def test_isomorphism(bgf_file1, bgf_file2, simple, silent=True):
"""
def test_isomorphism():
Function Parameters:
bgf_file A string of filename or BgfFile class.
"""
# initialize
pei1 = 0
pei2 = 0
# open BGF
if isinstance(bgf_file1, bgf.BgfFile):
myBGF = bgf_file1
else:
if not silent: print("opening bgf file 1.. " + str(bgf_file1))
myBGF = bgf.BgfFile(bgf_file1)
if isinstance(bgf_file2, bgf.BgfFile):
myBGF = bgf_file2
else:
if not silent: print("opening bgf file 2.. " + str(bgf_file2))
myBGF = bgf.BgfFile(bgf_file2)
if simple:
# reduce hydrogens
pei1 = bgftools.getBackbone(bgf_file1)
pei2 = bgftools.getBackbone(bgf_file2)
else:
pei1 = bgf_file1
pei2 = bgf_file2
# convert connection into dictionary
d_graph1 = bgftools.getConnectionDict(pei1)
d_graph2 = bgftools.getConnectionDict(pei2)
if not silent: print("structures are converted to graph.")
G1 = nx.Graph(d_graph1)
G2 = nx.Graph(d_graph2)
if not silent: print("graphs are loaded.")
# check isomorphism
if not silent: print("now comparing two graphs..")
GM = isomorphism.GraphMatcher(G1, G2)
result = GM.is_isomorphic()
if not silent:
print(result)
else:
return result
def calculate_wiener_index(body):
"""
Calculates Wiener index of the generated polymer.
:BgfAtom body
:returns int(index): Wiener index
Method from from PEI_calcWienerIndex_parallel.py
"""
myBGF = copy.deepcopy(body)
# remove hydrogens
pei = bgftools.getBackbone(myBGF, 0)
# convert connection into dictionary
d_graph = bgftools.getConnectionDict(pei)
# convert dictionary into Graph
G = nx.Graph(d_graph)
# calculate the all shortest length path
d_dist = nx.all_pairs_shortest_path_length(G)
# get Wiener Index
index = 0
for key in d_dist.iterkeys():
for key2 in d_dist[key].iterkeys():
index += d_dist[key][key2]
index = index / 2.0
return int(index)
def do_work(bgf_file, out_file, simple, silent=True):
#initialize
myPEI = bgf.BgfFile()
n_H = []
# open BGF
myBGF = bgf.BgfFile(bgf_file)
# remove hydrogens
pei = bgftools.getBackbone(myBGF, 0)
# convert connection into dictionary
d_graph = bgftools.getConnectionDict(pei)
# convert dictionary into Graph
G = nx.Graph(d_graph)
# calculate the all shortest length path
#d_dist = nx.floyd_warshall(G) # seems a problem
d_dist = nx.all_pairs_shortest_path_length(G)
# get Wiener Index
index = 0
for key in d_dist.iterkeys():
for key2 in d_dist[key].iterkeys():
index += d_dist[key][key2]
index = index / 2.0
return index
def check_isomorphism(directory, bgf_file, simple):
"""
check isomorphism of 'bgf_file' in all files in 'directory'
"""
#initialize
structure_dir = os.path.abspath(directory)
curr_dir = os.path.abspath(".")
pei_file = glob.glob(structure_dir + "/*.bgf")
for i in pei_file:
if bgf_file in i:
pei_file.remove(i) # remove self bgf_file from pei_file
pei_file.sort()
n_pei_file = len(pei_file)
print("The script will compare " + str(n_pei_file) + " files in the directory " + curr_dir)
joblist = []; # contains file pairlists
n_joblist = 0; # number of total jobs
count = 0; # job counter
t1 = time.time(); t2 = 0; # time progess
print("Queueing Jobs..")
print(n_pei_file)
if n_pei_file == 0:
return True;
for i in range(0, n_pei_file):
joblist.append([bgf_file, pei_file[i]]);
n_joblist = len(joblist)
myBGF1 = bgf.BgfFile(bgf_file)
if simple:
# remove hydrogens only in carbon atoms
pei1 = bgftools.getBackbone(bgf_file1)
else:
pei1 = bgf_file1
for job in joblist:
# initialize
bgf_file2 = job[1]; # filenames
pei1 = 0; pei2 = 0; myBGF2 = 0; n_H = [];
t2 = time.time()
elapsed = t2 - t1
estimated = elapsed * (n_joblist - count)
# open BGF
myBGF2 = bgf.BgfFile(bgf_file2)
if simple:
# remove hydrogens only in carbon atoms
pei2 = bgftools.getBackbone(bgf_file2)
else:
# toss whole structures
pei2 = bgf_file2
# convert connection into dictionary
d_graph1 = bgftools.getConnectionDict(pei1)
d_graph2 = bgftools.getConnectionDict(pei2)
# convert dictionary into Graph
G1 = nx.Graph(d_graph1)
G2 = nx.Graph(d_graph2)
# check isomorphism
GM = isomorphism.GraphMatcher(G1, G2)
result = GM.is_isomorphic()
if not result:
return bgf_file2;
#print bgf_file2
# count
count += 1;
# display the process
#sys.stdout.write("\rProgress: " + "{0:>8d}".format(count) + " / " + str(n_joblist) + " (" + str(estimated) + " sec left)"); sys.stdout.flush()
# completing
return True;
def do_work(directory, out_file, simple, silent=True):
#initialize
structure_dir = os.path.abspath(directory)
curr_dir = os.path.abspath(".")
pei_file = glob.glob(structure_dir + "/*.bgf")
pei_file.sort()
n_pei_file = len(pei_file)
f_out_file = open(out_file, 'w') # file for result
joblist = []
# contains filelists
n_joblist = 0
# number of total jobs
count = 0
# job counter
l_index = []
# for index
f_out_file.write(str(sys.argv[0]) + " version " + str(version) + "\n")
f_out_file.write("" + "\n")
f_out_file.write("Job started at " + time.asctime(time.gmtime()) + " on " +
os.environ["HOSTNAME"] + " by " + os.environ["USER"] +
"\n")
f_out_file.write("Command executed at " + os.getcwd() + "\n")
f_out_file.write("Requested options: " + str(sys.argv) + "\n")
f_out_file.write("" + "\n")
if not silent: print("Queueing Jobs..")
for i in range(0, n_pei_file):
joblist.append(pei_file[i])
n_joblist = len(joblist)
for job in joblist:
# initialize
bgf_file = job
myPEI = bgf.BgfFile()
n_H = []
# open BGF
myBGF = bgf.BgfFile(bgf_file)
# remove hydrogens
pei = bgftools.getBackbone(myBGF, 0)
# convert connection into dictionary
d_graph = bgftools.getConnectionDict(pei)
# convert dictionary into Graph
G = nx.Graph(d_graph)
# calculate the all shortest length path
#d_dist = nx.floyd_warshall(G) # seems a problem
d_dist = nx.all_pairs_shortest_path_length(G)
# count
count += 1
# get Wiener Index
index = 0
for key in d_dist.iterkeys():
for key2 in d_dist[key].iterkeys():
index += d_dist[key][key2]
index = index / 2.0
# find duplicates
if index in l_index:
output = os.path.basename(bgf_file) + "\t" + str(
index) + "\t" + "duplicate" + "\n"
else:
output = os.path.basename(bgf_file) + "\t" + str(index) + "\n"
# for index checking
l_index.append(index)
# write outputs
f_out_file.write(output)
# display the process
if not silent:
sys.stdout.write("\rProgress: " + "{0:>8d}".format(count) + " / " +
str(n_joblist))
sys.stdout.flush()
# check duplicates
tmp = []
dup = []
for i in l_index:
if i in tmp:
dup.append(i)
else:
tmp.append(i)
f_out_file.write("\nDuplicated Wiener Indices:\n")
f_out_file.write(str(dup) + "\n")
f_out_file.close()
return 1