def lreshape(data: DataFrame, groups, dropna=True, label=None):
"""
Reshape wide-format data to long. Generalized inverse of ``DataFrame.pivot``.
Accepts a dictionary, `groups`, in which each key is a new column name
and each value is a list of old column names that will be "melted" under
the new column name as part of the reshape.
Parameters
----------
data : DataFrame
The wide-format DataFrame.
groups : dict
Dictionary in the form: `{new_name : list_of_columns}`.
dropna : bool, default: True
Whether include columns whose entries are all NaN or not.
label : optional
Deprecated parameter.
Returns
-------
DataFrame
Reshaped DataFrame.
"""
if not isinstance(data, DataFrame):
raise ValueError("can not lreshape with instance of type {}".format(
type(data)))
ErrorMessage.default_to_pandas("`lreshape`")
return DataFrame(
pandas.lreshape(to_pandas(data), groups, dropna=dropna, label=label))
def import_table_doctopic(self):
"""Import data into doctopic table"""
src_file = self.get_source_file(
'output-doc-topics') #self.config['output-doc-topics']
doctopic = pd.read_csv(src_file, sep='\t', header=None)
doc = pd.DataFrame(doctopic.iloc[:, 1])
doc.columns = ['doc_tmp']
doc['src_doc_id'] = doc.doc_tmp.apply(lambda x: x.split(',')[0])
doc['doc_label'] = doc.doc_tmp.apply(lambda x: x.split(',')[1])
doc = doc[['src_doc_id', 'doc_label']]
doc.index.name = 'doc_id'
doctopic.drop(1, axis=1, inplace=True)
doctopic.rename(columns={0: 'doc_id'}, inplace=True)
y = [col for col in doctopic.columns[1:]]
doctopic_narrow = pd.lreshape(doctopic, {'topic_weight': y})
doctopic_narrow['topic_id'] = [
i for i in range(self.config['num-topics'])
for doc_id in doctopic['doc_id']
]
doctopic_narrow = doctopic_narrow[[
'doc_id', 'topic_id', 'topic_weight'
]]
doctopic_narrow.set_index(['doc_id', 'topic_id'], inplace=True)
doctopic_narrow['topic_weight_zscore'] = stats.zscore(
doctopic_narrow.topic_weight)
dtm = doctopic_narrow.reset_index()\
.set_index(['doc_id','topic_id'])['topic_weight'].unstack()
dtm.to_csv(self.tables_dir + 'DOCTOPIC.csv')
doc.to_csv(self.tables_dir + 'DOC.csv')
doctopic_narrow.to_csv(self.tables_dir + 'DOCTOPIC_NARROW.csv')
def lreshape(data: DataFrame, groups, dropna=True, label=None):
if not isinstance(data, DataFrame):
raise ValueError("can not lreshape with instance of type {}".format(
type(data)))
ErrorMessage.default_to_pandas("`lreshape`")
return DataFrame(
pandas.lreshape(to_pandas(data), groups, dropna=dropna, label=label))
def import_table_doctopic(self, src_file=None):
"""Import data into doctopic table"""
if not src_file: src_file = self.mallet['train-topics']['output-doc-topics']
if 'doc-topics-threshold' in self.mallet['train-topics']:
DOC = []
DOCTOPIC = []
src = PoloFile(src_file)
for line in src[1:]:
row = line.split('\t')
row.pop() # Pretty sure this is right
doc_id = row[0]
src_doc_id = int(row[1].split(',')[0])
doc_label = row[1].split(',')[1]
DOC.append([doc_id, src_doc_id, doc_label])
for i in range(2, len(row), 2):
topic_id = row[i]
topic_weight = row[i + 1]
DOCTOPIC.append([doc_id, topic_id, topic_weight])
doctopic = pd.DataFrame(DOCTOPIC, columns=['doc_id', 'topic_id', 'topic_weight'])
doctopic.set_index(['doc_id', 'topic_id'], inplace=True)
doctopic['topic_weight_zscore'] = stats.zscore(doctopic.topic_weight)
self.computed_thresh = round(doctopic.topic_weight.quantile(self.cfg_tw_quantile), 3)
doc = pd.DataFrame(DOC, columns=['doc_id', 'src_doc_id', 'doc_label'])
doc.set_index('doc_id', inplace=True)
self.put_table(doctopic, 'doctopic', index=True)
self.put_table(doc, 'doc', index=True)
else:
doctopic = pd.read_csv(src_file, sep='\t', header=None)
doc = pd.DataFrame(doctopic.iloc[:, 1])
doc.columns = ['doc_tmp']
doc['src_doc_id'] = doc.doc_tmp.apply(lambda x: int(x.split(',')[0]))
doc['doc_label'] = doc.doc_tmp.apply(lambda x: x.split(',')[1])
doc = doc[['src_doc_id', 'doc_label']]
doc.index.name = 'doc_id'
self.put_table(doc, 'doc', index=True)
doctopic.drop(1, axis = 1, inplace=True)
doctopic.rename(columns={0:'doc_id'}, inplace=True)
y = [col for col in doctopic.columns[1:]]
doctopic_narrow = pd.lreshape(doctopic, {'topic_weight': y})
doctopic_narrow['topic_id'] = [i for i in range(self.cfg_num_topics)
for doc_id in doctopic['doc_id']]
doctopic_narrow = doctopic_narrow[['doc_id', 'topic_id', 'topic_weight']]
doctopic_narrow.set_index(['doc_id', 'topic_id'], inplace=True)
doctopic_narrow['topic_weight_zscore'] = stats.zscore(doctopic_narrow.topic_weight)
self.computed_thresh = round(doctopic_narrow.topic_weight\
.quantile(self.cfg_tw_quantile), 3)
self.put_table(doctopic_narrow, 'doctopic', index=True)
# todo: Revisit this; in the best place to do this?
self.set_config_item('computed_thresh', self.computed_thresh)
def generalize_country_to_region(
workdata,
column: str,
countries=pd.read_csv(
'https://raw.githubusercontent.com/lukes/ISO-3166-Countries-with-Regional-Codes/master/all/all.csv',
usecols=['name', 'alpha-2', 'alpha-3', 'region'])):
"""
A DataFrame adott oszlopában lévő országneveket és kódokat cseréli le annak a régiónak a nevére,
ahol az ország található.
:param workdata: a WorkData példány, ami a DataFramet tartalmazza
:param column: az oszlop neve
:param countries: az országokat, kódokat, és régiókat tartalmazó file
:return:
"""
reshaped = pd.lreshape(countries, {
'country': ['name', 'alpha-2', 'alpha-3'],
'region': ['region', 'region', 'region']
},
dropna=False)
dictionary = dict(zip(reshaped['country'], reshaped['region']))
workdata.df[column] = workdata.df[column].map(dictionary)
def count_unique_heros():
# загружаем датасет
features = pd.read_csv('features.csv', index_col='match_id')
# получаем кол-во записей в датасете
rows_num = features.shape[0]
# заполняем пустоты в признаках
for f, n in features.count().items():
if n != rows_num:
features[f].fillna(features[f].mean(), inplace=True)
values = []
#заполняем массив наименований признаков
for f in list(features.columns):
if 'hero' in f:
values.append(f)
#решейпим матрицу в вектор
df1 = pd.lreshape(features, {'hero': values})
#выводим кол-во уникальных значений (героев)
print(df1['hero'].value_counts().shape[0])
def test_pairs(self):
data = {
'birthdt':
['08jan2009', '20dec2008', '30dec2008', '21dec2008', '11jan2009'],
'birthwt': [1766, 3301, 1454, 3139, 4133],
'id': [101, 102, 103, 104, 105],
'sex': ['Male', 'Female', 'Female', 'Female', 'Female'],
'visitdt1':
['11jan2009', '22dec2008', '04jan2009', '29dec2008', '20jan2009'],
'visitdt2':
['21jan2009', nan, '22jan2009', '31dec2008', '03feb2009'],
'visitdt3': ['05feb2009', nan, nan, '02jan2009', '15feb2009'],
'wt1': [1823, 3338, 1549, 3298, 4306],
'wt2': [2011.0, nan, 1892.0, 3338.0, 4575.0],
'wt3': [2293.0, nan, nan, 3377.0, 4805.0]
}
df = DataFrame(data)
spec = {
'visitdt': ['visitdt%d' % i for i in range(1, 4)],
'wt': ['wt%d' % i for i in range(1, 4)]
}
result = lreshape(df, spec)
exp_data = {
'birthdt': [
'08jan2009', '20dec2008', '30dec2008', '21dec2008',
'11jan2009', '08jan2009', '30dec2008', '21dec2008',
'11jan2009', '08jan2009', '21dec2008', '11jan2009'
],
'birthwt': [
1766, 3301, 1454, 3139, 4133, 1766, 1454, 3139, 4133, 1766,
3139, 4133
],
'id': [101, 102, 103, 104, 105, 101, 103, 104, 105, 101, 104, 105],
'sex': [
'Male', 'Female', 'Female', 'Female', 'Female', 'Male',
'Female', 'Female', 'Female', 'Male', 'Female', 'Female'
],
'visitdt': [
'11jan2009', '22dec2008', '04jan2009', '29dec2008',
'20jan2009', '21jan2009', '22jan2009', '31dec2008',
'03feb2009', '05feb2009', '02jan2009', '15feb2009'
],
'wt': [
1823.0, 3338.0, 1549.0, 3298.0, 4306.0, 2011.0, 1892.0, 3338.0,
4575.0, 2293.0, 3377.0, 4805.0
]
}
exp = DataFrame(exp_data, columns=result.columns)
tm.assert_frame_equal(result, exp)
result = lreshape(df, spec, dropna=False)
exp_data = {
'birthdt': [
'08jan2009', '20dec2008', '30dec2008', '21dec2008',
'11jan2009', '08jan2009', '20dec2008', '30dec2008',
'21dec2008', '11jan2009', '08jan2009', '20dec2008',
'30dec2008', '21dec2008', '11jan2009'
],
'birthwt': [
1766, 3301, 1454, 3139, 4133, 1766, 3301, 1454, 3139, 4133,
1766, 3301, 1454, 3139, 4133
],
'id': [
101, 102, 103, 104, 105, 101, 102, 103, 104, 105, 101, 102,
103, 104, 105
],
'sex': [
'Male', 'Female', 'Female', 'Female', 'Female', 'Male',
'Female', 'Female', 'Female', 'Female', 'Male', 'Female',
'Female', 'Female', 'Female'
],
'visitdt': [
'11jan2009', '22dec2008', '04jan2009', '29dec2008',
'20jan2009', '21jan2009', nan, '22jan2009', '31dec2008',
'03feb2009', '05feb2009', nan, nan, '02jan2009', '15feb2009'
],
'wt': [
1823.0, 3338.0, 1549.0, 3298.0, 4306.0, 2011.0, nan, 1892.0,
3338.0, 4575.0, 2293.0, nan, nan, 3377.0, 4805.0
]
}
exp = DataFrame(exp_data, columns=result.columns)
tm.assert_frame_equal(result, exp)
spec = {
'visitdt': ['visitdt%d' % i for i in range(1, 3)],
'wt': ['wt%d' % i for i in range(1, 4)]
}
pytest.raises(ValueError, lreshape, df, spec)
def test_pairs(self):
data = {
"birthdt": [
"08jan2009",
"20dec2008",
"30dec2008",
"21dec2008",
"11jan2009",
],
"birthwt": [1766, 3301, 1454, 3139, 4133],
"id": [101, 102, 103, 104, 105],
"sex": ["Male", "Female", "Female", "Female", "Female"],
"visitdt1": [
"11jan2009",
"22dec2008",
"04jan2009",
"29dec2008",
"20jan2009",
],
"visitdt2":
["21jan2009", np.nan, "22jan2009", "31dec2008", "03feb2009"],
"visitdt3":
["05feb2009", np.nan, np.nan, "02jan2009", "15feb2009"],
"wt1": [1823, 3338, 1549, 3298, 4306],
"wt2": [2011.0, np.nan, 1892.0, 3338.0, 4575.0],
"wt3": [2293.0, np.nan, np.nan, 3377.0, 4805.0],
}
df = DataFrame(data)
spec = {
"visitdt": [f"visitdt{i:d}" for i in range(1, 4)],
"wt": [f"wt{i:d}" for i in range(1, 4)],
}
result = lreshape(df, spec)
exp_data = {
"birthdt": [
"08jan2009",
"20dec2008",
"30dec2008",
"21dec2008",
"11jan2009",
"08jan2009",
"30dec2008",
"21dec2008",
"11jan2009",
"08jan2009",
"21dec2008",
"11jan2009",
],
"birthwt": [
1766,
3301,
1454,
3139,
4133,
1766,
1454,
3139,
4133,
1766,
3139,
4133,
],
"id": [101, 102, 103, 104, 105, 101, 103, 104, 105, 101, 104, 105],
"sex": [
"Male",
"Female",
"Female",
"Female",
"Female",
"Male",
"Female",
"Female",
"Female",
"Male",
"Female",
"Female",
],
"visitdt": [
"11jan2009",
"22dec2008",
"04jan2009",
"29dec2008",
"20jan2009",
"21jan2009",
"22jan2009",
"31dec2008",
"03feb2009",
"05feb2009",
"02jan2009",
"15feb2009",
],
"wt": [
1823.0,
3338.0,
1549.0,
3298.0,
4306.0,
2011.0,
1892.0,
3338.0,
4575.0,
2293.0,
3377.0,
4805.0,
],
}
exp = DataFrame(exp_data, columns=result.columns)
tm.assert_frame_equal(result, exp)
result = lreshape(df, spec, dropna=False)
exp_data = {
"birthdt": [
"08jan2009",
"20dec2008",
"30dec2008",
"21dec2008",
"11jan2009",
"08jan2009",
"20dec2008",
"30dec2008",
"21dec2008",
"11jan2009",
"08jan2009",
"20dec2008",
"30dec2008",
"21dec2008",
"11jan2009",
],
"birthwt": [
1766,
3301,
1454,
3139,
4133,
1766,
3301,
1454,
3139,
4133,
1766,
3301,
1454,
3139,
4133,
],
"id": [
101,
102,
103,
104,
105,
101,
102,
103,
104,
105,
101,
102,
103,
104,
105,
],
"sex": [
"Male",
"Female",
"Female",
"Female",
"Female",
"Male",
"Female",
"Female",
"Female",
"Female",
"Male",
"Female",
"Female",
"Female",
"Female",
],
"visitdt": [
"11jan2009",
"22dec2008",
"04jan2009",
"29dec2008",
"20jan2009",
"21jan2009",
np.nan,
"22jan2009",
"31dec2008",
"03feb2009",
"05feb2009",
np.nan,
np.nan,
"02jan2009",
"15feb2009",
],
"wt": [
1823.0,
3338.0,
1549.0,
3298.0,
4306.0,
2011.0,
np.nan,
1892.0,
3338.0,
4575.0,
2293.0,
np.nan,
np.nan,
3377.0,
4805.0,
],
}
exp = DataFrame(exp_data, columns=result.columns)
tm.assert_frame_equal(result, exp)
with tm.assert_produces_warning(FutureWarning):
lreshape(df, spec, dropna=False, label="foo")
spec = {
"visitdt": [f"visitdt{i:d}" for i in range(1, 3)],
"wt": [f"wt{i:d}" for i in range(1, 4)],
}
msg = "All column lists must be same length"
with pytest.raises(ValueError, match=msg):
lreshape(df, spec)
def test_pairs(self):
data = {'birthdt': ['08jan2009', '20dec2008', '30dec2008', '21dec2008',
'11jan2009'],
'birthwt': [1766, 3301, 1454, 3139, 4133],
'id': [101, 102, 103, 104, 105],
'sex': ['Male', 'Female', 'Female', 'Female', 'Female'],
'visitdt1': ['11jan2009', '22dec2008', '04jan2009',
'29dec2008', '20jan2009'],
'visitdt2':
['21jan2009', nan, '22jan2009', '31dec2008', '03feb2009'],
'visitdt3': ['05feb2009', nan, nan, '02jan2009', '15feb2009'],
'wt1': [1823, 3338, 1549, 3298, 4306],
'wt2': [2011.0, nan, 1892.0, 3338.0, 4575.0],
'wt3': [2293.0, nan, nan, 3377.0, 4805.0]}
df = DataFrame(data)
spec = {'visitdt': ['visitdt%d' % i for i in range(1, 4)],
'wt': ['wt%d' % i for i in range(1, 4)]}
result = lreshape(df, spec)
exp_data = {'birthdt':
['08jan2009', '20dec2008', '30dec2008', '21dec2008',
'11jan2009', '08jan2009', '30dec2008', '21dec2008',
'11jan2009', '08jan2009', '21dec2008', '11jan2009'],
'birthwt': [1766, 3301, 1454, 3139, 4133, 1766, 1454, 3139,
4133, 1766, 3139, 4133],
'id': [101, 102, 103, 104, 105, 101, 103, 104, 105, 101,
104, 105],
'sex': ['Male', 'Female', 'Female', 'Female', 'Female',
'Male', 'Female', 'Female', 'Female', 'Male',
'Female', 'Female'],
'visitdt': ['11jan2009', '22dec2008', '04jan2009',
'29dec2008', '20jan2009', '21jan2009',
'22jan2009', '31dec2008', '03feb2009',
'05feb2009', '02jan2009', '15feb2009'],
'wt': [1823.0, 3338.0, 1549.0, 3298.0, 4306.0, 2011.0,
1892.0, 3338.0, 4575.0, 2293.0, 3377.0, 4805.0]}
exp = DataFrame(exp_data, columns=result.columns)
tm.assert_frame_equal(result, exp)
result = lreshape(df, spec, dropna=False)
exp_data = {'birthdt':
['08jan2009', '20dec2008', '30dec2008', '21dec2008',
'11jan2009', '08jan2009', '20dec2008', '30dec2008',
'21dec2008', '11jan2009', '08jan2009', '20dec2008',
'30dec2008', '21dec2008', '11jan2009'],
'birthwt': [1766, 3301, 1454, 3139, 4133, 1766, 3301, 1454,
3139, 4133, 1766, 3301, 1454, 3139, 4133],
'id': [101, 102, 103, 104, 105, 101, 102, 103, 104, 105,
101, 102, 103, 104, 105],
'sex': ['Male', 'Female', 'Female', 'Female', 'Female',
'Male', 'Female', 'Female', 'Female', 'Female',
'Male', 'Female', 'Female', 'Female', 'Female'],
'visitdt': ['11jan2009', '22dec2008', '04jan2009',
'29dec2008', '20jan2009', '21jan2009', nan,
'22jan2009', '31dec2008', '03feb2009',
'05feb2009', nan, nan, '02jan2009',
'15feb2009'],
'wt': [1823.0, 3338.0, 1549.0, 3298.0, 4306.0, 2011.0, nan,
1892.0, 3338.0, 4575.0, 2293.0, nan, nan, 3377.0,
4805.0]}
exp = DataFrame(exp_data, columns=result.columns)
tm.assert_frame_equal(result, exp)
spec = {'visitdt': ['visitdt%d' % i for i in range(1, 3)],
'wt': ['wt%d' % i for i in range(1, 4)]}
pytest.raises(ValueError, lreshape, df, spec)
def main():
evaluate_run = False
results_folder = os.path.join(os.getcwd(),
"results_" + walk + "/" + experiment)
if not os.path.isdir(results_folder):
print(colored("Error, " + results_folder + " does not exist", 'red'))
else:
print(colored("OK, " + results_folder + " exists", 'green'))
rr = {}
rb = {}
br = {}
bb = {}
for timeout_folder in natsorted(os.listdir(os.path.join(results_folder))):
print(colored("Timeout folder:", 'blue'), timeout_folder)
if timeout_folder.endswith("pickle"):
continue
parameters = timeout_folder.split("_")
for param in parameters:
if param.startswith("timeout"):
timeout = int(param.split("#")[-1]) * 10
# print("\t timeoutR:",timeoutR)
if timeout == -1:
print(colored("\tWARNING: wrong timeout folder", 'red'))
continue
if os.path.isfile(
os.path.join(
results_folder, pickle_file_root + "_timeout#" +
str(timeout) + "_.pickle")):
run_memory_mean = pd.read_pickle(
os.path.join(
results_folder, pickle_file_root + "_timeout#" +
str(timeout) + "_.pickle"))
print(
colored(
pickle_file_root + "_timeout#" + str(timeout) +
"_.pickle already exists for timeout:" + str(timeout),
'green'))
else:
# print(colored(
# os.path.join(results_folder, pickle_file_root"_timeout#"+str(timeout*10)+"_.pickle"),
# 'red'))
# sys.exit()
for filename in natsorted(
os.listdir(os.path.join(results_folder, timeout_folder))):
filename_seed = filename.split("_")[0]
# print(filename)
if filename.endswith("areaLOG_client.tsv"):
if not os.path.getsize(
os.path.join(results_folder, timeout_folder,
filename)) > 0:
print(
colored("\tWARNING, empty file at:" + filename,
'red'))
continue
# print('\tfilename: ', filename)
df_area_client = pd.read_csv(os.path.join(
results_folder, timeout_folder, filename),
sep="\t",
header=None)
if filename.endswith("areaLOG_server.tsv"):
if not os.path.getsize(
os.path.join(results_folder, timeout_folder,
filename)) > 0:
print(
colored("\tWARNING, empty file at:" + filename,
'red'))
continue
# print('\tfilename: ', filename)
df_area_server = pd.read_csv(os.path.join(
results_folder, timeout_folder, filename),
sep="\t",
header=None)
# if filename.endswith("taskLOG_client.tsv"):
# if not os.path.getsize(os.path.join(results_folder, timeout_folder, filename)) > 0:
# print(colored("\tWARNING, empty file at:" + filename, 'red'))
# continue
# # print('\tfilename: ', filename)
# df_task_client = pd.read_csv(os.path.join(results_folder, timeout_folder, filename), sep="\t",
# header=None)
#
# if filename.endswith("taskLOG_server.tsv"):
# if not os.path.getsize(os.path.join(results_folder, timeout_folder, filename)) > 0:
# print(colored("\tWARNING, empty file at:" + filename, 'red'))
# continue
# # print('\tfilename: ', filename)
# df_task_server = pd.read_csv(os.path.join(results_folder, timeout_folder, filename), sep="\t",
# header=None)
if filename.endswith("kiloLOG_client.tsv"):
if not os.path.getsize(
os.path.join(results_folder, timeout_folder,
filename)) > 0:
print(
colored("\tWARNING, empty file at:" + filename,
'red'))
continue
# print('\tfilename: ', filename)
df_kilo_client = pd.read_csv(os.path.join(
results_folder, timeout_folder, filename),
sep="\t",
header=None)
if filename.endswith("kiloLOG_server.tsv"):
if not os.path.getsize(
os.path.join(results_folder, timeout_folder,
filename)) > 0:
print(
colored("\tWARNING, empty file at:" + filename,
'red'))
continue
# print('\tfilename: ', filename, end='\n')
df_kilo_server = pd.read_csv(os.path.join(
results_folder, timeout_folder, filename),
sep="\t",
header=None)
evaluate_run = True
if evaluate_run:
'''Kilo log part'''
if len(df_kilo_client.columns) > 145:
# print("Cutting null elements in client kilo df")
df_kilo_client.drop(
df_kilo_client.columns[len(df_kilo_client.columns)
- 1],
axis=1,
inplace=True)
if len(df_kilo_server.columns) > 145:
# print("Cutting null elements in server kilo df")
df_kilo_server.drop(
df_kilo_server.columns[len(df_kilo_server.columns)
- 1],
axis=1,
inplace=True)
col_kilo_labels = ['time']
for i in range(0, len(df_kilo_server.columns) - 1, 6):
# print(i,end=", ")
col_kilo_labels += [
'id' + str(i // 6), 'state' + str(i // 6),
'posx' + str(i // 6), 'posy' + str(i // 6),
'ori' + str(i // 6), 'same_state' + str(i // 6)
]
col_kilo_to_drop = []
for i in range((len(df_kilo_server.columns) - 1) // 6):
# print(i,end=", ")
col_kilo_to_drop += ['same_state' + str(i)]
df_kilo_server.columns = col_kilo_labels
df_kilo_client.columns = col_kilo_labels
df_kilo_server = df_kilo_server.drop(col_kilo_to_drop,
axis=1)
df_kilo_client = df_kilo_client.drop(col_kilo_to_drop,
axis=1)
'''Completed task LOG part'''
# task_label = ['time', 'id', 'creationTime', 'completitionTime', 'color', 'contained']
# df_task_client.columns = task_label
'''Area LOG part'''
col_area_labels = ['time']
for i in range(0, len(df_area_server.columns) - 2, 6):
# print(i, end=", ")
col_area_labels += [
'id' + str(i // 6), 'posx' + str(i // 6),
'posy' + str(i // 6), 'color' + str(i // 6),
'completed' + str(i // 6),
'contained' + str(i // 6)
]
# Remove last empty col and assign labels to df_area_server
if len(df_area_server.columns) > 49:
# print("Cutting null elements in area server df")
df_area_server.drop(
df_area_server.columns[len(df_area_server.columns)
- 1],
axis=1,
inplace=True)
df_area_server.columns = col_area_labels
# First df_area_client row contains garbage
# so is substituted with the second row except for the time,
# then remove Nan values in [:,49:]
if len(df_area_client.columns) > 49:
# print("Cutting null elements in area client df")
df_area_client.loc[0, 1:] = df_area_client.loc[1, 1:]
df_area_client = df_area_client.drop(np.arange(
49, len(df_area_client.columns)),
axis=1)
df_area_client.columns = col_area_labels
area_color_label = []
for i in range(num_areas):
area_color_label += ["color" + str(i)]
# print("color"+str(i))
areas_client_color = df_area_client[area_color_label].iloc[
0, :].values
areas_server_color = df_area_server[area_color_label].iloc[
0, :].values
# print(areas_client_color)
# print(areas_server_color)
area_pos_label = []
for i in range(num_areas):
area_pos_label += ["posx" + str(i)]
area_pos_label += ["posy" + str(i)]
areas_pos = df_area_client[area_pos_label].iloc[
0, :].values
# print(areas_pos)
areas_pos = areas_pos.reshape(-1, 2)
color_list = ["color" + str(i) for i in range(num_areas)]
df_area3_s = df_area_server.iloc[:1, :][color_list]
df_area3_c = df_area_client.iloc[:1, :][color_list]
for i, idx in enumerate(
range(1,
len(df_area3_c.columns) * 2, 2)):
# print(i, ' ', idx)
df_area3_c.insert(loc=idx,
column='other_col' + str(i),
value=df_area3_s.iloc[0][i])
client = [
col for col in df_area3_c.columns if 'color' in col
]
server = [
col for col in df_area3_c.columns if 'other_col' in col
]
df_area_colors = pd.lreshape(df_area3_c, {
'color_client': client,
'color_server': server
})
area_type = []
for area in df_area_colors.values:
if area[0] == 0 and area[1] == 0:
area_type += ['BB']
if area[0] == 0 and area[1] == 1:
area_type += ['BR']
if area[0] == 1 and area[1] == 0:
area_type += ['RB']
if area[0] == 1 and area[1] == 1:
area_type += ['RR']
df_area_colors.insert(loc=2,
column='area_type',
value=area_type)
'''Post process server'''
for i, kilo_id in enumerate(
np.arange(1, len(df_kilo_server.columns), 5)):
# print(colored("kilo_id:" + str((kilo_id - 1) // 5), 'blue'))
# print(df_kilo_client.iloc[:20, kilo_id+2:kilo_id+4].values, end='\n\n')
kilo_pos = df_kilo_server.iloc[:, kilo_id + i +
2:kilo_id + i +
4].values
# print(kilo_pos)
in_area = np.empty(kilo_pos.shape[0], dtype=int)
in_area.fill(-1)
for area_idx, area_pos in enumerate(areas_pos):
# print(area_idx, ' ', area_pos)
dist = np.linalg.norm(kilo_pos - area_pos, axis=1)
# print(dist, end='\n\n')
in_area = np.where(
dist < area_threshold,
df_area_colors.iloc[area_idx][-1][::-1],
in_area)
# in_area = np.where(in_area == -1, np.NaN, in_area)
# print(in_area)
df_kilo_server.insert(loc=int(kilo_id + i + 2),
column='area_type' + str(i),
value=in_area)
'''Post process client'''
for i, kilo_id in enumerate(
np.arange(1, len(df_kilo_client.columns), 5)):
# print(colored("kilo_id:" + str((kilo_id - 1) // 5), 'blue'))
# print(df_kilo_client.iloc[:20, kilo_id+2:kilo_id+4].values, end='\n\n')
kilo_pos = df_kilo_client.iloc[:, kilo_id + i +
2:kilo_id + i +
4].values
# print(kilo_pos)
in_area = np.empty(kilo_pos.shape[0], dtype=int)
in_area.fill(-1)
for area_idx, area_pos in enumerate(areas_pos):
# print(area_idx,' ', area_pos)
dist = np.linalg.norm(kilo_pos - area_pos, axis=1)
# print(dist, end='\n\n')
in_area = np.where(
dist < area_threshold,
df_area_colors.iloc[area_idx][-1], in_area)
# in_area = np.where(in_area == -1, np.NaN, in_area)
# print(in_area)
df_kilo_client.insert(loc=int(kilo_id + i + 2),
column='area_type' + str(i),
value=in_area)
'''Here finally evaluated in which area the timeout elapses'''
kilo_resume = [["state" + str(i), "area_type" + str(i)]
for i in range(num_robots)]
kilo_resume = np.reshape(kilo_resume, (-1))
server_kilo_resume = df_kilo_server.iloc[:][kilo_resume]
client_kilo_resume = df_kilo_client.iloc[:][kilo_resume]
total_exp_df = client_kilo_resume.join(server_kilo_resume,
lsuffix='_c',
rsuffix='_s')
if value_studied == "mean_timeout":
timeout_count = pd.DataFrame(
columns=['RR', 'RB', 'BR', 'BB'])
for i in range(0, len(total_exp_df.columns), 2):
# print(total_exp_df.iloc[:50,i:i+2])
kilo_state = total_exp_df.iloc[:, i:i + 2]
kilo_state = kilo_state.replace(2, 3)
mask = (kilo_state[
kilo_state.columns.values[0]].diff() == 2)
# print(kilo_state[mask])
# print(kilo_state[mask][kilo_state.columns.values[1]].value_counts(), end='\n\n')
robot_timeout = kilo_state[mask][
kilo_state.columns.values[1]].value_counts(
).to_frame().T
# robot_timeout = pd.DataFrame(kilo_state[mask][kilo_state.columns.values[1]].value_counts(), columns=['RR, RB,BR,BB'])
# print(robot_timeout)
timeout_count = timeout_count.append(robot_timeout)
# print(robot_timeout, end='\n\n')
timeout_count = timeout_count.fillna(0)
single_run_mean = timeout_count.mean(axis=0)
else:
completed_area_count = pd.DataFrame(
columns=['RR', 'RB', 'BR', 'BB'])
for i in range(0, len(total_exp_df.columns), 2):
# print(total_exp_df.iloc[:50,i:i+2])
kilo_state = total_exp_df.iloc[:, i:i + 2]
mask = (kilo_state[
kilo_state.columns.values[0]].diff() == -1)
# print(kilo_state[mask])
# print(kilo_state[mask][kilo_state.columns.values[1]].value_counts(), end='\n\n')
robot_completed_area = kilo_state[mask][
kilo_state.columns.values[1]].value_counts(
).to_frame().T
# robot_completed_area = pd.DataFrame(kilo_state[mask][kilo_state.columns.values[1]].value_counts(), columns=['RR, RB,BR,BB'])
# print(robot_completed_area)
completed_area_count = completed_area_count.append(
robot_completed_area)
# print(robot_completed_area, end='\n\n')
completed_area_count = completed_area_count.fillna(0)
single_run_mean = completed_area_count.mean(axis=0)
single_df = single_run_mean.to_frame().T
single_df.index = [filename_seed]
if os.path.isfile(
os.path.join(
results_folder, pickle_file_root +
"_timeout#" + str(timeout) + "_.pickle")):
run_memory_mean = pd.read_pickle(
os.path.join(
results_folder, pickle_file_root +
"_timeout#" + str(timeout) + "_.pickle"))
run_memory_mean = run_memory_mean.append(single_df)
run_memory_mean.to_pickle(
os.path.join(
results_folder, pickle_file_root +
"_timeout#" + str(timeout) + "_.pickle"))
print("Timeout:", timeout, end=", ")
print("Appending mean run, file size: ",
run_memory_mean.shape)
else:
print("Timeout:", timeout, end=", ")
print("Writing mean run")
single_df.to_pickle(
os.path.join(
results_folder, pickle_file_root +
"_timeout#" + str(timeout) + "_.pickle"))
evaluate_run = False
rr[timeout] = run_memory_mean['RR'].values
rb[timeout] = run_memory_mean['RB'].values
br[timeout] = run_memory_mean['BR'].values
bb[timeout] = run_memory_mean['BB'].values
if value_studied == "mean_timeout":
figureName = 'meanElapsedTimeout'
else:
figureName = 'meanCompletedAreas'
figureName += '_groupsize' + groupSize + '_' + experiment + '_' + walk
print("rr", rr)
boxplots_utils.grouped_4_boxplot(rr, rb, br, bb, y_lim, figureName)
wc_full_data = wc_full_data.sort_values(by=['Type', 'SKU'])
is_french = wc_full_data['SKU'].astype(str).str.contains('_fr')
wc_data_french = wc_full_data[is_french]
wc_data = wc_full_data[is_french == False]
# Clean the invalid name
wc_data['Name'] = wc_data['Name'].astype(str)
wc_data = wc_data[wc_data['Name'].str.contains('#REF!') == False]
# wc_data contains English rows
# wc_data_french contains French Rows
attribute_keys = wc_data.columns[wc_data.columns.str.endswith(' name')]
attribute_values = wc_data.columns[wc_data.columns.str.endswith('value(s)')]
wc_data_attributes = pd.lreshape(wc_data, {
'key': attribute_keys,
'value': attribute_values
})
wc_data_attributes = wc_data_attributes.pivot(index='ID',
columns='key',
values='value')
wc_data = pd.merge(wc_data, wc_data_attributes, on='ID')
wc_data['slug'] = wc_data['Name'].apply(lambda x: slugify(x))
# %%
slug_mask = wc_data['slug'].duplicated(keep=False)
wc_data.loc[slug_mask, ['slug', 'Name']].sort_values(by=['Name'])
wc_data.loc[slug_mask,
'slug'] += wc_data.groupby('slug').cumcount().add(1).astype(str)
wc_data['new_sku'] = wc_data['SKU'].fillna('').apply(lambda x: x.split('-')[0])
def make_link_and_node_df(sankey_df, num_steps: int, dropna=False):
"""Takes a df in the following format (output of make_sankey_df):
| 0 | 1 | ... | num | step_0 | step_1 | ...
----+-------+-------+-----+-----+--------+--------+-----
1 | cat_1 | cat_2 | ... | 2 | 0 | 8 | ...
2 | cat_2 | None | ... | 10 | 1 | 9 | ...
...
Returns link_df:
| source | target | num
---+--------+--------+-----
0 | 0 | 8 | 114
1 | 1 | 9 | 57
...
Returns node_df:
| source | label | step
---+--------+-------+--------
0 | 0 | cat_1 | step_0
1 | 1 | cat_2 | step_0
...
"""
# reshape into source-target
steps = range(num_steps)
link_df = pd.lreshape(sankey_df,
groups={
'source':
[f'step_{step}' for step in steps[:-1]],
'target': [f'step_{step}' for step in steps[1:]]
})[['source', 'target', 'num']]
link_df = link_df.groupby(['source', 'target']).sum().reset_index()
# get index labels
node_df = pd.lreshape(
sankey_df,
groups={
'source': [f'step_{step}' for step in steps],
'label': steps
})[['source', 'label'
]].drop_duplicates().sort_values('source').reset_index(drop=True)
# link source indices to step
step_source = sankey_df[[f'step_{step}'
for step in steps]].to_dict(orient='list')
step_source = {k: list(set(v)) for k, v in step_source.items()}
source_step_dict = {}
for k, v in step_source.items():
for source in v:
source_step_dict[source] = k
node_df['step'] = node_df['source'].apply(lambda x: source_step_dict[x])
if dropna is True:
# generate new indices for link_df
step_stack_df = pd.lreshape(
link_df, {'step_stack': ['source', 'target']})[['step_stack']]
step_stack_df['new_idx'] = step_stack_df['step_stack'].astype(
'category').cat.codes
step_stack_df = step_stack_df.drop_duplicates()
replace_dict = dict(
zip(step_stack_df['step_stack'], step_stack_df['new_idx']))
link_df.loc[:, ['source', 'target'
]] = link_df.loc[:, ['source', 'target']].replace(
replace_dict) # reassign missing keys
# filter out missing keys from node_df
node_df = node_df[(node_df['source'].isin(replace_dict.keys()))]
node_df.loc[:, 'source'] = node_df.loc[:, 'source'].replace(
replace_dict) # reassign missing keys
return link_df, node_df
def process():
input_json = json.load(open('input.json', 'r'))
m_i_df = pd.read_pickle(input_json['data_pickle'])
#input_df
#result = current_app.
print('proc till tolist')
m_i_df["away_players"] = m_i_df[[
'away1_id', 'away2_id', 'away3_id', 'away4_id', "away5_id", "away6_id",
"away7_id", 'away8_id', 'away9_id', 'away10_id', 'away11_id'
]].values.tolist()
m_i_df["home_players"] = m_i_df.loc[:, [
'home1_id', 'home2_id', 'home3_id', 'home4_id', 'home5_id', 'home6_id',
'home7_id', 'home8_id', 'home9_id', 'home10_id', 'home11_id'
]].values.tolist()
m_i_df["score_home"], m_i_df["score_away"] = m_i_df["score"].str.split(
'-', 1).str
m_i_df["score_home"] = m_i_df["score_home"].apply(int)
m_i_df["score_away"] = m_i_df["score_away"].apply(int)
m_i_df["away_players"] = m_i_df[[
'away1_id', 'away2_id', 'away3_id', 'away4_id', "away5_id", "away6_id",
"away7_id", 'away8_id', 'away9_id', 'away10_id', 'away11_id'
]].values.tolist()
m_i_df["home_players"] = m_i_df.loc[:, [
'home1_id', 'home2_id', 'home3_id', 'home4_id', 'home5_id', 'home6_id',
'home7_id', 'home8_id', 'home9_id', 'home10_id', 'home11_id'
]].values.tolist()
m_i_df["away_players_bd"] = m_i_df["away_players"].apply(eletkor_to_list)
m_i_df["home_players_bd"] = m_i_df["home_players"].apply(eletkor_to_list)
m_i_df["away_players_bd"] = m_i_df[[
'date',
'away_players_bd',
]].values.tolist()
m_i_df["home_players_bd"] = m_i_df[[
'date',
'home_players_bd',
]].values.tolist()
m_i_df["avg_home_age"] = m_i_df["home_players_bd"].apply(calculate_age)
m_i_df["avg_away_age"] = m_i_df["away_players_bd"].apply(calculate_age)
m_i_df["max_age"] = m_i_df["home_players_bd"].apply(calculate_max_age)
m_i_df["max_age_place"] = m_i_df["home_players_bd"].apply(
calculate_max_age_place)
m_i_df["AVG_HEIGHT_H"] = m_i_df["home_players"].apply(avg_height)
m_i_df["AVG_HEIGHT_A"] = m_i_df["away_players"].apply(avg_height)
m_i_df["stad_cap_a"] = m_i_df.away_team.apply(stad_cap)
m_i_df["stad_cap_h"] = m_i_df.home_team.apply(stad_cap)
m_i_df["all_goals"] = m_i_df["score_away"] + m_i_df["score_home"]
m_i_df["goal_difference"] = m_i_df["score_away"] - m_i_df["score_home"]
m_i_df_proba = pd.concat([
m_i_df[:], m_i_df["away_players"].apply(height_to_list).apply(
pd.Series).rename(columns=lambda x: 'heights_a_' + str(x + 1))[:]
],
axis=1)
m_i_df_proba = pd.concat([
m_i_df_proba[:], m_i_df["home_players"].apply(height_to_list).apply(
pd.Series).rename(columns=lambda x: 'heights_h_' + str(x + 1))[:]
],
axis=1)
pos_height_df = pd.lreshape(
m_i_df_proba, {
'position': [
'away1_pos', 'away2_pos', 'away3_pos', 'away4_pos',
"away5_pos", "away6_pos", "away7_pos", 'away8_pos',
'away9_pos', 'away10_pos', 'away11_pos', 'home1_pos',
'home2_pos', 'home3_pos', 'home4_pos', 'home5_pos',
'home6_pos', 'home7_pos', 'home8_pos', 'home9_pos',
'home10_pos', 'home11_pos'
],
'heights': [
'heights_a_1', 'heights_a_2', 'heights_a_3', 'heights_a_4',
'heights_a_5', 'heights_a_6', "heights_a_7", "heights_a_8",
"heights_a_9", 'heights_a_10', 'heights_a_11', 'heights_h_1',
'heights_h_2', 'heights_h_3', 'heights_h_4', 'heights_h_5',
'heights_h_6', "heights_h_7", "heights_h_8", "heights_h_9",
'heights_h_10', 'heights_h_11'
]
}).pipe(lambda x: x[["position", "heights"]])
pos_height_df["heights"] = pos_height_df["heights"].apply(try_height_2)
m_i_df = pd.concat([
m_i_df[:],
m_i_df["away_players_bd"].apply(calculate_valamelyik_age).apply(
pd.Series).rename(columns=lambda x: 'ages_a_' + str(x + 1))[:]
],
axis=1)
m_i_df = pd.concat([
m_i_df[:],
m_i_df["home_players_bd"].apply(calculate_valamelyik_age).apply(
pd.Series).rename(columns=lambda x: 'ages_h_' + str(x + 1))[:]
],
axis=1)
pos_age_df = pd.lreshape(
m_i_df, {
'position': [
'away1_pos', 'away2_pos', 'away3_pos', 'away4_pos',
"away5_pos", "away6_pos", "away7_pos", 'away8_pos',
'away9_pos', 'away10_pos', 'away11_pos', 'home1_pos',
'home2_pos', 'home3_pos', 'home4_pos', 'home5_pos',
'home6_pos', 'home7_pos', 'home8_pos', 'home9_pos',
'home10_pos', 'home11_pos'
],
'ages': [
'ages_a_1', 'ages_a_2', 'ages_a_3', 'ages_a_4', 'ages_a_5',
'ages_a_6', "ages_a_7", "ages_a_8", "ages_a_9", 'ages_a_10',
'ages_a_11', 'ages_h_1', 'ages_h_2', 'ages_h_3', 'ages_h_4',
'ages_h_5', 'ages_h_6', "ages_h_7", "ages_h_8", "ages_h_9",
'ages_h_10', 'ages_h_11'
]
}).pipe(lambda x: x)[["position", "ages"]]
print("EDDIG ELJUT - PROCESS")
new_df = pd.merge(m_i_df[["away_team","home_team","date","avg_home_age","avg_away_age","result","score"]]\
,current_app.m_o_df[["odds","odds1" ,"odds2" ,"oddsx","team_id_at","team_id_ht","date","score"]]\
,how='left', left_on=["away_team","home_team","date","score"], right_on = ["team_id_at","team_id_ht","date","score"])
new_df[["odds", "odds1", "odds2",
"oddsx"]] = new_df[["odds", "odds1", "odds2",
"oddsx"]].applymap(oddsx_to_float)
new_df = new_df.dropna(subset=['odds'])
new_df["result"] = new_df["result"].apply(lambda x: float(x))
new_df["win_as_udog_h"] = new_df.loc[:,
["odds1", "odds2", "result"]].apply(
win_udog_h, axis=1)
new_df["win_as_udog_a"] = new_df.loc[:,
["odds1", "odds2", "result"]].apply(
win_udog_a, axis=1)
new_df["lose_fav_h"] = new_df.loc[:, ["odds1", "odds2", "result"]].apply(
lose_fav_h, axis=1)
new_df["lose_fav_a"] = new_df.loc[:, ["odds1", "odds2", "result"]].apply(
lose_fav_a, axis=1)
new_df[
"looser"] = new_df.loc[:,
["team_id_ht", "team_id_at", "result"]].apply(
looser, axis=1)
new_df["looser_odds"] = new_df.loc[:, ["odds1", "odds2", "result"]].apply(
looser_odds, axis=1)
m_i_df["away_values"] = m_i_df.loc[:, ["date", "away_players"]].apply(
player_value, axis=1)
m_i_df["home_values"] = m_i_df.loc[:, ["date", "home_players"]].apply(
player_value, axis=1)
new_df = pd.merge(new_df,
m_i_df[[
"away_values", "home_values", "away_team",
"home_team", "date", "score"
]],
how='left',
left_on=["away_team", "home_team", "date", "score"],
right_on=["away_team", "home_team", "date", "score"])
value_a = m_i_df['away_values'].apply(pd.Series)
value_h = m_i_df['home_values'].apply(pd.Series)
value_a = value_a.rename(columns=lambda x: 'value_a_' + str(x + 1))
value_h = value_h.rename(columns=lambda x: 'value_h_' + str(x + 1))
out = [
{'most-used-formation':m_i_df[['home_formation','away_formation']].unstack().value_counts().index[0]},
{'number-of-players-with-no-games':str(len(current_app.p_i_df)-len(current_app.p_i_df.loc[current_app.p_i_df.playerid.isin(pd.DataFrame(m_i_df[['away10_id','away11_id','away1_id','away2_id','away3_id','away4_id',"away5_id","away6_id","away7_id",'away8_id','away9_id','home1_id','home2_id','home3_id','home4_id','home5_id','home6_id','home7_id','home8_id','home9_id','home10_id','home11_id']].unstack().unique())[0])]))},
{'player-with-highest-number-of-games':str(m_i_df[['away10_id','away11_id','away1_id','away2_id','away3_id','away4_id',"away5_id","away6_id","away7_id",'away8_id','away9_id','home1_id','home2_id','home3_id','home4_id','home5_id','home6_id','home7_id','home8_id','home9_id','home10_id','home11_id']].unstack().value_counts().index[0])},
{'player-with-highest-number-of-games-where-his-team-didnt-concede': int(pd.concat([pd.DataFrame(m_i_df.loc[(m_i_df['score_home'] == 0)][['home1_id','home2_id','home3_id','home4_id','home5_id','home6_id','home7_id','home8_id','home9_id','home10_id','home11_id']]),pd.DataFrame(m_i_df.loc[(m_i_df['score_away'] == 0)][['away10_id','away11_id','away1_id','away2_id','away3_id','away4_id',"away5_id", "away6_id","away7_id",'away8_id','away9_id']])]).unstack().value_counts().index[0])},
{'most-games-played-in-same-position-by-player':str(
pd.lreshape(m_i_df, {'player_id':['away10_id','away11_id','away1_id','away2_id','away3_id','away4_id',"away5_id","away6_id","away7_id",'away8_id','away9_id','home1_id','home2_id','home3_id','home4_id','home5_id','home6_id','home7_id','home8_id','home9_id','home10_id','home11_id'],
'position':['away10_pos','away11_pos','away1_pos','away2_pos','away3_pos','away4_pos',"away5_pos","away6_pos","away7_pos",'away8_pos','away9_pos','home1_pos','home2_pos','home3_pos','home4_pos','home5_pos','home6_pos','home7_pos','home8_pos','home9_pos','home10_pos','home11_pos']})\
.pipe(lambda x:x)[["player_id","position"]]\
.groupby(pd.lreshape(m_i_df, {'player_id':['away10_id','away11_id','away1_id','away2_id','away3_id','away4_id',"away5_id", "away6_id","away7_id",'away8_id','away9_id','home1_id','home2_id','home3_id','home4_id','home5_id','home6_id','home7_id','home8_id','home9_id','home10_id','home11_id'],
'position':['away10_pos','away11_pos','away1_pos','away2_pos','away3_pos','away4_pos',"away5_pos","away6_pos","away7_pos",'away8_pos','away9_pos','home1_pos','home2_pos','home3_pos','home4_pos','home5_pos','home6_pos','home7_pos','home8_pos','home9_pos','home10_pos','home11_pos']})\
.pipe(lambda x:x)[["player_id","position"]].columns.tolist(),as_index=False).size().max())},
{'most-different-positions-by-player':str(pd.lreshape(m_i_df, {'player_id':['away10_id','away11_id','away1_id','away2_id','away3_id','away4_id',"away5_id",
"away6_id","away7_id",'away8_id','away9_id','home1_id','home2_id','home3_id','home4_id','home5_id','home6_id','home7_id',
'home8_id','home9_id','home10_id','home11_id'],
'position':['away10_pos','away11_pos','away1_pos','away2_pos','away3_pos','away4_pos',"away5_pos",
"away6_pos","away7_pos",'away8_pos','away9_pos',
'home1_pos','home2_pos','home3_pos','home4_pos','home5_pos','home6_pos','home7_pos',
'home8_pos','home9_pos','home10_pos','home11_pos']}).pipe(lambda x:x)[["player_id","position"]][["player_id","position"]].groupby('player_id')["position"].nunique().max())},
{'most-different-formations-by-player':str(pd.lreshape(m_i_df, {'player_id':['away10_id','away11_id','away1_id','away2_id','away3_id','away4_id',"away5_id","away6_id","away7_id",'away8_id','away9_id','home1_id','home2_id','home3_id','home4_id','home5_id','home6_id','home7_id','home8_id','home9_id','home10_id','home11_id'],'formation':['away_formation','away_formation','away_formation','away_formation','away_formation','away_formation','away_formation','away_formation','away_formation','away_formation','away_formation',"home_formation","home_formation","home_formation","home_formation","home_formation","home_formation","home_formation","home_formation","home_formation","home_formation","home_formation"]}).pipe(lambda x: x)[["player_id","formation"]].groupby('player_id')["formation"].nunique().max())},
{'largest-odds-overcome-in-game':new_df[new_df["result"]!=0.0]["odds"].max()},
{'largest-height-difference-overcome-in-game':m_i_df.loc[:,["result","AVG_HEIGHT_H","AVG_HEIGHT_A"]].apply(height_diff_OC,axis=1).max()},
{'longest-time-in-days-between-two-games-for-player':None},
{'biggest-value-difference':str(m_i_df.loc[:,["away_values" ,"home_values"]].apply(values_diff,axis=1).max())},
{'biggest-value-difference-upset':int(abs(m_i_df.loc[:,["away_values" ,"home_values","result"]].apply(values_diff_ups,axis=1).min()))}, # an upset means the unexpected team won
{'biggest-value-difference-with-higher-odds':int(new_df.loc[:,["away_values" ,"home_values","odds1","odds2"]].apply(values_diff_ups_odds,axis=1).max())},
{'biggest-stadium-capacity-difference-upset':None},
{'capacity-of-stadium-of-team-with-most-games':pd.DataFrame(pd.lreshape(m_i_df, {'team_id':['away_team','home_team'],'seats':["stad_cap_a",'stad_cap_h']}).groupby(pd.lreshape(m_i_df, {'team_id':['away_team','home_team'],'seats':["stad_cap_a",'stad_cap_h']})[["team_id","seats"]].columns.tolist(),as_index=False).size()).idxmax()[0][1]},
{'id-of-oldest-team-to-win-a-game':id_of_oldest_team_to_win_a_game(m_i_df)},
{'biggest-age-difference-between-teams-match-id':int(m_i_df.iloc[abs(m_i_df["avg_away_age"]-m_i_df["avg_home_age"]).idxmax(),:]["mkey"])},
{'median-of-winning-team-average-age':(m_i_df.loc[:,["result","avg_away_age","avg_home_age"]].apply(gyoztes_kor,axis=1)).median()},
{'median-of-favorite-team-average-age':int(new_df.loc[:,["odds1" ,"odds2","avg_home_age","avg_away_age"]].apply(fav_age,axis=1).median())}, # favorite means has lower odds of winning
{'median-of-underdog-team-average-age':int(new_df.loc[:,["odds1" ,"odds2","avg_home_age","avg_away_age"]].apply(udog_age,axis=1).median())}, # underdog means has higher odds of winning
{'team-with-most-wins-as-underdog':pd.lreshape(new_df, {'team_id':['team_id_at','team_id_ht'],'wins_as_udog':["win_as_udog_a",'win_as_udog_h']}).pipe(lambda x: x[["team_id","wins_as_udog"]])\
.groupby("team_id").agg({"wins_as_udog":"sum"})["wins_as_udog"].idxmax()},
{'team-with-most-losses-as-favorite':pd.lreshape(new_df, {'team_id':['team_id_at','team_id_ht'],'lose_as_fav':["lose_fav_a",'lose_fav_h']}).pipe(lambda x: x[["team_id","lose_as_fav"]])\
.groupby("team_id").agg({"lose_as_fav":"sum"})["lose_as_fav"].idxmax()},
{'team-with-lowest-average-odds-of-draw':pd.lreshape(new_df, {'team_id':['team_id_at','team_id_ht'],'oddsx':["oddsx",'oddsx']}).pipe(lambda x: x[["team_id","oddsx"]])\
.groupby("team_id").agg({"oddsx":"mean"})["oddsx"].idxmin()},
{'position-with-highest-average-value':None},
{'position-with-largest-average-height':pos_height_df.groupby("position").agg({"heights":"mean"})["heights"].idxmax()},
{'position-with-youngest-average-age':pos_age_df.groupby("position").agg({"ages":"mean"})["ages"].idxmin()},
{'goalkeeper-with-most-clean-sheets':None},#az átlagosan legtöbb gólt kapó kapus születési dátuma
{'stadium-capactiy-of-team-with-most-avg-goals-in-a-game':None},#átlagosan leggólgazdagabb meccseket játszó csapat stadionjának befogadóképessége
{'team-with-highest-profit-for-losing':int(new_df[["looser_odds","looser"]].groupby("looser")["looser_odds"].sum().idxmax())},#a csapat, akinek, ha minden meccsén ellenük fogadsz, összesítve a legnagyobb profitot termeli (mindig ugyanakkora összeggel fogadsz rá)
{'largest-std-in-goal-difference-team':int(pd.lreshape(m_i_df, {'team_id':['away_team','home_team'],
'goal_difference':["goal_difference",'goal_difference']}).pipe(lambda x: x[["team_id","goal_difference"]])\
.groupby("team_id").agg({"goal_difference":"std"})["goal_difference"].idxmax())},#a legnagyobb gólkülönbség szórással rendelkező csapat
{'player-with-most-different-teams': int(pd.lreshape(m_i_df, {'player_id':['away10_id','away11_id','away1_id','away2_id','away3_id','away4_id',"away5_id",
"away6_id","away7_id",'away8_id','away9_id',
'home1_id','home2_id','home3_id','home4_id','home5_id','home6_id','home7_id',
'home8_id','home9_id','home10_id','home11_id'],
'team_id':['away_team','away_team','away_team','away_team','away_team','away_team','away_team','away_team','away_team','away_team','away_team',
'home_team','home_team','home_team','home_team','home_team','home_team','home_team','home_team','home_team','home_team','home_team']})\
.groupby("player_id")["team_id"].nunique().idxmax())},#a legtöbb csapatban pályára lépő játékos
{'longest-losing-streak-team':None},#a leghosszabb vesztes széria by team
{'longest-home-winning-streak-stadium-capacity':None },#leghosszabb hazai pályás győzelmi széria helyszínének befogadóképessége
{'win-ratio-of-actual-highest-rated-player':None},#az adott időpillanatban legértékesebb játékos átlagos win ratioja
{'oldest-player-to-win-a-home-game':m_i_df.iloc[m_i_df[m_i_df["result"]==1]["max_age"].idxmax(),:]["home_players"][m_i_df.iloc[m_i_df[m_i_df["result"]==1]["max_age"].idxmax(),:]["max_age_place"][0]]}#a legidősebb hazai pályán győztes meccset játszó játékos
]
json.dump(out, open('output.json', 'w'))
return 'FING'
def compute(hi,d_hrs,d_divo,period=3,transform=1):
#compute moments, period
#says how many years correspond to one
#period
#Get Date at Interview
hi.insert(0, 'IDN', range(0, len(hi)))
hi['res']=hi['NUMUNION']+hi['NUMCOHMR']
#Get Duration bins
bins_d=np.linspace(0,1200,int((100/period)+1))
bins_d_label=np.linspace(1,len(bins_d)-1,len(bins_d)-1)
##########################
#Gen cohabitation Dataset
#########################
#Get date at interview
hi['int']=hi['IDATMM']+(hi['IDATYY']-1900)*12
#Gen age at interview
hi['ageint']=round((((hi['IDATYY']-1900)*12+hi['IDATMM'])-hi['birth_month'])/12,0)
#Take only if cohabitations
coh=hi[(hi['NUMUNION']-hi['NUMMAR']>0) | (hi['NUMCOHMR']>0)].copy()
#Create number of cohabitations
coh['num']=0.0
for i in range(9):
if(np.any(coh['HOWBEG0'+str(i+1)])=='coh'):
coh.loc[coh['HOWBEG0'+str(i+1)]=='coh','num']=coh.loc[coh['HOWBEG0'+str(i+1)]=='coh','num']+1.0
#Expand the data
cohe=coh.loc[coh.index.repeat(np.array(coh.num,dtype=np.int32))]
#Link each cohabitation to relationship number
cohe['rell'] = cohe.groupby(['IDN']).cumcount()+1
cohe['cou']=1
cohe['rel']=None
for i in range(9):
if(np.any(coh['HOWBEG0'+str(i+1)])=='coh'):
cohe.loc[(cohe['HOWBEG0'+str(i+1)]=='coh') & (cohe['rell']==cohe['cou']),'rel']=i+1
cohe.loc[cohe['HOWBEG0'+str(i+1)]=='coh','cou']= cohe.loc[cohe['HOWBEG0'+str(i+1)]=='coh','cou']+1
#Get beginning and end of relationhip
cohe['beg']=-1
cohe['endd']=-1
cohe['how']=-1
cohe['mar']=-1
for i in range(9):
cohe.loc[(i+1==cohe['rel']),'beg']=cohe.loc[(i+1==cohe['rel']),'BEGDAT0'+str(i+1)]
cohe.loc[(i+1==cohe['rel']),'endd']=cohe.loc[(i+1==cohe['rel']),'ENDDAT0'+str(i+1)]
cohe.loc[(i+1==cohe['rel']),'how']=cohe.loc[(i+1==cohe['rel']),'HOWEND0'+str(i+1)]
cohe.loc[(i+1==cohe['rel']),'mar']=cohe.loc[(i+1==cohe['rel']),'MARDAT0'+str(i+1)]
#add here an indicator of whether it should be unilateral duvorce scenario
#Get how relationship end
cohe['fine']='censored'
cohe.loc[cohe['how']=='sep','fine']='sep'
cohe.loc[cohe['how']=='div','fine']='mar'
cohe.loc[(cohe['how']=='intact') & (cohe['mar']>1),'fine']='mar'
#Replace censored date if still together
cohe['end']=-1
cohe.loc[cohe['fine']=='sep','end']=cohe.loc[cohe['fine']=='sep','endd']
cohe.loc[cohe['fine']=='mar','end']=cohe.loc[cohe['fine']=='mar','mar']
cohe.loc[cohe['fine']=='censored','end']=cohe.loc[cohe['fine']=='censored','int']
#Duration
cohe['dur']=cohe['end']-cohe['beg']
#Keep if no error for duration
cohe=cohe[(cohe['dur']>0) & (cohe['dur']<2000)]
#Transform Duration in Years
cohe['dury'] = pd.cut(x=cohe['dur'], bins=bins_d,labels=bins_d_label)
cohe['dury']=cohe['dury'].astype(float)
#Eliminate non useful things
del coh
##########################
#Gen marriage Dataset
#########################
#Take only if marriages
mar=hi[hi['NUMMAR']>0].copy()
#Create number of cohabitations
mar['num']=0
for i in range(9):
mar.loc[mar['MARDAT0'+str(i+1)]>0,'num']=mar.loc[mar['MARDAT0'+str(i+1)]>0,'num']+1
#Expand the data
mare=mar.loc[mar.index.repeat(mar.num)]
#Link each marriage to relationship number
mare['rell'] = mare.groupby(['IDN']).cumcount()+1
mare['cou']=1
mare['rel']=None
for i in range(9):
mare.loc[(mare['MARDAT0'+str(i+1)]>0) & (mare['rell']==mare['cou']),'rel']=i+1
mare.loc[mare['MARDAT0'+str(i+1)]>0,'cou']= mare.loc[mare['MARDAT0'+str(i+1)]>0,'cou']+1
#Get beginning and end of relationhip
mare['beg']=-1
mare['endd']=-1
mare['how']=-1
mare['mar']=-1
for i in range(9):
mare.loc[(i+1==mare['rel']),'beg']=mare.loc[(i+1==mare['rel']),'MARDAT0'+str(i+1)]
mare.loc[(i+1==mare['rel']),'endd']=mare.loc[(i+1==mare['rel']),'ENDDAT0'+str(i+1)]
mare.loc[(i+1==mare['rel']),'how']=mare.loc[(i+1==mare['rel']),'HOWEND0'+str(i+1)]
#Get how relationship end
mare['fine']='censored'
mare.loc[mare['how']=='div','fine']='div'
#Replace censored date if still together
mare['end']=-1
mare.loc[mare['fine']=='div','end']=mare.loc[mare['fine']=='div','endd']
mare.loc[mare['fine']=='censored','end']=mare.loc[mare['fine']=='censored','int']
#Duration
mare['dur']=mare['end']-mare['beg']
#Keep if no error for duration
mare=mare[(mare['dur']>0) & (mare['dur']<2000)]
#Transform Duration in Years
mare['dury'] = pd.cut(x=mare['dur'], bins=bins_d,labels=bins_d_label)
mare['dury']=mare['dury'].astype(float)
del mar
#############################
#Build relationship by month
##############################
#Eliminate observation if info on beg-end not complete
#for i in range(9):
# hi=hi[(np.isfinite(hi['BEGDAT0'+str(i+1)])) & (hi['BEGDAT0'+str(i+1)]<3999)]
#Get date in time at which the guy is 20,25...,50 (9)
for j in range(7):
hi['time_'+str(20+(j)*5)]=hi['DOBY']*12+hi['DOBM']+(20+(j)*5)*12
#Get the status
for j in range(7):
#Create the variable of Status
hi['status_'+str(20+(j)*5)]='single'
for i in range(9):
if(np.any(hi['HOWBEG0'+str(i+1)])!=None):
#Get if in couple
hi.loc[(hi['time_'+str(20+(j)*5)]>=hi['BEGDAT0'+str(i+1)]) & (hi['BEGDAT0'+str(i+1)]<3999) &
(((hi['time_'+str(20+(j)*5)]<=hi['ENDDAT0'+str(i+1)]) & (hi['ENDDAT0'+str(i+1)]>0)) |
(hi['ENDDAT0'+str(i+1)]==0) | (hi['WIDDAT0'+str(i+1)]>0) )
,'status_'+str(20+(j)*5)]='mar'
if(np.any(hi['HOWBEG0'+str(i+1)])=='coh'):
#Substitute if actually cohabitation
hi.loc[(hi['time_'+str(20+(j)*5)]>=hi['BEGDAT0'+str(i+1)]) & (hi['BEGDAT0'+str(i+1)]<3999) &
(((hi['time_'+str(20+(j)*5)]<=hi['ENDDAT0'+str(i+1)]) & (hi['ENDDAT0'+str(i+1)]>0)) |
(hi['ENDDAT0'+str(i+1)]==0) | (hi['WIDDAT0'+str(i+1)]>0) ) &
(hi['status_'+str(20+(j)*5)]=='mar') &
(hi['HOWBEG0'+str(i+1)]=='coh') &
((hi['MARDAT0'+str(i+1)]==0) | (hi['MARDAT0'+str(i+1)]>hi['time_'+str(20+(j)*5)]))
,'status_'+str(20+(j)*5)]='coh'
#Create the variables ever cohabited and ever married
for j in range(7):
#Create the variable of ever married or cohabit
hi['everm_'+str(20+(j)*5)]=0.0
hi['everc_'+str(20+(j)*5)]=0.0
for i in range(9):
#if(np.any(hi['HOWBEG0'+str(i+1)])=='coh'):
#Get if ever cohabited
#hi.loc[((hi['everc_'+str(20+(max(j-1,0))*5)]>=0.1) | ((hi['HOWBEG0'+str(i+1)]=='coh') & (hi['time_'+str(20+(j)*5)]>=hi['BEGDAT0'+str(i+1)]))),'everc_'+str(20+(j)*5)]=1.0
hi.loc[(hi['everc_'+str(20+(max(j-1,0))*5)]>=0.1),'everc_'+str(20+(j)*5)]=1.0
try:
hi.loc[((hi['HOWBEG0'+str(i+1)]=='coh') & (hi['time_'+str(20+(j)*5)]>=hi['BEGDAT0'+str(i+1)])),'everc_'+str(20+(j)*5)]=1.0
except:
pass
#Get if ever married
hi.loc[((hi['everm_'+str(20+(max(j-1,0))*5)]>=0.1) | (hi['time_'+str(20+(j)*5)]>=hi['MARDAT0'+str(i+1)])),'everm_'+str(20+(j)*5)]=1.0
######################################
#Build employment by status in 1986
######################################
empl=hi[(hi['M2DP01']=='FEMALE') & (hi['weeks']<99)].copy()
empl['stat']='single'
empl['dist']=99999
for j in range(7):
empl.loc[np.abs(empl['time_'+str(20+(j)*5)]-86*12)<empl['dist'],'stat']=hi['status_'+str(20+(j)*5)]
##########################
#BUILD HAZARD RATES
#########################
#Hazard of Separation
hazs=list()
hazs=hazards(cohe,'sep','dury','fine',hazs,int(6/period),'SAMWT')
#Hazard of Marriage
hazm=list()
hazm=hazards(cohe,'mar','dury','fine',hazm,int(6/period),'SAMWT')
#Hazard of Divorce
hazd=list()
hazd=hazards(mare,'div','dury','fine',hazd,int(12/period),'SAMWT')
#Eventually transform Hazards pooling more years together
if transform>1:
#Divorce
hazdp=list()
pop=1
for i in range(int(12/(period*transform))):
haz1=hazd[transform*i]*pop
haz2=hazd[transform*i+1]*(pop-haz1)
hazdp=[(haz1+haz2)/pop]+hazdp
pop=pop-(haz1+haz2)
hazdp.reverse()
hazdp=np.array(hazdp).T
hazd=hazdp
#Separation and Marriage
hazsp=list()
hazmp=list()
pop=1
for i in range(int(6/(period*transform))):
hazs1=hazs[transform*i]*pop
hazm1=hazm[transform*i]*pop
hazs2=hazs[transform*i+1]*(pop-hazs1-hazm1)
hazm2=hazm[transform*i+1]*(pop-hazs1-hazm1)
hazsp=[(hazs1+hazs2)/pop]+hazsp
hazmp=[(hazm1+hazm2)/pop]+hazmp
pop=pop-(hazs1+hazs2+hazm1+hazm2)
hazsp.reverse()
hazsp=np.array(hazsp).T
hazs=hazsp
hazmp.reverse()
hazmp=np.array(hazmp).T
hazm=hazmp
########################################
#Construct share of each relationship
#######################################
mar=np.zeros(6)
coh=np.zeros(6)
emar=np.zeros(6)
ecoh=np.zeros(6)
for j in range(6):
mar[j]=np.average(hi['status_'+str(20+(j)*5)]=='mar', weights=np.array(hi['SAMWT']))
coh[j]=np.average(hi['status_'+str(20+(j)*5)]=='coh', weights=np.array(hi['SAMWT']))
emar[j]=np.average(hi['everm_'+str(20+(j)*5)], weights=np.array(hi['SAMWT']))
ecoh[j]=np.average(hi['everc_'+str(20+(j)*5)], weights=np.array(hi['SAMWT']))
#########################################
#Create the age at unilateral divorce+
#regression on the effect of unilateral divorce
###########################################
#Number of relationships for the person
hi['numerl']=0.0
#List of variables to keep
keep_var=list()
keep_var=keep_var+['numerl']+['state']+['SAMWT']
for i in range(9):
#Make sure that some relationship of order i exist
if (np.any(hi['BEGDAT0'+str(i+1)])):
#Add relationship order
hi['order'+str(i+1)]=np.nan
hi.loc[np.isnan(hi['BEGDAT0'+str(i+1)])==False,'order'+str(i+1)]=i+1
#Add number of relationships
hi.loc[np.isnan(hi['BEGDAT0'+str(i+1)])==False,'numerl']+=1.0
#Get whether the relationship started in marriage or cohabitation
hi['imar'+str(i+1)]=np.nan
hi.loc[hi['HOWBEG0'+str(i+1)]=='coh','imar'+str(i+1)]=0.0
hi.loc[hi['HOWBEG0'+str(i+1)]=='mar','imar'+str(i+1)]=1.0
#Get age at relationship
hi['iage'+str(i+1)]=np.nan
hi.loc[np.isnan(hi['BEGDAT0'+str(i+1)])==False,'iage'+str(i+1)]=round((hi['BEGDAT0'+str(i+1)]-hi['birth_month'])/12)
#Get if unilateral divorce when relationship started
hi['unid'+str(i+1)]=np.nan
hi.loc[np.isnan(hi['BEGDAT0'+str(i+1)])==False,'unid'+str(i+1)]=0.0
hi.loc[(round(hi['BEGDAT0'+str(i+1)]/12+1900)>=hi['unil']) & (hi['unil']>0.1),'unid'+str(i+1)]=1.0
#Year Realationship Started
hi['year'+str(i+1)]=np.nan
hi.loc[np.isnan(hi['BEGDAT0'+str(i+1)])==False,'year'+str(i+1)]=round(hi['BEGDAT0'+str(i+1)]/12+1900)
#Keep variables
keep_var=keep_var+['year'+str(i+1)]+['unid'+str(i+1)]+['iage'+str(i+1)]+['imar'+str(i+1)]+['order'+str(i+1)]
#New Dataset to reshape
hi2=hi[keep_var]
#Reahspe Dataset
years = ([col for col in hi2.columns if col.startswith('year')])
unids = ([col for col in hi2.columns if col.startswith('unid')])
iages = ([col for col in hi2.columns if col.startswith('iage')])
imars = ([col for col in hi2.columns if col.startswith('imar')])
order = ([col for col in hi2.columns if col.startswith('order')])
hi3 = pd.lreshape(hi2, {'year' : years,'unid' : unids,'iage' : iages,'imar' : imars,'order' : order})
#Eliminate if missing
hi3.replace([np.inf, -np.inf], np.nan)
hi3.dropna(subset=['imar','unid'])
#Regression
FE_ols = smf.wls(formula='imar ~ unid+C(iage)+C(state)+C(year)',weights=hi3['SAMWT'], data = hi3.dropna()).fit()
#FE_ols = smf.ols(formula='imar ~ unid+C(iage)+C(state)+C(year)', data = hi3.dropna()).fit()
beta_unid=FE_ols.params['unid']
#Get age at which unilateral divorced was introduced
hi['age_unid']=0.0
hi.loc[hi['unil']==0,'age_unid']=1000.0
hi.loc[hi['unil']!=0,'age_unid']=hi['unil']-hi['birth']
#Get age in the second survey
date_age=pd.read_csv('age_drop.csv')
#From hi make '-1' if law changed before the guy starts
hi.loc[hi['age_unid']<0,'age_unid']=-1
##############################
#Compute hours using the psid
################################
#Account for the survey to be retrospective
d_hrs['age']=d_hrs['age']-1.0
#Trim if hrs>2000
d_hrs.loc[d_hrs['wls']>=2000,'wls']=2000
#First keep the right birth cohorts
d_hrs['birth']=d_hrs['year']-d_hrs['age']
d_hrs=d_hrs[(d_hrs['birth']>=1940) & (d_hrs['birth']<1955)]
#Generate variables of interest
d_hrs['mar']=-1.0
d_hrs.loc[(d_hrs['mls']==1),'mar']=1.0
d_hrs.loc[(d_hrs['mls']>1) & (d_hrs['mls']<100),'mar']=0.0
#Get mean labor supply
mean_fls=np.average(d_hrs.loc[(d_hrs['age']>=20) & (d_hrs['age']<=60),'wls'])/2000
#New dataset
d_hrs2=d_hrs[(d_hrs['mar']>=0) & (d_hrs['year']>=1977)]
#Get Ratio of Female to Male FLP #23-38-53
fls_ratio=np.zeros((2))
fls_ratio[0]=np.average(d_hrs2.loc[(d_hrs2['mar']==1.0) & (d_hrs['age']>=23) &
(d_hrs['age']<=38),'wls'])/np.average(d_hrs2.loc[(d_hrs2['mar']==0.0) &
(d_hrs['age']>=23) & (d_hrs['age']<=38),'wls'])
fls_ratio[1]=np.average(d_hrs2.loc[(d_hrs2['mar']==1.0) & (d_hrs['age']>=38) &
(d_hrs['age']<=53),'wls'])/np.average(d_hrs2.loc[(d_hrs2['mar']==0.0) &
(d_hrs['age']>=38) & (d_hrs['age']<=53),'wls'])
#Get difference in male wages in marriage and cohabitation
weightm=d_hrs2.loc[(d_hrs2['mar']==1.0) & (np.isnan(d_hrs2['ln_ly'])==False),'wls']
weightc=d_hrs2.loc[(d_hrs2['mar']==0.0) & (np.isnan(d_hrs2['ln_ly'])==False),'wls']
wage_ratio=np.average(d_hrs2.loc[(d_hrs2['mar']==1.0) & (np.isnan(d_hrs2['ln_ly'])==False),'ln_ly'],weights=weightm)-np.average(d_hrs2.loc[(d_hrs2['mar']==0.0) & (np.isnan(d_hrs2['ln_ly'])==False),'ln_ly'],weights=weightc)
#######################################
#Get divorce by income using PSID
########################################
divR=np.average(d_divo.loc[(d_divo['ln_ly']>d_divo['wtmedian']),'div'])
divP=np.average(d_divo.loc[(d_divo['ln_ly']<d_divo['wtmedian']),'div'])
marR=np.average(d_divo.loc[(d_divo['ln_ly']>d_divo['wtmedian']),'mar'])
marP=np.average(d_divo.loc[(d_divo['ln_ly']<d_divo['wtmedian']),'mar'])
div_ratio=(divR/marR)/(divP/marP)
########################################
#FREQENCIES
#######################################
def CountFrequency(my_list):
# Creating an empty dictionary
freq = {}
for item in my_list:
if (item in freq):
freq[item] += 1
else:
freq[item] = 1
#for key, value in freq.items():
# print ("% d : % d"%(key, value))
return freq
#Modify age unid
freq_pc=dict()
freq_pc['male'] = CountFrequency(hi.loc[hi['M2DP01']=='MALE','age_unid'].tolist())
freq_pc['female'] = CountFrequency(hi.loc[hi['M2DP01']=='FEMALE','age_unid'].tolist())
freq_pc['share_female']=np.mean(hi['M2DP01']=='FEMALE')
#Frequencies for age in the second wave
freq_i= CountFrequency(date_age['age'].tolist())
#Frequencies for age at intervire
freq_ai=CountFrequency(hi['ageint'].tolist())
#Frequencies of agents by age at unid and gender
freq_nsfh = hi[['M2DP01','age_unid','SAMWT']]#hi.groupby(['M2DP01','age_unid'])['SAMWT'].count()
#Get distribution of types using the psid
freq_psid_tot=d_hrs[['age','unid']]
freq_psid_par=d_hrs2[['age','unid','mar']]
freq_psid_div=d_divo[['age','unid']]
#Create a dictionary for saving simulated moments
listofTuples = [("hazs" , hazs), ("hazm" , hazm),("hazd" , hazd),("emar" , emar),
("ecoh" , ecoh), ("fls_ratio" , fls_ratio),("wage_ratio" , wage_ratio),("div_ratio" , div_ratio),
("mean_fls" , mean_fls),("mar" , mar),("coh" , coh),
("freq_pc" , freq_pc), ("freq_i" , freq_i),("beta_unid" , beta_unid),("freq_ai" , freq_ai),
("freq_nsfh" , freq_nsfh),("freq_psid_tot" , freq_psid_tot),("freq_psid_par" , freq_psid_par),("freq_psid_div" , freq_psid_div)]
dic_mom=dict(listofTuples)
del hi,hi2,hi3
return dic_mom
def main():
evaluate_run = False
results_folder = os.path.join(os.getcwd(), "results_"+walk+"/" + experiment)
if not os.path.isdir(results_folder):
print(colored("Error, " + results_folder + " does not exist", 'red'))
else:
print(colored("OK, " + results_folder + " exists", 'green'))
for timeout_folder in natsorted(os.listdir(os.path.join(results_folder))):
if timeout_folder.endswith("pickle"):
continue
print(colored("Timeout folder:", 'blue'), timeout_folder)
df_kilo_timeout = pd.DataFrame()
timeout = -1
parameters = timeout_folder.split("_")
for param in parameters:
if param.startswith("timeout"):
timeout = int(param.split("#")[-1]) * 10
# print("\t timeoutR:",timeoutR)
if timeout == -1:
print(colored("\tWARNING: wrong timeout folder", 'red'))
continue
if os.path.isfile(os.path.join(results_folder, timeout_folder, "kiloLOG_timeout#" + str(timeout) + "_.pickle")):
print("Already exists ",
os.path.join(results_folder, timeout_folder, "kiloLOG_timeout#" + str(timeout) + "_.pickle"))
else:
# print(colored(
# os.path.join(results_folder, pickle_file_root"_timeout#"+str(timeout*10)+"_.pickle"),
# 'red'))
# sys.exit()
for filename in natsorted(os.listdir(os.path.join(results_folder, timeout_folder))):
filename_seed = filename.split("_")[0].split("#")[-1]
# print(filename)
if filename.endswith("areaLOG_client.tsv"):
if not os.path.getsize(os.path.join(results_folder, timeout_folder, filename)) > 0:
print(colored("\tWARNING, empty file at:" + filename, 'red'))
continue
# print('\tfilename: ', filename)
df_area_client = pd.read_csv(os.path.join(results_folder, timeout_folder, filename), sep="\t",
header=None)
if filename.endswith("areaLOG_server.tsv"):
if not os.path.getsize(os.path.join(results_folder, timeout_folder, filename)) > 0:
print(colored("\tWARNING, empty file at:" + filename, 'red'))
continue
# print('\tfilename: ', filename)
df_area_server = pd.read_csv(os.path.join(results_folder, timeout_folder, filename), sep="\t",
header=None)
if filename.endswith("kiloLOG_client.tsv"):
if not os.path.getsize(os.path.join(results_folder, timeout_folder, filename)) > 0:
print(colored("\tWARNING, empty file at:" + filename, 'red'))
continue
# print('\tfilename: ', filename)
df_kilo_client = pd.read_csv(os.path.join(results_folder, timeout_folder, filename), sep="\t",
header=None)
if filename.endswith("kiloLOG_server.tsv"):
if not os.path.getsize(os.path.join(results_folder, timeout_folder, filename)) > 0:
print(colored("\tWARNING, empty file at:" + filename, 'red'))
continue
# print('\tfilename: ', filename, end='\n')
df_kilo_server = pd.read_csv(os.path.join(results_folder, timeout_folder, filename), sep="\t",
header=None)
evaluate_run = True
if evaluate_run:
print(colored("\tEvaluating run:" + filename_seed, 'blue'))
'''Kilo log part'''
if len(df_kilo_client.columns) > 145:
# print("Cutting null elements in client kilo df")
df_kilo_client.drop(df_kilo_client.columns[len(df_kilo_client.columns) - 1], axis=1, inplace=True)
if len(df_kilo_server.columns) > 145:
# print("Cutting null elements in server kilo df")
df_kilo_server.drop(df_kilo_server.columns[len(df_kilo_server.columns) - 1], axis=1, inplace=True)
col_kilo_labels = ['time']
for i in range(0, len(df_kilo_server.columns) - 1, 6):
# print(i,end=", ")
col_kilo_labels += ['id' + str(i // 6), 'state' + str(i // 6), 'posx' + str(i // 6),
'posy' + str(i // 6),
'ori' + str(i // 6), 'same_state' + str(i // 6)]
col_kilo_to_drop = []
for i in range((len(df_kilo_server.columns) - 1) // 6):
# print(i,end=", ")
col_kilo_to_drop += ['same_state' + str(i)]
df_kilo_server.columns = col_kilo_labels
df_kilo_client.columns = col_kilo_labels
df_kilo_server = df_kilo_server.drop(col_kilo_to_drop, axis=1)
df_kilo_client = df_kilo_client.drop(col_kilo_to_drop, axis=1)
'''Area LOG part'''
col_area_labels = ['time']
for i in range(0, len(df_area_server.columns) - 2, 6):
# print(i, end=", ")
col_area_labels += ['id' + str(i // 6), 'posx' + str(i // 6), 'posy' + str(i // 6),
'color' + str(i // 6),
'completed' + str(i // 6), 'contained' + str(i // 6)]
# Remove last empty col and assign labels to df_area_server
if len(df_area_server.columns) > 49:
# print("Cutting null elements in area server df")
df_area_server.drop(df_area_server.columns[len(df_area_server.columns) - 1], axis=1, inplace=True)
df_area_server.columns = col_area_labels
# First df_area_client row contains garbage
# so is substituted with the second row except for the time,
# then remove Nan values in [:,49:]
if len(df_area_client.columns) > 49:
# print("Cutting null elements in area client df")
df_area_client.loc[0, 1:] = df_area_client.loc[1, 1:]
df_area_client = df_area_client.drop(np.arange(49, len(df_area_client.columns)), axis=1)
df_area_client.columns = col_area_labels
area_pos_label = []
for i in range(num_areas):
area_pos_label += ["posx" + str(i)]
area_pos_label += ["posy" + str(i)]
areas_pos = df_area_client[area_pos_label].iloc[0, :].values
# print(areas_pos)
areas_pos = areas_pos.reshape(-1, 2)
color_list = ["color" + str(i) for i in range(num_areas)]
df_area3_s = df_area_server.iloc[:1, :][color_list]
df_area3_c = df_area_client.iloc[:1, :][color_list]
for i, idx in enumerate(range(1, len(df_area3_c.columns) * 2, 2)):
# print(i, ' ', idx)
df_area3_c.insert(loc=idx, column='other_col' + str(i), value=df_area3_s.iloc[0][i])
client = [col for col in df_area3_c.columns if 'color' in col]
server = [col for col in df_area3_c.columns if 'other_col' in col]
df_area_colors = pd.lreshape(df_area3_c, {'color_client': client, 'color_server': server})
area_type = []
for area in df_area_colors.values:
if area[0] == 0 and area[1] == 0:
area_type += ['BB']
if area[0] == 0 and area[1] == 1:
area_type += ['BR']
if area[0] == 1 and area[1] == 0:
area_type += ['RB']
if area[0] == 1 and area[1] == 1:
area_type += ['RR']
df_area_colors.insert(loc=2, column='area_type', value=area_type)
'''Post process server'''
for i_c, kilo_id in enumerate(np.arange(1, len(df_kilo_server.columns), 5)):
# print(colored("kilo_id:" + str((kilo_id - 1) // 5), 'blue'))
# print(df_kilo_client.iloc[:20, kilo_id+2:kilo_id+4].values, end='\n\n')
kilo_pos = df_kilo_server.iloc[:, kilo_id + i_c + 2:kilo_id + i_c + 4].values
# print(kilo_pos)
in_area = np.empty(kilo_pos.shape[0], dtype=int)
in_area.fill(-1)
for area_idx, area_pos in enumerate(areas_pos):
# print(area_idx, ' ', area_pos)
dist = np.linalg.norm(kilo_pos - area_pos, axis=1)
# print(dist, end='\n\n')
in_area = np.where(dist < area_threshold, df_area_colors.iloc[area_idx][-1][::-1], in_area)
# in_area = np.where(in_area == -1, np.NaN, in_area)
# print(in_area)
df_kilo_server.insert(loc=int(kilo_id + i_c + 2), column='area_type' + str(i_c), value=in_area)
'''Post process client'''
for i_s, kilo_id in enumerate(np.arange(1, len(df_kilo_client.columns), 5)):
# print(colored("kilo_id:" + str((kilo_id - 1) // 5), 'blue'))
# print(df_kilo_client.iloc[:20, kilo_id+2:kilo_id+4].values, end='\n\n')
kilo_pos = df_kilo_client.iloc[:, kilo_id + i_s + 2:kilo_id + i_s + 4].values
# print(kilo_pos)
in_area = np.empty(kilo_pos.shape[0], dtype=int)
in_area.fill(-1)
for area_idx, area_pos in enumerate(areas_pos):
# print(area_idx,' ', area_pos)
dist = np.linalg.norm(kilo_pos - area_pos, axis=1)
# print(dist, end='\n\n')
in_area = np.where(dist < area_threshold, df_area_colors.iloc[area_idx][-1], in_area)
# in_area = np.where(in_area == -1, np.NaN, in_area)
# print(in_area)
df_kilo_client.insert(loc=int(kilo_id + i_s + 2), column='area_type' + str(i_s), value=in_area)
df_kilo_single_run = df_kilo_client.join(df_kilo_server, lsuffix='_c', rsuffix='_s')
df_kilo_single_run = df_kilo_single_run.set_index(df_kilo_single_run.index.astype(str) + '_' + filename_seed)
df_kilo_timeout = df_kilo_timeout.append(df_kilo_single_run)
evaluate_run = False
'''Save pickle file'''
df_kilo_timeout.to_pickle(os.path.join(results_folder, timeout_folder, "kiloLOG_timeout#"+str(timeout)+"_.pickle"))
print("Saving at: ", os.path.join(results_folder, timeout_folder, "kiloLOG_timeout#"+str(timeout)+"_.pickle"))
print("Changing dir")
