def MolToJSON(mol, size=(300,300), kekulize=True, wedgeBonds=True,
fitImage=False, options=None, **kwargs):
if not mol:
raise ValueError,'Null molecule provided'
canvas = Canvas(size=size)
if options is None:
options = DrawingOptions()
if fitImage:
options.dotsPerAngstrom = int(min(size) / 10)
options.wedgeDashedBonds = wedgeBonds
try:
drawer = MolDrawing(canvas=canvas,drawingOptions=options)
except TypeError:
drawer = MolDrawing(canvas=canvas)
if kekulize:
mol = Chem.Mol(mol.ToBinary())
Chem.Kekulize(mol)
if not mol.GetNumConformers():
Compute2DCoords(mol)
drawer.AddMol(mol,**kwargs)
try:
drawer.AddLegend(kwargs.get('legend', ''))
except AttributeError:
pass
canvas.flush()
return canvas.json
#------------------------------------------------------------------------------
def _parseMolData(data):
"""Imports a molfile and verifies if all of the coordinates are set to zeros.
if they are set to zeros then we know there are no real coordinates in the molfile
In this case we allow RDKit to recaculate the positions of the atoms and come up with its own pictorial representation of the molecule
If not we use the molecule as drawn"""
suppl = SDMolSupplier()
suppl.SetData(str(data), sanitize=False)
data = [x for x in suppl if x]
for x in data:
if not x.HasProp("_drawingBondsWedged"):
SanitizeMol(x)
ctab = MolToMolBlock(x)
ctablines = [
item.split("0.0000") for item in ctab.split("\n")
if "0.0000" in item
]
needs_redraw = 0
for line in ctablines:
if len(line) > 3:
needs_redraw += 1
if needs_redraw == len(ctablines):
#check for overlapping molecules in the CTAB
SanitizeMol(x)
Compute2DCoords(x)
return data
def annotate_chemical_svg(network):
"""Annotate chemical nodes with SVGs depiction.
:param network: dict, network of elements as outputted by the sbml_to_json method
:return: dict, network annotated
"""
from rdkit.Chem import MolFromInchi
from rdkit.Chem.Draw import rdMolDraw2D
from rdkit.Chem.AllChem import Compute2DCoords
from urllib import parse
for node in network['elements']['nodes']:
if node['data']['type'] == 'chemical' and node['data'][
'inchi'] is not None:
inchi = node['data']['inchi']
try:
mol = MolFromInchi(inchi)
# if mol is None:
# raise BaseException('Mol is None')
Compute2DCoords(mol)
drawer = rdMolDraw2D.MolDraw2DSVG(200, 200)
drawer.DrawMolecule(mol)
drawer.FinishDrawing()
svg_draft = drawer.GetDrawingText().replace("svg:", "")
svg = 'data:image/svg+xml;charset=utf-8,' + parse.quote(
svg_draft)
node['data']['svg'] = svg
except BaseException as e:
msg = 'SVG depiction failed from inchi: "{}"'.format(inchi)
logging.warning(msg)
logging.warning("Below the RDKit backtrace...")
logging.warning(e)
node['data']['svg'] = None
return network
def update_mol_image(self, mol_name, save=False, verbosity=0):
""" Generate mol images, if needed
:param str mol_name: name of the molecule to be updated
:param bool save: automatically save data
:param int verbosity: controls verbosity level
"""
import rdkit.Chem
from rdkit.Chem.Draw import MolDraw2DSVG
from rdkit.Chem.AllChem import Compute2DCoords
this_mol = self.ligands_data[mol_name]['molecule']
self.ligands_data[mol_name].setdefault('images', {})
if not {'2d_hs', '2d_nohs'}.issubset(
self.ligands_data[mol_name]['images']):
draw_2d_svg = MolDraw2DSVG(300, 300)
temp_mol = rdkit.Chem.Mol(this_mol)
Compute2DCoords(temp_mol)
draw_2d_svg.drawOptions().addStereoAnnotation = True
draw_2d_svg.DrawMolecule(temp_mol)
draw_2d_svg.FinishDrawing()
svg_data_hs = draw_2d_svg.GetDrawingText()
temp_mol = rdkit.Chem.RemoveHs(temp_mol)
Compute2DCoords(temp_mol)
draw_2d_svg = MolDraw2DSVG(300, 300)
draw_2d_svg.DrawMolecule(temp_mol)
draw_2d_svg.FinishDrawing()
svg_data_no_hs = draw_2d_svg.GetDrawingText()
self.ligands_data[mol_name]['images'].update({
'2d_hs':
svg_data_hs,
'2d_nohs':
svg_data_no_hs
})
if save:
self.save_data()
def toDataFrame(self, fields=None, molAsImage=True, sketch=True):
"""
Returns a pandas.DataFrame of the SmallMolLib object.
Parameters
----------
fields: list
The list of fields to convert into a pandas DataFrame column
molAsImage: bool
If True, the rdkit.Chem.rdchem.Mol is converted into an image
sketch: bool
If True, the molecule are rendered to be 2D
Returns
-------
dataframe: pandas.DataFrame
The pandas DataFrame
"""
from rdkit.Chem import PandasTools
from rdkit.Chem.AllChem import Compute2DCoords
import pandas as pd
from copy import deepcopy
if fields is not None:
if not isinstance(fields, list):
raise TypeError(
'The argument fields passed {} should be a list '.format(
type(fields)))
else:
fields = ['ligname', '_mol'] if molAsImage else ['ligname']
records = []
indexes = []
for i, m in enumerate(self._mols):
row = dict((f, m.__getattribute__(f)) for f in fields)
if sketch:
mm = deepcopy(m._mol)
Compute2DCoords(mm)
row['_mol'] = mm
records.append(row)
indexes.append(i)
df = pd.DataFrame(records, columns=fields, index=indexes)
if molAsImage:
Chem.PandasTools.ChangeMoleculeRendering(df)
return df
def testMakePNG(self):
# smoke test only--that it doesn't fail, not that it looks correct (that's outside the scope of this package)
# The choices shown resulted (at last check) in 3 fragments, one of which has a branch
try:
silent_remove(TEST_PNG)
result = create_sample_kmc_result(num_initial_monos=24, max_monos=24, seed=1, max_time=SHORT_TIME)
summary = analyze_adj_matrix(result[ADJ_MATRIX])
adj_analysis_to_stdout(summary, break_co_bonds=False)
nodes = result[MONO_LIST]
adj = result[ADJ_MATRIX]
block = generate_mol(adj, nodes)
mol = MolFromMolBlock(block)
Compute2DCoords(mol)
MolToFile(mol, TEST_PNG, size=(2000, 1200))
self.assertTrue(os.path.isfile(TEST_PNG))
finally:
silent_remove(TEST_PNG, disable=DISABLE_REMOVE)
pass
def display_smiles(smiles_list,colour_type,params,molSize=(450,150)):
for smile in smiles_list:
highlight =[]
colors={}
highlightBonds=[]
highlightAtomRadii ={}
bond_colors={}
mol = smile_to_mol(smile, params)
mol = Chem.RemoveHs(mol,sanitize = False)
if(colour_type=='t'):
color = (1,1,0.925)
elif(colour_type=='g'):
color = (0.925,1,0.925)
elif(colour_type=='w'):
color = (1,1,1)
else:
color = (0.925,0.925,1)
for i in range(mol.GetNumAtoms()):
highlight.append(i)
colors[i] = color
highlightAtomRadii[i] =1000000
for i in range(mol.GetNumBonds()):
highlightBonds.append(i)
bond_colors[i] = color
Compute2DCoords(mol)
drawer = rdMolDraw2D.MolDraw2DCairo(molSize[0],molSize[1])
drawer.DrawMolecule(mol,highlightAtoms=highlight,highlightAtomColors=colors,highlightBonds=highlightBonds,highlightBondColors=bond_colors,highlightAtomRadii=highlightAtomRadii)
drawer.FinishDrawing()
fp = BytesIO()
#with open(fp,'wb') as f:
fp.write(drawer.GetDrawingText())
img_top = PIL.Image.open(fp)
display(img_top)
def get_batch_from_xls_row(index, row, structure_col, project):
if structure_col:
smiles_str = row[structure_col]
try:
struc = Chem.MolFromSmiles(smiles_str)
if struc:
Compute2DCoords(struc)
b = CBHCompoundBatch.objects.from_rd_mol(struc,
smiles=smiles_str,
project=project,
reDraw=True)
b.blinded_batch_id = None
else:
raise Exception("Smiles not processed")
except Exception, e:
b = CBHCompoundBatch.objects.blinded(project=project)
b.original_smiles = smiles_str
b.warnings["parseerror"] = "true"
b.properties["action"] = "Ignore"
def produce_output(adj_matrix, mono_list, cfg):
if cfg[SUPPRESS_SMI] and not (cfg[SAVE_JSON] or cfg[SAVE_PNG] or cfg[SAVE_SVG]):
format_list = [SAVE_TCL]
mol = None # Make IDE happy
else:
# Default out is SMILES, which requires getting an rdKit molecule object; also required for everything
# except the TCL format
format_list = [SAVE_TCL, SAVE_JSON, SAVE_PNG, SAVE_SVG]
block = generate_mol(adj_matrix, mono_list)
mol = MolFromMolBlock(block)
try:
smi_str = MolToSmiles(mol) + '\n'
except:
raise InvalidDataError("Error in producing SMILES string.")
# if SMI is to be saved, don't output to stdout
if cfg[SAVE_SMI]:
fname = create_out_fname(cfg[BASENAME], base_dir=cfg[OUT_DIR], ext=SAVE_SMI)
str_to_file(smi_str, fname, print_info=True)
else:
print("\nSMILES representation: \n", MolToSmiles(mol), "\n")
if cfg[SAVE_PNG] or cfg[SAVE_SVG] or cfg[SAVE_JSON]:
# PNG and SVG make 2D images and thus need coordinates
# JSON will save coordinates--zero's if not computed; might as well compute and save non-zero values
Compute2DCoords(mol)
for save_format in format_list:
if cfg[save_format]:
fname = create_out_fname(cfg[BASENAME], base_dir=cfg[OUT_DIR], ext=save_format)
if save_format == SAVE_TCL:
gen_tcl(adj_matrix, mono_list, tcl_fname=fname, chain_id=cfg[CHAIN_ID],
psf_fname=cfg[PSF_FNAME], toppar_dir=cfg[TOPPAR_DIR], out_dir=cfg[OUT_DIR])
if save_format == SAVE_JSON:
json_str = MolToJSON(mol)
str_to_file(json_str + '\n', fname)
elif save_format == SAVE_PNG or save_format == SAVE_SVG:
MolToFile(mol, fname, size=cfg[IMAGE_SIZE])
print(f"Wrote file: {fname}")
def depict(self, sketch=True, filename=None, ipython=False, optimize=False, optimizemode='std', removeHs=True,
atomlabels=None, highlightAtoms=None, resolution=(400, 200)):
"""
Depicts the molecules. It is possible to save it into an svg file and also generates a jupiter-notebook rendering
Parameters
----------
sketch: bool
Set to True for 2D depiction
filename: str
Set the filename for the svg file
ipython: bool
Set to True to return the jupiter-notebook rendering
optimize: bool
Set to True to optimize the conformation. Works only with 3D.
optimizemode: ['std', 'mmff']
Set the optimization mode for 3D conformation
removeHs: bool
Set to True to hide hydrogens in the depiction
atomlabels: str
Accept any combinations of the following pararemters as unique string '%a%i%c%*' a:atom name, i:atom index,
c:atom formal charge (+/-), *:chiral (* if atom is chiral)
highlightAtoms: list
List of atom to highlight. It can be also a list of atom list, in this case different colors will be used
resolution: tuple of integers
Resolution in pixels: (X, Y)
Returns
-------
ipython_svg: SVG object if ipython is set to True
Example
-------
>>> sm.depict(ipython=True, optimize=True, optimizemode='std') # doctest: +SKIP
>>> sm.depict(ipython=True, sketch=True) # doctest: +SKIP
>>> sm.depict(ipython=True, sketch=True) # doctest: +SKIP
>>> sm.depict(ipython=True, sketch=True, atomlabels="%a%i%c") # doctest: +SKIP
>>> ids = np.intersect1d(sm.get('idx', 'hybridization SP2'), sm.get('idx', 'element C')) # doctest: +SKIP
>>> sm.depict(ipython=True, sketch=True,highlightAtoms=ids.tolist(), removeHs=False) # doctest: +SKIP
"""
from rdkit import Chem
from rdkit.Chem.AllChem import Compute2DCoords, EmbedMolecule, MMFFOptimizeMolecule, ETKDG
from copy import deepcopy
if sketch and optimize:
raise ValueError('Impossible to use optimization in 2D sketch representation')
if optimizemode not in ['std', 'mmff']:
raise ValueError('Optimization mode {} not understood. Can be "std" or "ff"'.format(optimizemode))
elements = self._element
indexes = self._idx
formalcharges = self._formalcharge
chirals = self._chiral
_mol = deepcopy(self._mol)
if sketch:
Compute2DCoords(_mol)
if removeHs:
_mol = Chem.RemoveHs(_mol)
elements = self.get('element', 'element H', invert=True)
indexes = self.get('idx', 'element H', invert=True)
formalcharges = self.get('formalcharge', 'element H', invert=True)
chirals = self.get('chiral', 'element H', invert=True)
_labelsFunc = ['a', 'i', 'c', '*']
if atomlabels is not None:
labels = atomlabels.split('%')[1:]
formalcharges = ['' if c == 0 else "+" if c == 1 else "-" for c in formalcharges]
chirals = ['' if c == '' else '*' for c in chirals]
values = [elements, indexes, formalcharges, chirals]
idxs = [_labelsFunc.index(l) for l in labels]
labels_required = [values[i] for i in idxs]
atomlabels = ["".join([str(i) for i in a]) for a in list(zip(*labels_required))]
if optimize:
if optimizemode == 'std':
EmbedMolecule(_mol, ETKDG())
elif optimizemode == 'mmff':
MMFFOptimizeMolecule(_mol)
return _depictMol(_mol, filename=filename, ipython=ipython, atomlabels=atomlabels,
highlightAtoms=highlightAtoms, resolution=resolution)
def _computeCoords(mol, force=False):
if force or (not mol.GetNumConformers() or mol.GetConformer().Is3D()):
Compute2DCoords(mol)
return mol
def post_validate_list(self, request, **kwargs):
"""
When a list of SMILES patterns are submitted, save them to elasticsearhc and
call the normal validate multi batch function to give the normal statistics page
"""
session_key = request.COOKIES[settings.SESSION_COOKIE_NAME]
deserialized = self.deserialize(request,
request.body,
format=request.META.get(
'CONTENT_TYPE',
'application/json'))
deserialized = self.alter_deserialized_detail_data(
request, deserialized)
bundle = self.build_bundle(data=dict_strip_unicode_keys(deserialized),
request=request)
if bundle.obj.pk:
self.authorized_update_detail(self.get_object_list(bundle.request),
bundle)
else:
self.authorized_create_detail(self.get_object_list(bundle.request),
bundle)
smilesdata = bundle.data.get("smilesdata", "")
objects = [smi.strip() for smi in smilesdata.splitlines(True)]
batches = []
# first assume smiles
allmols = [(obj, Chem.MolFromSmiles(str(obj))) for obj in objects]
# Next test for inchi
multiple_batch = CBHCompoundMultipleBatch.objects.create(
project=bundle.data["project"], batch_count=len(allmols))
for index, m in enumerate(allmols):
if m[1] is None:
inchimol = Chem.MolFromInchi(
str(m[0].encode('ascii', 'ignore')))
if inchimol is not None:
allmols[index] = (Chem.MolToSmiles(inchimol), inchimol)
for m in allmols:
if (m[1]):
Compute2DCoords(m[1])
batches = []
for mol2 in allmols:
if mol2[1]:
b = CBHCompoundBatch.objects.from_rd_mol(
mol2[1],
smiles=mol2[0],
project=bundle.data["project"],
reDraw=True)
else:
b = CBHCompoundBatch.objects.blinded(
project=bundle.data["project"])
b.warnings["smilesParseError"] = "true"
b.properties["action"] = "Ignore"
b.original_smiles = mol2[0]
b.multiple_batch_id = multiple_batch.pk
b.created_by = bundle.request.user.username
b.created_by_id = bundle.request.user.id
batches.append(b)
bundle.data["current_batch"] = multiple_batch.pk
bundle.data["headers"] = []
validate_multi_batch(self, multiple_batch, bundle.data, session_key,
batches)
return self.create_response(request,
bundle,
response_class=http.HttpAccepted)
def display_encoded_decoded(params,tanimoto_values,encode_decode_mols,num_F2=100,pic_size=(1190/2,1684/2),start_num=1):
text_size =12
text_font_size = 12
orange = (255/255,165/255,0)
green = (0,1,0)
light_blue = (0,1,1)
png_files =[]
w_len = 1
'''
stop_while = True
while stop_while:
h_len_min = int(num_F2/w_len)+int(np.mod(num_F2,w_len)>0)
h_temp = pic_size[0]/w_len*h_len_min
if(h_temp>pic_size[1]) :
w_len = w_len+1
else:
stop_while = False
'''
w_len = 5
h_len_min = int(num_F2/w_len)+int(np.mod(num_F2,w_len)>0)
h_len = h_len_min
molSize= (int(pic_size[0]/w_len)*2,int(pic_size[0]/w_len))
x = np.array(range(len(libExamples)))
#random.shuffle(x)
for mol_idx1 in range(h_len*w_len):
if( mol_idx1 < num_F2):
#mol = Chem.RemoveHs(mol,sanitize = False)
mol = Chem.RemoveHs(encode_decode_mols[mol_idx1][0],sanitize = False)
Compute2DCoords(mol)
drawer = rdMolDraw2D.MolDraw2DCairo(molSize[0],int(molSize[1]/2)*2)
drawer.DrawMolecule(mol)
drawer.FinishDrawing()
fp = BytesIO()
#with open(fp,'wb') as f:
fp.write(drawer.GetDrawingText())
img_top = PIL.Image.open(fp)
mol = Chem.RemoveHs(encode_decode_mols[mol_idx1][1],sanitize = False)
Compute2DCoords(mol)
drawer = rdMolDraw2D.MolDraw2DCairo(molSize[0],int(molSize[1]/2)*2)
drawer.DrawMolecule(mol)
drawer.FinishDrawing()
fp = BytesIO()
#with open(fp,'wb') as f:
fp.write(drawer.GetDrawingText())
img_bottom = PIL.Image.open(fp)
text_works = Image.new('RGB', (molSize[0],text_size), color = (255, 255, 255))
d = ImageDraw.Draw(text_works)
d.text((0,0),str(mol_idx1+start_num), font=ImageFont.truetype("arialbd.ttf", text_font_size),fill=(0,0,0))
d.text((0,0),' Tanimoto '+str(int(tanimoto_values[mol_idx1]*100)/100), font=ImageFont.truetype("arial.ttf", text_font_size),fill=(0,0,0))
comb_img = vstack_img(text_works,vstack_img(img_top,img_bottom))
fp1 = BytesIO()
comb_img.save(fp1, 'PNG')
png_files.append(fp1)
#display(comb_img)
#print(comb_img.size)
'''
img, svg = gen_img_svg_mol(mol,molSize=(500,250))
fp = BytesIO()
display(img)
print(libExamples['smiles'][mol_idx])
svg2png(bytestring=svg,write_to=fp,scale=1)
text_selected_frags_flash = Image.new('RGB', (half_pic_size[0],single_space*2), color = (255, 255, 255))
d = ImageDraw.Draw(text_selected_frags_flash)
d.text((0,0), "Bag of sampled first order fragments", font=ImageFont.truetype("arial.ttf", font_size),fill=(0,0,0))
d.text((0,single_space), "Status: Sampling node", font=ImageFont.truetype("arial.ttf", font_size),fill=(0,0,0))
'''
else:
fp2 = BytesIO()
png_files.append(fp2)
#print((int(molSize[1]/2.0)*4))
image = Image.new('RGB', (molSize[0],(int(molSize[1]/2.0)*4)+text_size), color = 'white')
draw = ImageDraw.Draw(image)
image.save(fp2, 'PNG')
#print(image.size)
imgs = [ PIL.Image.open(i) for i in png_files ]
imgs_comb = []
for y_idx in range(0,h_len):
y_idx_list = png_files[y_idx*w_len:w_len+y_idx*w_len]
imgs_comb.append( np.hstack( (np.asarray(PIL.Image.open(i) ) for i in y_idx_list ) ))
imgs_comb = np.vstack(imgs_comb)
imgs_comb = PIL.Image.fromarray( imgs_comb)
imgs_comb_size = imgs_comb.size
'''
if(imgs_comb_size[0]!=pic_size[0]):
fp = BytesIO()
image = Image.new('RGB', (pic_size[0]-imgs_comb_size[0],imgs_comb_size[1]), color = 'white')
draw = ImageDraw.Draw(image)
image.save(fp, 'PNG')
imgs_comb=( np.hstack( (np.asarray(i) for i in [imgs_comb,PIL.Image.open(fp)] ) ))
imgs_comb = PIL.Image.fromarray( imgs_comb)
imgs_comb_size = imgs_comb.size
'''
'''
if(imgs_comb_size[1]!=pic_size[1]):
fp = BytesIO()
image = Image.new('RGB', (pic_size[0],pic_size[1]-imgs_comb_size[1]), color = 'white')
draw = ImageDraw.Draw(image)
image.save(fp, 'PNG')
imgs_comb=( np.vstack( (np.asarray(i) for i in [imgs_comb,PIL.Image.open(fp)] ) ))
imgs_comb = PIL.Image.fromarray( imgs_comb)
'''
fp = BytesIO()
imgs_comb.save( fp,'PNG' )
return imgs_comb
def display_training_data(params,num_F2=80,pic_size=(int(1190*0.9),int(1684*0.9)),train=True):
text_size =20
text_font_size = 20
if(train==True):
libExamples = pd.read_csv(params['model_dir']+'Experimental_Training_set_NN.csv')
libExamples = libExamples.reset_index(drop=True)
else:
libExamples = pd.read_csv(params['model_dir']+'Experimental_Test_NN.csv')
libExamples = libExamples.reset_index(drop=True)
orange = (255/255,165/255,0)
green = (0,1,0)
light_blue = (0,1,1)
png_files =[]
w_len = 1
stop_while = True
while stop_while:
h_len_min = int(num_F2/w_len)+int(np.mod(num_F2,w_len)>0)
h_temp = pic_size[0]/w_len*h_len_min
if(h_temp>pic_size[1]) :
w_len = w_len+1
else:
stop_while = False
h_len = h_len_min
molSize= (int(pic_size[0]/w_len),int(pic_size[0]/w_len))
x = np.array(range(len(libExamples)))
#random.shuffle(x)
for mol_idx1 in range(h_len*w_len):
if( mol_idx1 < num_F2):
mol_idx = x[mol_idx1]
highlight =[]
colors={}
highlightBonds=[]
highlightAtomRadii ={}
bond_colors={}
mol =smile_to_mol(libExamples['smiles'][mol_idx], params)
mol = Chem.RemoveHs(mol,sanitize = False)
if(libExamples['metrics'][mol_idx]==1):
for i in range(mol.GetNumAtoms()):
highlight.append(i)
if(libExamples['metrics'][mol_idx]==1):
colors[i] = (0.75,1,0.75)
else:
colors[i] = (1,0.75,0.75)
highlightAtomRadii[i] =1000000
for i in range(mol.GetNumBonds()):
highlightBonds.append(i)
if(libExamples['metrics'][mol_idx]==1):
bond_colors[i] = (0.75,1,0.75)
else:
bond_colors[i] = (1,0.75,0.75)
Compute2DCoords(mol)
drawer = rdMolDraw2D.MolDraw2DCairo(molSize[0],molSize[1])
drawer.DrawMolecule(mol,highlightAtoms=highlight,highlightAtomColors=colors,highlightBonds=highlightBonds,highlightBondColors=bond_colors,highlightAtomRadii=highlightAtomRadii)
drawer.FinishDrawing()
fp = BytesIO()
#with open(fp,'wb') as f:
fp.write(drawer.GetDrawingText())
img_top = PIL.Image.open(fp)
if(libExamples['metrics'][mol_idx]==1):
text_works = Image.new('RGB', (molSize[0],text_size), color = (191,255,191))
else:
text_works = Image.new('RGB', (molSize[0],text_size), color = (255, 255, 255))
d = ImageDraw.Draw(text_works)
if((libExamples['ECFP'][mol_idx]==0 and libExamples['metrics'][mol_idx]==1)or(libExamples['ECFP'][mol_idx]==1 and libExamples['metrics'][mol_idx]==0)):
symbol = 'E'
else:
symbol = ""
d.text((int(molSize[0]/2)+text_size*0.5,0), symbol, font=ImageFont.truetype("arial.ttf", text_font_size),fill=(0,0,255))
if((libExamples['ChemVAE'][mol_idx]==0 and libExamples['metrics'][mol_idx]==1)or(libExamples['ChemVAE'][mol_idx]==1 and libExamples['metrics'][mol_idx]==0)):
symbol = 'C'
else:
symbol = ""
d.text((int(molSize[0]/2)+text_size*1.5,0), symbol, font=ImageFont.truetype("arial.ttf", text_font_size),fill=(0,0,255))
if((libExamples['rnd_FragVAE'][mol_idx]==0 and libExamples['metrics'][mol_idx]==1)or(libExamples['rnd_FragVAE'][mol_idx]==1 and libExamples['metrics'][mol_idx]==0)):
symbol = 'R'
else:
symbol = ""
d.text((int(molSize[0]/2)-text_size*0.5,0), symbol, font=ImageFont.truetype("arial.ttf", text_font_size),fill=(0,0,255))
if((libExamples['FragVAE'][mol_idx]==0 and libExamples['metrics'][mol_idx]==1)or(libExamples['FragVAE'][mol_idx]==1 and libExamples['metrics'][mol_idx]==0)):
symbol = 'F'
else:
symbol = ""
d.text((int(molSize[0]/2)-text_size*1.5,0),symbol, font=ImageFont.truetype("arial.ttf", text_font_size),fill=(0,0,255))
d.text((0,0),str(mol_idx1+1), font=ImageFont.truetype("arialbd.ttf", text_font_size),fill=(0,0,0))
comb_img = vstack_img(img_top,text_works)
fp1 = BytesIO()
comb_img.save(fp1, 'PNG')
png_files.append(fp1)
'''
img, svg = gen_img_svg_mol(mol,molSize=(500,250))
fp = BytesIO()
display(img)
print(libExamples['smiles'][mol_idx])
svg2png(bytestring=svg,write_to=fp,scale=1)
text_selected_frags_flash = Image.new('RGB', (half_pic_size[0],single_space*2), color = (255, 255, 255))
d = ImageDraw.Draw(text_selected_frags_flash)
d.text((0,0), "Bag of sampled first order fragments", font=ImageFont.truetype("arial.ttf", font_size),fill=(0,0,0))
d.text((0,single_space), "Status: Sampling node", font=ImageFont.truetype("arial.ttf", font_size),fill=(0,0,0))
'''
else:
fp2 = BytesIO()
png_files.append(fp2)
image = Image.new('RGB', (molSize[0],molSize[1]+text_size), color = 'white')
draw = ImageDraw.Draw(image)
image.save(fp2, 'PNG')
imgs = [ PIL.Image.open(i) for i in png_files ]
imgs_comb = []
for y_idx in range(0,h_len):
y_idx_list = png_files[y_idx*w_len:w_len+y_idx*w_len]
imgs_comb.append( np.hstack( (np.asarray(PIL.Image.open(i) ) for i in y_idx_list ) ))
imgs_comb = np.vstack(imgs_comb)
imgs_comb = PIL.Image.fromarray( imgs_comb)
imgs_comb_size = imgs_comb.size
if(imgs_comb_size[0]!=pic_size[0]):
fp = BytesIO()
image = Image.new('RGB', (pic_size[0]-imgs_comb_size[0],imgs_comb_size[1]), color = 'white')
draw = ImageDraw.Draw(image)
image.save(fp, 'PNG')
imgs_comb=( np.hstack( (np.asarray(i) for i in [imgs_comb,PIL.Image.open(fp)] ) ))
imgs_comb = PIL.Image.fromarray( imgs_comb)
imgs_comb_size = imgs_comb.size
'''
if(imgs_comb_size[1]!=pic_size[1]):
fp = BytesIO()
image = Image.new('RGB', (pic_size[0],pic_size[1]-imgs_comb_size[1]), color = 'white')
draw = ImageDraw.Draw(image)
image.save(fp, 'PNG')
imgs_comb=( np.vstack( (np.asarray(i) for i in [imgs_comb,PIL.Image.open(fp)] ) ))
imgs_comb = PIL.Image.fromarray( imgs_comb)
'''
fp = BytesIO()
imgs_comb.save( fp,'PNG' )
return imgs_comb
def depict(
self,
ids=None,
sketch=True,
filename=None,
ipython=False,
optimize=False,
optimizemode="std",
removeHs=True,
legends=None,
highlightAtoms=None,
mols_perrow=3,
):
"""
Depicts the molecules into a grid. It is possible to save it into an svg file and also generates a
jupiter-notebook rendering
Parameters
----------
ids: list
The index of the molecules to depict
sketch: bool
Set to True for 2D depiction
filename: str
Set the filename for the svg file
ipython: bool
Set to True to return the jupiter-notebook rendering
optimize: bool
Set to True to optimize the conformation. Works only with 3D.
optimizemode: ['std', 'mmff']
Set the optimization mode for 3D conformation
removeHs: bool
Set to True to hide hydrogens in the depiction
legends: str
The name to used for each molecule. Can be 'names':the name of themselves; or 'items': a incremental id
highlightAtoms: list
A List of atom to highligh for each molecule. It can be also a list of atom list, in this case different
colors will be used
mols_perrow: int
The number of molecules to depict per row of the grid
Returns
-------
ipython_svg: SVG object if ipython is set to True
"""
from rdkit.Chem.AllChem import (
Compute2DCoords,
EmbedMolecule,
MMFFOptimizeMolecule,
ETKDG,
)
from rdkit.Chem import RemoveHs
from moleculekit.smallmol.util import depictMultipleMols
if sketch and optimize:
raise ValueError(
"Impossible to use optmization in 2D sketch representation")
if legends is not None and legends not in ["names", "items"]:
raise ValueError('The "legends" should be "names" or "items"')
_smallmols = self.getMols(ids)
if ids is None:
_mols = [m._mol for m in self._mols]
else:
_mols = [m._mol for m in self.getMols(ids)]
if highlightAtoms is not None:
if len(highlightAtoms) != len(_mols):
raise ValueError(
"The highlightAtoms {} should have the same length of the "
"mols {}".format(len(highlightAtoms), len(_mols)))
if sketch:
for _m in _mols:
Compute2DCoords(_m)
if removeHs:
_mols = [RemoveHs(_m) for _m in _mols]
# activate 3D coords optimization
if optimize:
if optimizemode == "std":
for _m in _mols:
EmbedMolecule(_m)
elif optimizemode == "mmff":
for _m in _mols:
MMFFOptimizeMolecule(_m, ETKDG())
legends_list = []
if legends == "names":
legends_list = [_m.getProp("ligname") for _m in _smallmols]
elif legends == "items":
legends_list = [str(n + 1) for n in range(len(_smallmols))]
return depictMultipleMols(
_mols,
ipython=ipython,
legends=legends_list,
highlightAtoms=highlightAtoms,
filename=filename,
mols_perrow=mols_perrow,
)
def update_pertubation_image(self,
mol_a_name,
mol_b_name,
core_smarts=None,
save=False,
verbosity=0,
**kwargs):
""" Generate mol images describing a pertubation between the ligand pair
:param str mol_a_name: name of the molecule A
:param str mol_b_name: name of the molecule B
:param str core_smarts: use this smarts as common core
:param bool save: automatically save data
:param int verbosity: controls verbosity level
"""
# verbosity = 5
self.ligands_data[mol_a_name].setdefault('images', {})
self.ligands_data[mol_a_name]['images'].setdefault('perturbations', {})
self.ligands_data[mol_b_name].setdefault('images', {})
self.ligands_data[mol_b_name]['images'].setdefault('perturbations', {})
import rdkit.Chem
this_mol_a = rdkit.Chem.Mol(self.ligands_data[mol_a_name]['molecule'])
this_mol_b = rdkit.Chem.Mol(self.ligands_data[mol_b_name]['molecule'])
if core_smarts is None:
# Get core_smarts using find_mcs
from merge_topologies import find_mcs
this_mol_a.RemoveAllConformers()
this_mol_b.RemoveAllConformers()
core_smarts = find_mcs([this_mol_a, this_mol_b],
savestate=self,
verbosity=verbosity,
**kwargs).smartsString
try:
# Test whether the correct data structure is already present
assert len(self.ligands_data[mol_a_name]['images']['perturbations']
[mol_b_name][core_smarts]) > 0
assert len(self.ligands_data[mol_b_name]['images']['perturbations']
[mol_a_name][core_smarts]) > 0
except (KeyError, AssertionError):
# It isn't, go on and create the images
os_util.local_print(
'Perturbation images for molecules {} and {} with common core "{}" were not found. '
'Generating it.'.format(mol_a_name, mol_b_name, core_smarts),
msg_verbosity=os_util.verbosity_level.debug,
current_verbosity=verbosity)
else:
return None
from rdkit.Chem.Draw import MolDraw2DSVG
from rdkit.Chem.AllChem import Compute2DCoords, GenerateDepictionMatching2DStructure
core_mol = rdkit.Chem.MolFromSmarts(core_smarts)
Compute2DCoords(core_mol)
for each_name, each_mol, other_mol in zip([mol_a_name, mol_b_name],
[this_mol_a, this_mol_b],
[mol_b_name, mol_a_name]):
GenerateDepictionMatching2DStructure(each_mol,
core_mol,
acceptFailure=True)
# Draw mol with hydrogens
draw_2d_svg = MolDraw2DSVG(300, 150)
draw_2d_svg.drawOptions().addStereoAnnotation = True
not_common_atoms = [
i.GetIdx() for i in each_mol.GetAtoms()
if i.GetIdx() not in each_mol.GetSubstructMatch(core_mol)
]
draw_2d_svg.DrawMolecule(each_mol,
legend=each_name,
highlightAtoms=not_common_atoms)
draw_2d_svg.FinishDrawing()
svg_data_hs = draw_2d_svg.GetDrawingText()
# Draw mol without hydrogens
draw_2d_svg = MolDraw2DSVG(300, 150)
draw_2d_svg.drawOptions().addStereoAnnotation = True
each_mol = rdkit.Chem.RemoveHs(each_mol)
not_common_atoms = [
i.GetIdx() for i in each_mol.GetAtoms()
if i.GetIdx() not in each_mol.GetSubstructMatch(
rdkit.Chem.RemoveHs(core_mol))
]
draw_2d_svg.DrawMolecule(each_mol,
legend=each_name,
highlightAtoms=not_common_atoms)
draw_2d_svg.FinishDrawing()
svg_data_nohs = draw_2d_svg.GetDrawingText()
perturbation_imgs = self.ligands_data[each_name]['images'][
'perturbations']
perturbation_imgs.setdefault(other_mol, {})[core_smarts] = {
'2d_hs': svg_data_hs,
'2d_nohs': svg_data_nohs
}
if save:
self.save_data()