def __init__(self,
cik_lookup,
filing_type,
start_date=None,
end_date=datetime.datetime.today(),
client=None,
**kwargs):
self._start_date = start_date
self._end_date = end_date
self._accession_numbers = []
if not isinstance(filing_type, FilingType):
raise FilingTypeError
self._filing_type = filing_type
# make CIKLookup object for users if not given
if not isinstance(cik_lookup, CIKLookup):
cik_lookup = CIKLookup(cik_lookup)
self._cik_lookup = cik_lookup
self._params = {
'action': 'getcompany',
'dateb': sanitize_date(self.end_date),
'output': 'xml',
'owner': 'include',
'start': 0,
'type': self.filing_type.value
}
if kwargs.get('count') is not None:
self._params['count'] = kwargs.get('count')
if start_date is not None:
self._params['datea'] = sanitize_date(start_date)
# Make default client NetworkClient and pass in kwargs
self._client = client if client is not None else NetworkClient(
**kwargs)
def params(self):
""":obj:`dict`: Parameters to include in requests."""
if self.start_date:
self._params["datea"] = sanitize_date(self.start_date)
else:
self._params.pop("datea", None) # if no start date, make sure it isn't in params
if self.end_date:
self._params["dateb"] = sanitize_date(self.end_date)
else:
self._params.pop("dateb", None) # if no end date, make sure it isn't in params
self._params["ownership"] = self.ownership
return self._params
def _fetch_report(self, company_code, cik, priorto, count, filing_type):
"""Fetch filings.
Args:
company_code (str): Code used to help find company filings.
Often the company's ticker is used.
cik (Union[str, int]): Central Index Key assigned by SEC.
See https://www.sec.gov/edgar/searchedgar/cik.htm to search for
a company's CIK.
priorto (Union[str, datetime.datetime]): Most recent report to consider.
Must be in form 'YYYYMMDD' or
valid ``datetime.datetime`` object.
filing_type (str): Choose from list of valid filing types.
Includes '10-Q', '10-K', '8-K', '13-F', 'SD'.
Returns:
None
"""
priorto = sanitize_date(priorto)
cik = self._check_cik(cik)
self._make_directory(company_code, cik, priorto, filing_type)
# generate the url to crawl
base_url = "http://www.sec.gov/cgi-bin/browse-edgar"
params = {
'action': 'getcompany',
'owner': 'exclude',
'output': 'xml',
'CIK': cik,
'type': filing_type,
'dateb': priorto,
'count': count
}
print("started {filing_type} {company_code}".format(
filing_type=filing_type, company_code=company_code))
r = requests.get(base_url, params=params)
if r.status_code == 200:
data = r.text
# get doc list data
docs = self._create_document_list(data)
try:
self._save_in_directory(company_code, cik, priorto,
filing_type, docs)
except Exception as e:
print(str(e)) # Need to use str for Python 2.5
else:
raise EDGARQueryError(r.status_code)
print("Successfully downloaded all the files")
def end_date(self, val):
self._end_date = val
self._params['dateb'] = sanitize_date(val)
def start_date(self, val):
if val is not None:
self._start_date = val
self._params['datea'] = sanitize_date(val)
else:
self._start_date = None
def end_date(self, val):
sanitize_date(val) # make sure end date is valid
self._end_date = val
def start_date(self, val):
sanitize_date(val) # make sure start date is valid
self._start_date = val
def start_date(self, val):
self._start_date = val
self._params['datea'] = sanitize_date(val)
def test_good_formats_datetime(self, dt_date, expected):
assert sanitize_date(dt_date) == expected
def test_good_formats_no_change(self, good_date):
"""Tests formats that should not change from what is given. """
assert sanitize_date(good_date) == good_date
def test_bad_date_formats(self, bad_date):
with pytest.raises(TypeError):
sanitize_date(bad_date)
def end_date(self, val):
self._params["dateb"] = sanitize_date(val)
self._end_date = val
def start_date(self, val):
if val is not None:
self._params["datea"] = sanitize_date(val)
self._start_date = val
else:
self._start_date = None