def getDay(day, derived_day, month, year):
if (day == '*'):
if not math.isnan((derived_day)):
return int(derived_day)
else:
return calendar.monthrange(year, month)[1]
return day
def reposicionar_centroids(self, centroids: list, matriz: pd.DataFrame):
print("===================")
print(f"Centroid Entrada: {centroids}")
self.centroids: list = centroids
for _indice in range(0, len(centroids)):
# Filtrar pro grupo de centroids
filtro = f"centroid_id == 'centroid_{_indice}'"
sub_conjunto: pd.DataFrame = matriz.query(filtro)
if math.isnan(sub_conjunto.mean().ponto_x) or math.isnan(sub_conjunto.mean().ponto_y):
continue
# self.centroids[_indice] = [round(sub_conjunto.mean().ponto_x), round(sub_conjunto.mean().ponto_y)]
self.centroids[_indice] = [sub_conjunto.mean().ponto_x, sub_conjunto.mean().ponto_y]
print(f"Centroid Saida: {self.centroids}")
return self.centroids
def getYear(year, derived_year):
if (year == '*'):
if not math.isnan(derived_year):
return int(derived_year)
else:
#raise ValueError('Can not derive year')
return -1
else:
return int(year)
def getClosureForFolder(v):
folder = v['folder']
if (folder == 'folder'):
piece = v['piece']
item = v['item']
s1 = df.loc[df['piece'].eq(piece)
& (df['item'].eq(item) if not math.isnan(item) else True) &
(df['closure_type'].str.startswith('open_on_transfer'))]
if (s1.empty):
#check if I have closed_until or unknown_status
s2 = df.loc[df['piece'].eq(piece) &
(df['item'].eq(item) if not math.isnan(item) else True)
& (df['closure_type'].str.startswith('closed_until'))]
if (s2.empty):
return 'unknown_status'
else:
return 'closed_until'
else:
return 'open_on_transfer'
else:
return v['closure_type']
def populateClosureStartDate(v):
from datetime import datetime
dateOfBirth = v['date_of_birth']
piece = v['piece']
item = v['item']
closure_type = v['closure_type']
if (closure_type == 'closed_until'):
if (pd.isnull(dateOfBirth)):
#if I don't have a date of birth, I must copy the latest date of birth from the files
df1 = df.loc[df['piece'].eq(piece)
& (df['item'].eq(item) if not math.isnan(item) else
True)]['date_of_birth'].copy()
df1.sort(ascending=False)
latestDate = df1.iloc[0]
return (datetime.strftime(latestDate, '%Y-%m-%dT%H:%M:%S'))
else:
return str(datetime.strftime(dateOfBirth, '%Y-%m-%dT%H:%M:%S'))
def angle_between(v1, v2):
""" Returns the angle in radians between vectors 'v1' and 'v2'::
>>> angle_between((1, 0, 0), (0, 1, 0))
1.5707963267948966
>>> angle_between((1, 0, 0), (1, 0, 0))
0.0
>>> angle_between((1, 0, 0), (-1, 0, 0))
3.141592653589793
"""
v1_u = unit_vector(v1)
v2_u = unit_vector(v2)
angle = np.arccos(np.dot(v1_u, v2_u))
if math.isnan(angle):
if (v1_u == v2_u).all():
return 0.0
else:
return 180
return math.degrees(angle)
def main():
args = parse_args()
creator.create("Maximum", base.Fitness, weights=(1.0, ))
creator.create("Minimum", base.Fitness, weights=(-1.0, ))
fitness = creator.Minimum if args.minimum else creator.Maximum
creator.create("Particle",
list,
exemplar=list,
best=None,
no_improvement_counter=0,
fitness=fitness,
fragrance=float)
best_histories = []
epochs = []
accuracies = []
best_values = []
for i in range(args.iteration):
result = run_boa(args)
best_values.append(result[1])
save_fitness_history("../results/" + args.logCatalog + "/", result[4])
best_histories.append(result[3])
accuracies.append(result[2])
if accuracies[i] <= args.accuracy:
epochs.append(result[0])
save_best_fitness_history("../results/" + args.logCatalog + "/",
best_histories)
filtered_best_values = [
value for value in best_values
if not (math.isinf(value) or math.isnan(value))
]
display_and_save_results(epochs, filtered_best_values, accuracies,
args.accuracy,
"../results/" + args.logCatalog + "/")
dict_data['fsa:CashAndCashEquivalents'] = datalisten[i, dataslut]
dict_data['fsa:CashAndCashEquivalents_prev'] = datalisten[
i, dataslut - 1]
if datalisten[i, 1] == 20000:
dict_data['fsa:Equity'] = datalisten[i, dataslut]
dict_data['fsa:Equity_prev'] = datalisten[i, dataslut - 1]
if datalisten[i, 1] == 49500:
dict_data['fsa:ProfitLoss'] = datalisten[i, dataslut]
dict_data['fsa:ProfitLoss_prev'] = datalisten[i, dataslut - 1]
if datalisten[i, 1] == 49100:
dict_data[
'fsa:ProfitLossFromOrdinaryOperatingActivities'] = datalisten[
i, dataslut]
dict_data[
'fsa:ProfitLossFromOrdinaryOperatingActivities_prev'] = datalisten[
i, dataslut - 1]
slettes = []
for noegle in dict_data.keys():
if math.isnan(dict_data[noegle]):
slettes.append(noegle)
for noegle in slettes:
del dict_data[noegle]
transform_spain = spain_to_dict()
transformed_data = transform_spain.transform([dict_data])
output.append(transformed_data[0])
print(clf_EW_spain.predict([dict_data]),
clf_EW_spain.predict_proba([dict_data]))
def sectionSpeed(numOfHours,numOfDays,trainData,firstTime,n,minLon, lonLen, minLat, latLen):
'''计算路段平均速度[地区号][时段]
n表示将整个地图分成n*n个网格，地区号是其一维下标'''
#缺乏数据默认的速度
defaultVel=5.0
#初始化三维数组speed[地区号][时段][天数],值为i地区j时段第k天的速度
speed=[]
for i in range(n*n):
li=[]
speed.append(li)
for j in range(numOfHours):
lj=[]
li.append(lj)
for k in range(numOfDays):
lj.append(0.0)
for file in trainData:
df=pandas.read_csv(file,
header=None,
names=["taxiId","lat","lon","busy","time"],
dtype={"taxiId":numpy.int16,"lat":numpy.double,"lon":numpy.double,
"busy":numpy.int8,"time":numpy.str})
#df按照出租车id，时间点升序排序
df.sort_values(by=["taxiId","time"], axis=0, ascending=[True,True], inplace=True)
taxiId1=-1
lat1=0
lon1=0
sectionId1=0
time1=firstTime
temp=[]
for i in range(n*n):
li=[]
for row in df.itertuples(index=False):
taxiId2=row[0]
lat2=row[1]
lon2=row[2]
time2=datetime.datetime.strptime(row[4],"%Y/%m/%d %H:%M:%S")
if taxiId1==taxiId2 and time1.hour==time2.hour:
v=calDistance(lon1, lat1, lon2, lat2)/((time2-time1).seconds)
# print(calDistance(lon1, lat1, lon2, lat2))
# print((time2-time1).seconds)
# print("seid="+str(sectionId1)+"hourid="+str(time1.hour-firstTime.hour)+"dayId="+str((time1-firstTime).days))
speed[sectionId1][time1.hour-firstTime.hour][(time1-firstTime).days].append(v)
taxiId1=taxiId2;lat1=lat2;lon1=lon2;time1=time2
# print(sectionId1)
sectionId1=calSectionId(lon1, lat1,minLon, lonLen, minLat, latLen, n)
res=[]
for i in range(n*n):
li=[]
res.append(li)
for j in range(numOfHours):
temp=[]
for k in range(numOfDays):
temp.append(numpy.mean(speed[i][j][k]))
if math.isnan(temp[0]):
li.append(defaultVel)
else:
li.append(numpy.mean(temp))
return res
def is_nan(x):
return isinstance(x, float) and math.isnan(x)
import json
from numpy import math
try:
litindex_json_files = glob.glob("./rawdata/*.json")
conn = psycopg2.connect(
"dbname='litindex' user='******' host='0.0.0.0' password='******'")
cur = conn.cursor()
for onefile in litindex_json_files:
json_records = pd.read_json(onefile, lines=True)
for index, row in json_records.iterrows():
print(row['id'])
# print(row['institution_id'])
rowId = row['id']
if math.isnan(rowId):
rowId = 0
# print("NAN row_id =",rowId)
rowInstitutionId = row['institution_id']
if math.isnan(rowInstitutionId):
rowInstitutionId = 0.0
# print("NAN rowInstitutionId =",rowInstitutionId)
year = row['year']
if math.isnan(year):
year = 0
# print("NAN year =",year)
cur.execute(
"INSERT INTO open_syllabi(id, source_url, source_anchor, syllabus_probability, year, field_name, institution_id, grid_name, grid_country_code, text_md5, text) VALUES (%s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s);",
(
rowId,
row['source_url'],
('fsa:ProfitLoss_prev', 92),
('fsa:ProfitLoss', 93),
):
vaerdi = datalist[virksomhed][col]
dict_input_poland[tekst] = vaerdi
dict_input_poland['fsa:Assets'] = np.nan_to_num(
dict_input_poland['fsa:NoncurrentAssets']) + np.nan_to_num(
dict_input_poland['fsa:CurrentAssets']) + np.nan_to_num(
dict_input_poland['fsa:Prepayments'])
dict_input_poland['fsa:Assets_prev'] = np.nan_to_num(
dict_input_poland['fsa:NoncurrentAssets_prev']) + np.nan_to_num(
dict_input_poland['fsa:CurrentAssets_prev']) + np.nan_to_num(
dict_input_poland['fsa:Prepayments_prev'])
slettes = []
for noegle in dict_input_poland.keys():
if math.isnan(dict_input_poland[noegle]):
slettes.append(noegle)
for noegle in slettes:
del dict_input_poland[noegle]
transform_poland = Polish_to_dict()
transformed_data = transform_poland.transform([dict_input_poland])
output.append(transformed_data[0])
#print(dict_input_poland)
print(virksomhed, clf_EW_poland.predict_proba([dict_input_poland]))
#df_out = pd.DataFrame.from_dict(output)
#
#import matplotlib.pyplot as plt
##print(df_out.columns)