def drop(self,databases):
items = map(int, self.LB1.curselection())
ids = self.LB1.curselection()
list_directory = []
for id in ids:
list_directory.append(databases[int(id)])
database.drop(list_directory)
self.list_directory_refresh()
infoCallBack("ok")
def rescan():
scanner = Scanner()
scanner.existing_tracks = Track.indexByPath(Track().getAllByPath(
root_dir, ''))
database.drop()
database.connect()
Track().create_table()
Folder().create_table()
scanner.parse(root_dir)
help='Get all pages')
args = parser.parse_args()
# print args
if args.debug:
logger.setLevel(logging.DEBUG)
elif args.verbose:
logger.setLevel(logging.INFO)
if args.verbose == 1:
logger.setLevel(logging.WARNING)
if args.quiet:
logger.setLevel(logging.CRITICAL)
if args.drop:
database.drop()
if args.create:
database.create()
if not args.drop:
items_to_process = []
if args.update:
items_to_update = update.read_items_to_update(args.upfile)
# logger.debug(items_to_update)
items_to_process = get_items_with_info(items_to_update)
else:
items_to_process = get_items_from_cli(args.items)
save_authority_codes(items=items_to_process)
def similarity_classifier_knn(db, test, nncl, **kwargs):
"""Compares the test point with the database and returns a pre-defined
number of best matches, according to the chosen similarity metrics.
Args:
db (pandas df): Reference database to use for the search
test (pandas df): Test cluster
nncl (int): Number of total clusters
kwargs metrics (string): 'Label1' to use probability or
'Label2' to use similarity to find best matches
kwargs flatten (float): n-root to take for all the values in the list of
best matches. Makes the difference between the first
guess and the last guess smaller
kwargs first_values (int): How many of the best guesses to keep in the list to return
kwargs function (string): Similarity function. 'cosine', 'euclidean', 'correlation'
and 'manhattan' are accepted.
Returns:
pandas dataframe: List of the N most probable locations, with coordinates
"""
metrics = kwargs['metrics'] if 'metrics' in kwargs else 'Label1'
flatten = kwargs['flatten'] if 'flatten' in kwargs else 1.0
first_values = kwargs['first_values'] if 'first_values' in kwargs else 10
function = kwargs['function'] if 'function' in kwargs else 'cosine'
db_sparse = db.drop(columns=['cLat','cLon','rLat','rLon'])#,'Label1' if metrics == 'Label2' else 'Label2'])
test = test.drop(['cLat','cLon','rLat','rLon','Label1'])#,'Label2'])
v_test = test.tolist()
#calculate the similarity for every point in the comparison database
sim = [[] for i in range(0,nncl+1)]
for i, row in db_sparse.iterrows():
label = int(row[metrics])
row = row.drop(metrics)
#create vector
v1 = row.tolist()
if(function == 'cosine'):
sim[label].append(cosine_similarity(v1,v_test))
elif(function == 'euclidean'):
sim[label].append(euclidean_similarity(v1,v_test))
elif(function == 'correlation'):
sim[label].append(correlation_similarity(v1,v_test))
elif(function == 'manhattan'):
sim[label].append(manhattan_similarity(v1,v_test))
else:
print("ERROR: Please specify a correct classification function. Accepted functions: 'cosine', 'euclidean', 'manhattan', correlation'")
#list the most probable clusters with similarity index
cluster_sim = []
for c in sim:
if len(c) > 0:
cluster_sim.append((np.mean(c),np.var(c)))
else:
cluster_sim.append((0,0))
cluster_stat = pd.DataFrame(data=cluster_sim,columns=['Mean Similarity','Variance'])
cluster_stat.sort_values(by='Mean Similarity',ascending=False,inplace=True)
#ignore warnings in the next part
warnings.filterwarnings("ignore")
#rescale and norm for the first 10 clusters --> probabilities
head = cluster_stat.head(first_values+1)
head['rescaled'] = ((head.loc[:,'Mean Similarity'] - head.loc[:,'Mean Similarity'].min()) / (head.loc[:,'Mean Similarity'].max()-head.loc[:,'Mean Similarity'].min()))**(1/flatten)
head['Probability'] = head.loc[:,'rescaled'] / head.loc[:,'rescaled'].sum()
head.drop(columns='rescaled',inplace=True)
head.drop(head.index[first_values],inplace=True) #drop last index with probability 0
#print(head)
#add coordinates of center
center_coords = []
for index in head.index.values:
c = db.loc[np.isclose(db[metrics],index)].reset_index(drop=True).loc[1,['cLat','cLon']]
#print("Cluster {} - {} {}".format(index,c.loc['cLat'],c.loc['cLon']))
center_coords.append((c.loc['cLat'],c.loc['cLon']))
head = head.reset_index()
coords_pd = pd.DataFrame(data=center_coords,columns=['Lat','Lon'])
best_with_coords = pd.concat([head,coords_pd],axis=1)
best_with_coords.rename(index=str, columns={"index":"Cluster ID"},inplace=True)
return best_with_coords
print(f"SKIPPING - TABLE ALREADY EXISTS - {names} \n")
dataList = [
('Sipra Banerjee', '*****@*****.**', 9831790182, 55),
('Subir Banerjee', '*****@*****.**', 9831626265, 62),
('Swapna Mitra', '*****@*****.**', 9433851479, 57),
('Subhadeep Banerjee', '*****@*****.**', 7980207055, 21)
]
database.insertData(TABLENAME, dataList=dataList)
database.showAll(TABLENAME)
dataList = [
('Subhadeep Banerjee', '*****@*****.**', 7980207055, 21),
('Ria Gupta', '*****@*****.**', 5123456891, 22),
('Souma Mitra', '*****@*****.**', 6444848265, 21),
]
database.insertData(TABLENAME, dataList=dataList)
database.showAll(TABLENAME)
data = ('Ria Gupta', '*****@*****.**', 5123456891, 22)
database.deleteOne(TABLENAME, data)
print("\nAfter Deleting\n")
database.showAll(TABLENAME)
database.drop(TABLENAME)
print(f"Remaining Tables - {database.listTables()}")
help='Get all pages')
args = parser.parse_args()
# print args
if args.debug:
logger.setLevel(logging.DEBUG)
elif args.verbose:
logger.setLevel(logging.INFO)
if args.verbose == 1:
logger.setLevel(logging.WARNING)
if args.quiet:
logger.setLevel(logging.CRITICAL)
if args.drop:
database.drop()
if args.create:
database.create()
if not args.drop:
items_to_process = []
if args.update:
items_to_update = update.read_items_to_update(args.upfile)
# logger.debug(items_to_update)
items_to_process = get_items_with_info(items_to_update)
else:
items_to_process = get_items_from_cli(args.items)
save_authority_codes(items=items_to_process)
