def run_test(self):
fs = self.corp.extract_fs(bgn_de=self.bgn_de,
separate=self.separate,
report_tp=self.report_tp)
for test in self.test_set:
tp = test['fs_tp']
date = test['date']
column = test['column']
item = test['item']
expected = test['expected']
df = fs[tp]
date_column = find_all_columns(df=df, query=date)[0]
label_column = find_all_columns(df=df, query=column)[0]
actual = None
for idx in range(len(df)):
text = df[label_column].iloc[idx].replace(' ', '')
if str_compare(text, item):
actual = df[date_column].iloc[idx]
if actual != expected:
pytest.fail("Test failed: corp_code='{}', ".format(
self.corp.corp_code) +
"corp_name='{}', fs_tp='{}', ".format(
self.corp.corp_name, tp) +
"start_dt='{}', report_tp='{}', ".format(
self.bgn_de, fs.info['report_tp']) +
"date='{}', column='{}',".format(date, column) +
"item='{}', actual='{}', expected='{}'".format(
item, actual, expected))
def _extract_dataset(self, reports: List[Report]):
"""
Report에 포함된 XBRL 파일에서 Concept_id 와 Label 값 추출
Parameters
----------
reports: list of Report
추출할 Report 리스트
"""
if is_notebook():
from tqdm import tqdm_notebook as tqdm
else:
from tqdm import tqdm
dataset = []
for report in tqdm(reports,
desc='Extracting concept_id and label_ko',
unit='report'):
df_fs = analyze_xbrl(report)
if df_fs is None:
continue
for tp in df_fs:
df = df_fs[tp]
if df is not None:
concept_column = find_all_columns(df, 'concept_id')[0]
label_ko_column = find_all_columns(df, 'label_ko')[0]
for idx in range(len(df)):
concept_id = df[concept_column].iloc[idx]
label_ko = df[label_ko_column].iloc[idx]
if concept_id and label_ko:
try:
label = self.extract_nouns(label_ko)
dataset.append((concept_id, label))
except BaseException:
continue
self._dataset = dataset
def compare_df_and_ndf_cnn(
column: Tuple[Union[str, Tuple[str]]], df: DataFrame, ndf: DataFrame,
ldf: DataFrame, ndata: List[Union[float, str, None]],
nlabels: List[str]) -> Tuple[List[Union[float, str]], List[str]]:
"""
Convolutional neural network 를 시용하여 데이터를 검색하는 함수
Parameters
----------
column: tuple
추가할 column Name
df: dict of { str: DataFrame }
데이터를 추가할 DataFrame, 추출된 결과값이 누적된 DataFrame
ndf: dict of { str: DataFrame }
데이터를 검색할 DataFrame, Report에서 추출한 새로운 DataFrame
ndata: list of float
추가할 column의 데이터 리스트
nlabels: list of str
추가할 column의 label 리스트
Returns
-------
tuple of list
추가할 column의 데이터 리스트, 추가할 column의 label 리스트
"""
# CNN 처리시 사용
concept_none_data = {}
df_label_column = find_all_columns(df, 'label_ko')[0]
is_concept = True
df_concept_column = find_all_columns(df, 'concept_id')
if len(df_concept_column) == 0:
is_concept = False
else:
df_concept_column = df_concept_column[0]
ndf_label_column = find_all_columns(ndf, 'label_ko')[0]
for idx, value in enumerate(ndata):
if isinstance(value, str):
pass
elif value is None:
pass
elif math.isnan(value):
pass
else:
continue
label = df[df_label_column].iloc[idx]
label = re.sub(r'\s+', '', label)
label = extract_account_title(label)
if is_concept:
concept_id = df[df_concept_column].iloc[idx]
else:
concept_id = guess_concept_id(label)
if concept_id is not None:
concept_none_data[concept_id] = idx
matched = []
used = []
for idx in range(len(ndf)):
if idx in matched:
continue
label = extract_account_title(ndf[ndf_label_column].iloc[idx])
concept_id = guess_concept_id(label)
index = concept_none_data.get(concept_id)
if index is not None and index not in used:
value = ndf[column].iloc[idx]
if isinstance(value, str):
pass
else:
used.append(index)
matched.append(idx)
ndata[index] = value
nlabels[index] = label
return ndata, nlabels
def test_fs_show_depth(fs_report):
df = fs_report.show('bs', show_depth=1)
columns = find_all_columns(df, 'class')
actual = len(columns)
expected = 2
assert actual == expected
def test_fs_concept_false(fs_report):
df = fs_report.show('bs', show_concept=False)
columns = find_all_columns(df, 'concept')
actual = len(columns)
expected = 0
assert actual == expected
def show(self,
tp,
show_class: bool = True,
show_depth: int = 10,
show_concept: bool = True) -> DataFrame:
"""
재무제표 정보를 표시해주는 Method
Parameters
----------
tp: str
표시할 재무제표 타입: 'fs' 재무상태표, 'is' 손익계산서, 'ci' 포괄손익계산서, 'cf' 현금흐름표
show_class: bool
class 표시 여부
show_depth: bool
표시할 class의 깊이
show_concept: bool
concept_id 표시 여부
Returns
-------
DataFrame
재무제표
"""
from dart_fss.fs.extract import find_all_columns
df = self._statements[tp]
if df is None:
return df
class_columns = find_all_columns(df, 'class')
if show_class is False:
ncolumns = []
columns = df.columns.tolist()
for column in columns:
if column not in class_columns:
ncolumns.append(column)
df = df[ncolumns]
else:
drop_rows = []
columns = df.columns.tolist()
cdf = df[class_columns]
for idx in range(len(cdf)):
for class_idx, item in enumerate(cdf.iloc[idx]):
if class_idx > show_depth and item is not None:
drop_rows.append(idx)
ncolumns = []
for column in columns:
if column not in class_columns[show_depth + 1:]:
ncolumns.append(column)
df = df[ncolumns].drop(drop_rows)
if show_concept is False:
concept_colmuns = find_all_columns(df, 'concept_id')
if len(concept_colmuns) == 1:
ncolumns = []
columns = df.columns.tolist()
for column in columns:
if column not in concept_colmuns:
ncolumns.append(column)
df = df[ncolumns]
return df
def test_xbrl_get_author_information(samsung_xbrl):
author = samsung_xbrl.get_author_information()
column = find_all_columns(author, '공시담당자')[0]
actual = author[column][3]
expected = '031-277-7227'
assert actual == expected