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 extract(corp_code: str,
bgn_de: str,
end_de: str = None,
fs_tp: Tuple[str] = ('bs', 'is', 'cis', 'cf'),
separate: bool = False,
report_tp: str = 'annual',
lang: str = 'ko',
separator: bool = True) -> FinancialStatement:
"""
재무제표 검색
Parameters
----------
corp_code: str
공시대상회사의 고유번호(8자리)
bgn_de: str
검색 시작일자(YYYYMMDD)
end_de: str, optional
검색 종료일자(YYYYMMDD)
fs_tp: tuple of str, optional
'bs' 재무상태표, 'is' 손익계산서, 'cis' 포괄손익계산서, 'cf' 현금흐름표
separate: bool, optional
개별재무제표 여부
report_tp: str, optional
'annual' 1년, 'half' 반기, 'quarter' 분기
lang: str, optional
'ko' 한글, 'en' 영문
separator: bool, optional
1000단위 구분자 표시 여부
Returns
-------
FinancialStatement
제무제표 검색 결과
"""
if is_notebook():
from tqdm import tqdm_notebook as tqdm
else:
from tqdm import tqdm
import dart_fss as dart
dart.utils.spinner.spinner_enable = False
reports = search_annual_report(corp_code=corp_code, bgn_de=bgn_de, end_de=end_de, separate=separate)
try:
length = len(reports)
statements = None
label_df = None
# Spinner disable
for _ in tqdm(range(length), desc='Annual reports', unit='report'):
report = reports.pop(0)
if statements is None:
statements = analyze_report(report=report,
fs_tp=fs_tp,
separate=separate,
lang=lang,
separator=separator)
if separate is False and all([statements[tp] is None for tp in statements]):
raise NotFoundConsolidated('Could not find consolidated financial statements')
# initialize label dictionary
label_df = init_label(statements, fs_tp=fs_tp)
else:
nstatements = analyze_report(report=report,
fs_tp=fs_tp,
separate=separate,
lang=lang,
separator=separator)
statements, label_df = merge_fs(statements, nstatements, fs_tp=fs_tp, label_df=label_df)
if str_compare(report_tp, 'half') or str_compare(report_tp, 'quarter'):
half = search_filings(corp_code=corp_code, bgn_de=bgn_de, end_de=end_de,
pblntf_detail_ty='A002', page_count=100, last_reprt_at='Y')
length = len(half)
for _ in tqdm(range(length), desc='Semiannual reports', unit='report'):
report = half.pop(0)
nstatements = analyze_report(report=report,
fs_tp=fs_tp,
separate=separate,
lang=lang,
separator=separator)
statements, label_df = merge_fs(statements, nstatements, fs_tp=fs_tp, label_df=label_df)
if str_compare(report_tp, 'quarter'):
quarter = search_filings(corp_code=corp_code, bgn_de=bgn_de, end_de=end_de,
pblntf_detail_ty='A003', page_count=100, last_reprt_at='Y')
length = len(quarter)
for _ in tqdm(range(length), desc='Quarterly report', unit='report'):
report = quarter.pop(0)
nstatements = analyze_report(report=report,
fs_tp=fs_tp,
separate=separate,
lang=lang,
separator=separator)
statements, label_df = merge_fs(statements, nstatements, fs_tp=fs_tp, label_df=label_df)
statements = drop_empty_columns(statements)
label_df = drop_empty_columns(label_df)
statements = sorting_columns(statements)
label_df = sorting_columns(label_df)
info = {
'corp_code': corp_code,
'bgn_de': bgn_de,
'end_de': end_de,
'separate': separate,
'report_tp': report_tp,
'lang': lang,
'separator': separator
}
# Spinner enable
dart.utils.spinner.spinner_enable = True
return FinancialStatement(statements, label_df, info)
except Exception as e:
msg = 'An error occurred while fetching or analyzing {}.'.format(report.to_dict())
e.args = (*e.args, msg, )
raise e
finally:
dart.utils.spinner.spinner_enable = True
def extract(corp_code: str,
bgn_de: str,
end_de: str = None,
fs_tp: Tuple[str] = ('bs', 'is', 'cis', 'cf'),
separate: bool = False,
report_tp: Union[str, List[str]] = 'annual',
lang: str = 'ko',
separator: bool = True,
dataset: str = 'xbrl') -> FinancialStatement:
"""
재무제표 검색
Parameters
----------
corp_code: str
공시대상회사의 고유번호(8자리)
bgn_de: str
검색 시작일자(YYYYMMDD)
end_de: str, optional
검색 종료일자(YYYYMMDD)
fs_tp: tuple of str, optional
'bs' 재무상태표, 'is' 손익계산서, 'cis' 포괄손익계산서, 'cf' 현금흐름표
separate: bool, optional
개별재무제표 여부
report_tp: str or list, optional
str: 'annual' 연간, 'half' 연간 + 반기, 'quarter' 연간 + 반기 + 분기
list: ['annual'] : 연간, ['half']: 반기, ['quarter'] 분기, ['annual', 'half']: 연간 + 반기
['annual', 'quarter']: 연간 + 분기, ['half', 'quarter']: 반기 + 분기, ['annual', 'half', 'quarter']: 연간 + 반기 + 분기
lang: str, optional
'ko' 한글, 'en' 영문
separator: bool, optional
1000단위 구분자 표시 여부
dataset: str, optional
'xbrl': xbrl 파일 우선 데이터 추출, 'web': web page 우선 데이터 추출(default: 'xbrl')
Returns
-------
FinancialStatement
제무제표 검색 결과
"""
if is_notebook():
from tqdm import tqdm_notebook as tqdm
else:
from tqdm import tqdm
if dataset not in ['xbrl', 'web']:
raise ValueError('invalid dataset type: only xbrl or web are allowed')
all_report_tp = ('annual', 'half', 'quarter')
all_report_name = ('Annual', 'Semiannual', 'Quarterly')
all_pblntf_detail_ty = ('A001', 'A002', 'A003')
def check_report_tp(req_tp, tp):
if isinstance(req_tp, str):
index = all_report_tp.index(req_tp) + 1
if tp in all_report_tp[:index]:
return True
else:
return False
elif isinstance(req_tp, list) and tp in req_tp:
return True
else:
return False
# Spinner disable
import dart_fss as dart
dart.utils.spinner.spinner_enable = False
statements = None
label_df = None
report = None
try:
for idx, tp in enumerate(all_report_tp):
if check_report_tp(report_tp, tp):
if tp == 'annual':
reports = search_annual_report(corp_code=corp_code, bgn_de=bgn_de, end_de=end_de, separate=separate)
else:
reports = search_filings(corp_code=corp_code, bgn_de=bgn_de, end_de=end_de,
pblntf_detail_ty=all_pblntf_detail_ty[idx], page_count=100, last_reprt_at='Y')
if reports == 0:
continue
length = len(reports)
for _ in tqdm(range(length), desc='{} reports'.format(all_report_name[idx]), unit='report'):
report = reports.pop(0)
if statements is None:
statements = analyze_report(report=report,
fs_tp=fs_tp,
separate=separate,
lang=lang,
separator=separator)
if statements is None:
warnings_text = 'Unable to extract financial statements: {}.'.format(report.to_dict())
warnings.warn(warnings_text, RuntimeWarning)
else:
if separate is False and all([statements[tp] is None for tp in statements]):
raise NotFoundConsolidated('Could not find consolidated financial statements')
# initialize label dictionary
label_df = init_label(statements, fs_tp=fs_tp)
else:
nstatements = analyze_report(report=report,
fs_tp=fs_tp,
separate=separate,
lang=lang,
separator=separator,
dataset=dataset)
if nstatements is None:
warnings_text = 'Unable to extract financial statements: {}.'.format(report.to_dict())
warnings.warn(warnings_text, RuntimeWarning)
else:
statements, label_df = merge_fs(statements, nstatements, fs_tp=fs_tp, label_df=label_df)
# Spinner enable
dart.utils.spinner.spinner_enable = True
if separate is False and (statements is None or all([statements[tp] is None for tp in statements])):
raise NotFoundConsolidated('Could not find consolidated financial statements')
statements = drop_empty_columns(statements)
label_df = drop_empty_columns(label_df)
statements = sorting_columns(statements)
label_df = sorting_columns(label_df)
info = {
'corp_code': corp_code,
'bgn_de': bgn_de,
'end_de': end_de,
'separate': separate,
'report_tp': report_tp,
'lang': lang,
'separator': separator
}
return FinancialStatement(statements, label_df, info)
except Exception as e:
if report is not None:
msg = 'An error occurred while fetching or analyzing {}.'.format(report.to_dict())
else:
msg = 'Unexpected Error'
e.args = (*e.args, msg, )
raise e
finally:
dart.utils.spinner.spinner_enable = True
def extract(corp_code: str,
bgn_de: str,
end_de: str = None,
fs_tp: Tuple[str] = ('bs', 'is', 'cis', 'cf'),
separate: bool = False,
report_tp: str = 'annual',
lang: str = 'ko',
separator: bool = True) -> FinancialStatement:
"""
재무제표 검색
Parameters
----------
corp_code: str
공시대상회사의 고유번호(8자리)
bgn_de: str
검색 시작일자(YYYYMMDD)
end_de: str, optional
검색 종료일자(YYYYMMDD)
fs_tp: tuple of str, optional
'bs' 재무상태표, 'is' 손익계산서, 'cis' 포괄손익계산서, 'cf' 현금흐름표
separate: bool, optional
개별재무제표 여부
report_tp: str, optional
'annual' 1년, 'half' 반기, 'quarter' 분기
lang: str, optional
'ko' 한글, 'en' 영문
separator: bool, optional
1000단위 구분자 표시 여부
Returns
-------
FinancialStatement
제무제표 검색 결과
"""
if is_notebook():
from tqdm import tqdm_notebook as tqdm
else:
from tqdm import tqdm
# 재무제표 검색 결과
statements = None
reports = []
try:
# 사업보고서 검색(최종보고서)
reports = search_filings(corp_code=corp_code,
bgn_de=bgn_de,
end_de=end_de,
pblntf_detail_ty='A001',
page_count=100,
last_reprt_at='Y')
except NoDataReceived:
# 감사보고서 검색
if separate:
pblntf_detail_ty = 'F001'
else:
pblntf_detail_ty = 'F002'
reports = search_filings(corp_code=corp_code,
bgn_de=bgn_de,
end_de=end_de,
pblntf_detail_ty=pblntf_detail_ty,
page_count=100,
last_reprt_at='Y')
finally:
if len(reports) == 0:
raise RuntimeError('Could not find an annual report')
next_index = 0
for idx, _ in enumerate(reports):
# 가장 최근 보고서의 경우 XBRL 파일을 이용하여 재무제표 검색
latest_report = reports[idx]
latest_xbrl = latest_report.xbrl
# XBRL 파일이 존재할 때
if latest_xbrl is not None:
if separate is False and not latest_xbrl.exist_consolidated():
raise NotFoundConsolidated(
'Could not find consolidated financial statements')
# XBRL 정보를 이용하여 재무제표 정보 초기화
analyzed_results = analyze_xbrl(latest_report,
fs_tp=fs_tp,
separate=separate,
lang=lang,
show_abstract=False,
show_class=True,
show_depth=10,
show_concept=True,
separator=separator)
statements = copy.deepcopy(analyzed_results)
else:
statements = analyze_html(latest_report,
fs_tp=fs_tp,
separate=separate,
lang=lang)
# Report 에 재무제표 정보 없이 수정 사항만 기록된 경우 다음 리포트 검색
if statements is not None:
next_index = idx + 1
break
if separate is False and all(
[statements[tp] is None for tp in statements]):
raise NotFoundConsolidated(
'Could not find consolidated financial statements')
label_df = None
for report in tqdm(reports[next_index:],
desc='Annual reports',
unit='report'):
statements, label_df = merge_fs(statements,
label_df,
report,
fs_tp=fs_tp,
separate=separate,
lang=lang)
if str_compare(report_tp, 'half') or str_compare(report_tp, 'quarter'):
half = search_filings(corp_code=corp_code,
bgn_de=bgn_de,
end_de=end_de,
pblntf_detail_ty='A002',
page_count=100,
last_reprt_at='Y')
for report in tqdm(half, desc='Semiannual reports', unit='report'):
statements, label_df = merge_fs(statements,
label_df,
report,
fs_tp=fs_tp,
separate=separate,
lang=lang)
if str_compare(report_tp, 'quarter'):
quarter = search_filings(corp_code=corp_code,
bgn_de=bgn_de,
end_de=end_de,
pblntf_detail_ty='A003',
page_count=100,
last_reprt_at='Y')
for report in tqdm(quarter, desc='Quarterly report',
unit='report'):
statements, label_df = merge_fs(statements,
label_df,
report,
fs_tp=fs_tp,
separate=separate,
lang=lang)
statements = drop_empty_columns(statements)
label_df = drop_empty_columns(label_df)
statements = sorting_columns(statements)
label_df = sorting_columns(label_df)
info = {
'corp_code': corp_code,
'bgn_de': bgn_de,
'end_de': end_de,
'separate': separate,
'report_tp': report_tp,
'lang': lang,
'separator': separator
}
return FinancialStatement(statements, label_df, info)