifier"] not in ids_completed:
ids_completed.append(iden["identifier"])
for kegg_com in gnomics.objects.compound.Compound.kegg_compound_db_entry(com):
# Returns PubChem SID.
pubchem_sid = str(kegg_com["DBLINKS"]["PubChem"])
# Get CIDs from SIDs.
server = "https://pubchem.ncbi.nlm.nih.gov/rest/pug"
ext = "/substance/sid/" + str(pubchem_sid) + "/JSONP"
r = requests.get(server+ext, headers={"Content-Type": "application/json"})
if not r.ok:
print("Something went wrong when trying to access the PubChem PUG REST.")
else:
str_r = r.text
try:
l_index = str_r.index("(") + 1
r_index = str_r.rindex(")")
res = str_r[l_index:r_index]
decoded = json.loads(res)
for temp_com in decoded["PC_Substances"][0]["compound"]:
if "id" in temp_com:
if temp_com["id"]["type"] == 1 and "id" in temp_com["id"]:
if "cid" in temp_com["id"]["id"]:
if temp_com["id"]["id"]["cid"] not in cid_array:
cid_array.append(temp_com["id"]["id"]["cid"])
gnomics.objects.compound.Compound.add_identifier(com, identifier=temp_com["id"]["id"]["cid"], identifier_type="PubChem CID", source="KEGG", language=None)
except ValueError:
print("Input is not in a JSONP format.")
for iden in gnomics.objects.auxiliary_files.identifier.filter_identifiers(com.identifiers, ["wikidata", "wikidata accession", "wikidata id", "wikidata identifier"]):
if iden["identifier"] not in ids_completed:
ids_completed.append(iden["identifier"])
for wikidata_object in gnomics.objects.compound.Compound.wikidata(com):
found_array = gnomics.objects.auxiliary_files.wiki.wikidata_property_check(wikidata_object, "pubchem cid", wikidata_property_language = "en")
for x in found_array:
if x not in cid_array:
cid_array.append(x)
gnomics.objects.compound.Compound.add_identifier(com, identifier = x, identifier_type = "PubChem CID", language = None, source = "Wikidata")
return cid_array
import pandas as pd
data = pd.read_csv("compound_names.csv")
#data = data[0:11]
outdf = pd.DataFrame(columns=('Compound Name', 'CID'))
i = 0
for compound in data.values:
for item in get_compounds(compound, 'name'):
#print "Compound Name : "+compound+", "+"CID : "+str(item.cid)
#outdf["Compound Name"] = compound
#outdf["CID"] = item.cid
outdf.loc[i] = [compound, item.cid]
i = i +1
outdf.to_csv('mydb.csv')
def prediction_algorithm2(target_smiles, df):
'''iterates through an entire df instead of a target smiles list'''
start = timeit.default_timer()
TestUni = df
chol = str(target_smiles)
tally = []
for z in range(len(TestUni)):
try:
t = Chem.MolFromSmiles( chol )
s = Chem.MolFromSmiles( TestUni['Substrates'].irow(z) )
p = Chem.MolFromSmiles( TestUni['Products'].irow(z) )
tally.append( scan_atoms(t, s, p) )
if z % 10 == 0 and z != 0:
print z
except:
tally.append( 'do not proceed' )
TestUni['Anabolic Compatible'] = tally
TestUni2 = TestUni[TestUni['Anabolic Compatible'] == 'proceed']
print "Potential hits..."
print len(TestUni2)
anabolic = chol
substrates = []
products = []
targets = []
results = []
enzymes = []
for x in range(len(TestUni2)):
try:
results.append( explore_substrate(anabolic, x, TestUni2) )
substrates.append( TestUni2['Substrates'].irow(x) )
products.append( TestUni2['Products'].irow(x) )
targets.append( anabolic )
enzymes.append( TestUni2['Enzymes'].irow(x) )
except:
pass
stop = timeit.default_timer()
#unique_prods, unique_substrates, unique_products, unique_enzymes = filter_results(results, substrates, products, TestUni, enzymes)
unique_prods, unique_substrates, unique_products, unique_enzymes = results, substrates, products, enzymes
print "Finished reactions... searching hits on PubChem... "
#search pubchem for unique hits
novel_compounds = []
for i in unique_prods:
searches = pcp.get_compounds('CanonicalSMILES', str(Chem.MolToSmiles(i)), 'smiles')
if str(searches) == '[Compound()]':
novel_compounds.append( i )
#put things in a df
unique_prods_smiles = []
novel_smiles = []
for i in unique_prods:
unique_prods_smiles.append( Chem.MolToSmiles(i) )
for i in novel_compounds:
novel_smiles.append( Chem.MolToSmiles(i) )
novels =[ ]
for i in unique_prods_smiles:
if i in novel_smiles:
novels.append( 'Novel' )
else:
novels.append( 'Found in Pubchem' )
Results = pd.DataFrame({'Native Substrate': unique_substrates, 'Native Product': unique_products, 'Products': unique_prods_smiles, 'Novel Compound?': novels, 'Enzymes': unique_enzymes})
Results = Results.drop_duplicates(cols=['Products'])
print 'Novel Compounds found...'
print len(novel_compounds)
print "Runtime..."
print stop - start
product_pictures = Draw.MolsToGridImage(unique_prods,molsPerRow=8, includeAtomNumbers=False)
return Results, product_pictures
def search(query, user=None, search_type=None, source="chemspider", mass_plus_minus=0.001):
result_set = []
if source.lower() in ["chemspider", "all"]:
if user is not None and user.chemspider_security_token is not None:
cs = chemspider(user.chemspider_security_token)
if search_type == None:
for result in cs.search(query):
temp_com = gnomics.objects.compound.Compound(identifier = result.csid, identifier_type = "ChemSpider ID", source = "ChemSpider", name = None)
result_set.append(temp_com)
elif search_type == "formula" or search_type == "molecular formula":
for result in cs.simple_search_by_formula(query):
temp_com = gnomics.objects.compound.Compound(identifier = result.csid, identifier_type = "ChemSpider ID", source = "ChemSpider", name = None)
result_set.append(temp_com)
elif search_type == "mass" and mass_plus_minus is not None:
for result in cs.simple_search_by_mass(query, mass_plus_minus):
temp_com = gnomics.objects.compound.Compound(identifier = result.csid, identifier_type = "ChemSpider ID", source = "ChemSpider", name = None)
result_set.append(temp_com)
else:
print("No valid search type for ChemSpider was provided.")
print("Continuing with search type 'None'...")
return search(query, user = user, search_type = None, source = "chemspider")
elif source.lower() == "chemspider":
print("Searching with ChemSpider requires the creation of a User object with a valid ChemSpider security token. Information on obtaining such a token can be found here: 'http://www.chemspider.com/AboutServices.aspx?'.\n")
print("Continuing with PubChem search...\n")
return search(query, source = "pubchem")
if source.lower() in ["pubchem", "all"]:
if search_type == None:
try:
server = "https://pubchem.ncbi.nlm.nih.gov/rest/pug"
ext = "/compound/name/" + str(query) + "/synonyms/JSONP"
r = requests.get(server+ext, headers={"Content-Type": "application/json"})
if not r.ok:
r.raise_for_status()
print("No results found.")
return result_set
str_r = r.text
try:
l_index = str_r.index("(") + 1
r_index = str_r.rindex(")")
except ValueError:
print("Input is not in a JSONP format.")
exit()
res = str_r[l_index:r_index]
decoded = json.loads(res)
for result in decoded["InformationList"]["Information"]:
result_cid = result["CID"]
temp_com = gnomics.objects.compound.Compound(identifier = result_cid, identifier_type = "PubChem CID", source = "PubChem", name = result["Synonym"][0])
result_set.append(temp_com)
except requests.exceptions.RequestException as e:
print(e)
print("No results found.")
elif search_type == "substructure":
return pubchem.get_compounds(query, "substructure")
elif search_type == "superstructure":
return pubchem.get_compounds(query, "superstructure")
elif search_type == "similarity":
return pubchem.get_compounds(query, "similarity")
elif search_type == "identity":
return pubchem.get_compounds(query, "identity")
elif search_type.lower() == "smiles":
for temp_com in pubchem.get_compounds(query, "smiles"):
if temp_com.synonyms:
new_com = gnomics.objects.compound.Compound(identifier = temp_com.cid, identifier_type = "PubChem CID", source = "PubChem", name = temp_com.synonyms[0])
result_set.append(new_com)
else:
new_com = gnomics.objects.compound.Compound(identifier = temp_com.cid, identifier_type = "PubChem CID", source = "PubChem", name = temp_com.iupac_name)
result_set.append(new_com)
elif search_type.lower() == "inchi":
for temp_com in pubchem.get_compounds(query, "inchi"):
if temp_com.synonyms:
new_com = gnomics.objects.compound.Compound(identifier = temp_com.cid, identifier_type = "PubChem CID", source = "PubChem", name = temp_com.synonyms[0])
result_set.append(new_com)
else:
new_com = gnomics.objects.compound.Compound(identifier = temp_com.cid, identifier_type = "PubChem CID", source = "PubChem", name = temp_com.iupac_name)
result_set.append(new_com)
elif search_type == "sdf":
return pubchem.get_compounds(query, "sdf")
elif search_type == "cid":
return pubchem.get_compounds(query, "cid")
elif search_type == "cas":
server = "https://pubchem.ncbi.nlm.nih.gov/rest/pug"
ext = "/compound/name/" + str(query) + "/synonyms/JSONP"
r = requests.get(server+ext, headers={"Content-Type": "application/json"})
if not r.ok:
print("There was a problem attempting to access the PubChem PUG REST service.")
else:
str_r = r.text
try:
l_index = str_r.index("(") + 1
r_index = str_r.rindex(")")
except ValueError:
print("Input is not in a JSONP format.")
exit()
res = str_r[l_index:r_index]
decoded = json.loads(res)
for result in decoded["InformationList"]["Information"]:
for syn in result["Synonym"]:
if syn == query:
result_cid = result["CID"]
temp_com = gnomics.objects.compound.Compound(identifier = result_cid, identifier_type = "PubChem CID", source = "PubChem", name = result["Synonym"][0])
result_set.append(temp_com)
else:
print("No valid search type for PubChem was provided.")
print("Continuing with search type 'None'...")
return search(query, user = None, search_type = None, source = "pubchem")
if source.lower() != "chemspider" and source.lower() != "pubchem" and source.lower() != "all":
print("No valid search source was provided.")
if user is not None and user.chemspider_security_token is not None:
print("Because user and ChemSpider security token are provided, continuing with ChemSpider search...")
return search(query, user = user, search_type = None, source = "chemspider")
elif user.chemspider_security_token is not None:
print("Because either user not provided or ChemSpider security token is not valid, continuing with PubChem search...")
return search(query, user = None, search_type = None, source = "pubchem")
else:
return result_set
return result_set
