def get_filings(companyCode, date='20200101', cik=None, count=100):
if cik is None:
with open('company_list.txt', 'r') as f:
for line in f:
if companyCode in line:
line_arr = line.rstrip().split(' ')
cik = line_arr[-1]
if cik is None:
print("cik not provided and not found in list. please try again.")
return
# create object
seccrawler = SecCrawler()
seccrawler.filing_10K(str(companyCode), str(cik), str(date), str(count))
dest_dir = companyCode + "/"
src_dir = dest_dir + cik + "/10-K/"
years_downloaded = []
for old_filename in os.listdir(src_dir):
parts = old_filename.split('-')
old_year = parts[1]
if int(old_year) > 50:
new_year = '19' + old_year
else:
new_year = '20' + old_year
years_downloaded.append(new_year)
os.rename(src_dir + old_filename, dest_dir + companyCode + '_10K_' + new_year + '.txt')
shutil.rmtree(dest_dir + cik + '/')
# create object
seccrawler = SecCrawler()
seccrawler.filing_10Q(str(companyCode), str(cik), str(date), str(count))
dest_dir = companyCode + "/"
src_dir = dest_dir + cik + "/10-Q/"
years_downloaded = []
for old_filename in os.listdir(src_dir):
parts = old_filename.split('-')
old_year = parts[1]
if int(old_year) > 20:
new_year = '19' + old_year
else:
new_year = '20' + old_year
years_downloaded.append(new_year)
os.rename(src_dir + old_filename, dest_dir + companyCode + '_10Q_' + new_year + '.txt')
shutil.rmtree(dest_dir + cik + '/')
return years_downloaded
def get_filings(cik, ticker):
t1 = time.time()
# create object
seccrawler = SecCrawler()
companyCode = ticker
cik = cik
date = '20170101'
count = '10'
seccrawler.filing_10K(str(companyCode), str(cik), str(date), str(count))
t2 = time.time()
print("Total Time taken: "),
print(t2 - t1)
def get_filings(cik, ticker):
t1 = time.time()
# create object
seccrawler = SecCrawler()
companyCode = ticker # company code for apple
cik = cik # cik code for apple
date = '20170101' # date from which filings should be downloaded
count = '10' # no of filings
seccrawler.filing_10K(str(companyCode), str(cik), str(date), str(count))
t2 = time.time()
print("Total Time taken: "),
print(t2 - t1)
def getfilings():
t1 = time.time()
# create object
seccrawler = SecCrawler()
companyCode = 'AAPL' # compnay ticker symbol for Apple
cik = '0000320193' # cik code for Apple
date = '20010101' # date from which filings should be downloaded (01/01/2001 in this case)
count = '10' # the number of filings
seccrawler.filing_10Q(str(companyCode), str(cik), str(date), str(count))
seccrawler.filing_10K(str(companyCode), str(cik), str(date), str(count))
seccrawler.filing_8K(str(companyCode), str(cik), str(date), str(count))
seccrawler.filing_13F(str(companyCode), str(cik), str(date), str(count))
t2 = time.time()
print("Total time taken: ")
print(t2 - t1)
def get_all_filings(self, deadline):
''' get the SEC filings for all companies in US stock universe'''
t1 = time.time()
# create object
seccrawler = SecCrawler()
companyCode = 'AAPL' # company code for apple
cik = '0000320193' # cik code for apple
date = '20160101' # date from which filings should be downloaded
count = '100' # no of filings
seccrawler.filing_10Q(str(companyCode), str(cik), str(date), str(count))
seccrawler.filing_10K(str(companyCode), str(cik), str(date), str(count))
seccrawler.filing_8K(str(companyCode), str(cik), str(date), str(count))
seccrawler.filing_13F(str(companyCode), str(cik), str(date), str(count))
t2 = time.time()
print "Total Time taken: " + str(t2-t1),
def get_filings():
t1 = time.time()
# create object
seccrawler = SecCrawler()
companyCode = "AAPL" # company code for apple
cik = "0000320193" # cik code for apple
date = "20010101" # date from which filings should be downloaded
count = "10" # no of filings
seccrawler.filing_10Q(str(companyCode), str(cik), str(date), str(count))
seccrawler.filing_10K(str(companyCode), str(cik), str(date), str(count))
seccrawler.filing_8K(str(companyCode), str(cik), str(date), str(count))
seccrawler.filing_13F(str(companyCode), str(cik), str(date), str(count))
t2 = time.time()
print
"Total Time taken: ",
print(t2 - t1)
def get_filings():
t1 = time.time()
DEFAULT_DATA_PATH = os.path.join(os.path.abspath(os.getcwd()), 'output')
# create object
seccrawler = SecCrawler()
companyCode = 'AAPL' # company code for apple
cik = '0000320193' # cik code for apple
date = '20010101' # date from which filings should be downloaded
count = '10' # no of filings
seccrawler.filing_10Q(str(companyCode), str(cik), str(date), str(count))
seccrawler.filing_10K(str(companyCode), str(cik), str(date), str(count))
seccrawler.filing_8K(str(companyCode), str(cik), str(date), str(count))
seccrawler.filing_13F(str(companyCode), str(cik), str(date), str(count))
t2 = time.time()
print("Total Time taken: "),
print(t2 - t1)
def get_filings(): t1 = time.time() # create object seccrawler = SecCrawler() companyCode = 'AAPL' # company code for apple cik = '0000320193' # cik code for apple date = '20010101' # date from which filings should be downloaded count = '10' # no of filings seccrawler.filing_10Q(str(companyCode), str(cik), str(date), str(count)) seccrawler.filing_10K(str(companyCode), str(cik), str(date), str(count)) seccrawler.filing_8K(str(companyCode), str(cik), str(date), str(count)) seccrawler.filing_13F(str(companyCode), str(cik), str(date), str(count)) t2 = time.time() print "Total Time taken: ", print (t2-t1) if __name__ == '__main__': get_filings()
"""
#%%
import re
from SECEdgar.crawler import SecCrawler
#%% GEt names of companies
sector_code = '10'
list_companies = []
with open('tic_company_gics.txt') as file:
for line in file:
sector = line[-3:-1]
if sector == sector_code:
name_company = re.search('[\w\.]+', line).group()#[:-1]
list_companies.append(name_company)
list_companies = list(set(list(list_companies)))
#%% Download the 10K
number_of_companies = 50
date_retrieve = '20150101'
number_10K = 5
secCrawler = SecCrawler()
for company in list_companies[0:number_of_companies]:
secCrawler.filing_10K(company, date_retrieve, number_10K)
cik = ''
results = CIK_RE.findall(get(URL.format(name)).content)
if len(results):
cik = str(results[0])
return cik
list_companies = []
with open('tic_company_gics.txt') as file:
for line in file:
sector = line[-3:-1]
if sector == sector_code:
name_company = re.search('[\w\.]+', line).group()#[:-1]
list_companies.append(name_company)
list_companies = list(set(list(list_companies)))
#%% GEt SIC code for each company
list_companies_cik = []
for company in list_companies:
cik = get_cik(company)
if len(cik) > 0:
list_companies_cik.append((company, cik))
#%% Get 10K
secCrawler = SecCrawler()
for name, cik in list_companies_cik:
secCrawler.filing_10K(name, cik, '20150101', '1')
