def get_new_pricing_data(self, work_dict):
"""get_new_pricing_data
Get Ticker information on:
- prices - turn off with ``work_dict.get_pricing = False``
- news - turn off with ``work_dict.get_news = False``
- options - turn off with ``work_dict.get_options = False``
:param work_dict: dictionary for key/values
"""
label = 'get_new_pricing_data'
log.info('task - {} - start ' 'work_dict={}'.format(label, work_dict))
ticker = ae_consts.TICKER
ticker_id = ae_consts.TICKER_ID
rec = {
'pricing': None,
'options': None,
'calls': None,
'puts': None,
'news': None,
'daily': None,
'minute': None,
'quote': None,
'stats': None,
'peers': None,
'iex_news': None,
'financials': None,
'earnings': None,
'dividends': None,
'company': None,
'exp_date': None,
'publish_pricing_update': None,
'date': ae_utils.utc_now_str(),
'updated': None,
'version': ae_consts.DATASET_COLLECTION_VERSION
}
res = {'status': ae_consts.NOT_RUN, 'err': None, 'rec': rec}
try:
ticker = work_dict.get('ticker', ticker)
ticker_id = work_dict.get('ticker_id', ae_consts.TICKER_ID)
s3_bucket = work_dict.get('s3_bucket', ae_consts.S3_BUCKET)
s3_key = work_dict.get('s3_key', ae_consts.S3_KEY)
redis_key = work_dict.get('redis_key', ae_consts.REDIS_KEY)
exp_date = work_dict.get('exp_date', None)
cur_date = datetime.datetime.utcnow()
cur_strike = work_dict.get('strike', None)
contract_type = str(work_dict.get('contract', 'C')).upper()
label = work_dict.get('label', label)
iex_datasets = work_dict.get('iex_datasets',
iex_consts.DEFAULT_FETCH_DATASETS)
td_datasets = work_dict.get('td_datasets',
td_consts.DEFAULT_FETCH_DATASETS_TD)
fetch_mode = work_dict.get('fetch_mode', ae_consts.FETCH_MODE_ALL)
# control flags to deal with feed issues:
get_iex_data = True
get_td_data = True
if (fetch_mode == ae_consts.FETCH_MODE_ALL
or str(fetch_mode).lower() == 'all'):
get_iex_data = True
get_td_data = True
elif (fetch_mode == ae_consts.FETCH_MODE_YHO
or str(fetch_mode).lower() == 'yahoo'):
get_iex_data = False
get_td_data = False
elif (fetch_mode == ae_consts.FETCH_MODE_IEX
or str(fetch_mode).lower() == 'iex'):
get_iex_data = True
get_td_data = False
elif (fetch_mode == ae_consts.FETCH_MODE_TD
or str(fetch_mode).lower() == 'td'):
get_iex_data = False
get_td_data = True
else:
log.debug('{} - unsupported fetch_mode={} value'.format(
label, fetch_mode))
"""
as of Thursday, Jan. 3, 2019:
https://developer.yahoo.com/yql/
Important EOL Notice: As of Thursday, Jan. 3, 2019
the YQL service at query.yahooapis.com will be retired
"""
get_yahoo_data = False
if not exp_date:
exp_date = opt_dates.option_expiration(date=exp_date)
else:
exp_date = datetime.datetime.strptime(exp_date, '%Y-%m-%d')
rec['updated'] = cur_date.strftime('%Y-%m-%d %H:%M:%S')
log.info('{} getting pricing for ticker={} '
'cur_date={} exp_date={} '
'yahoo={} iex={}'.format(label, ticker, cur_date, exp_date,
get_yahoo_data, get_iex_data))
yahoo_rec = {
'ticker': ticker,
'pricing': None,
'options': None,
'calls': None,
'puts': None,
'news': None,
'exp_date': None,
'publish_pricing_update': None,
'date': None,
'updated': None
}
# disabled on 2019-01-03
if get_yahoo_data:
log.info('{} yahoo ticker={}'.format(label, ticker))
yahoo_res = yahoo_data.get_data_from_yahoo(work_dict=work_dict)
if yahoo_res['status'] == ae_consts.SUCCESS:
yahoo_rec = yahoo_res['rec']
log.info('{} yahoo ticker={} '
'status={} err={}'.format(
label, ticker,
ae_consts.get_status(status=yahoo_res['status']),
yahoo_res['err']))
rec['pricing'] = yahoo_rec.get('pricing', '{}')
rec['news'] = yahoo_rec.get('news', '{}')
rec['options'] = yahoo_rec.get('options', '{}')
rec['calls'] = rec['options'].get('calls',
ae_consts.EMPTY_DF_STR)
rec['puts'] = rec['options'].get('puts',
ae_consts.EMPTY_DF_STR)
else:
log.error('{} failed YAHOO ticker={} '
'status={} err={}'.format(
label, ticker,
ae_consts.get_status(status=yahoo_res['status']),
yahoo_res['err']))
# end of get from yahoo
if get_iex_data:
num_iex_ds = len(iex_datasets)
log.debug('{} iex datasets={}'.format(label, num_iex_ds))
for idx, ft_type in enumerate(iex_datasets):
dataset_field = iex_consts.get_ft_str(ft_type=ft_type)
log.info('{} iex={}/{} field={} ticker={}'.format(
label, idx, num_iex_ds, dataset_field, ticker))
iex_label = '{}-{}'.format(label, dataset_field)
iex_req = copy.deepcopy(work_dict)
iex_req['label'] = iex_label
iex_req['ft_type'] = ft_type
iex_req['field'] = dataset_field
iex_req['ticker'] = ticker
iex_res = iex_data.get_data_from_iex(work_dict=iex_req)
if iex_res['status'] == ae_consts.SUCCESS:
iex_rec = iex_res['rec']
log.info(
'{} iex ticker={} field={} '
'status={} err={}'.format(
label, ticker, dataset_field,
ae_consts.get_status(status=iex_res['status']),
iex_res['err']))
if dataset_field == 'news':
rec['iex_news'] = iex_rec['data']
else:
rec[dataset_field] = iex_rec['data']
else:
log.debug(
'{} failed IEX ticker={} field={} '
'status={} err={}'.format(
label, ticker, dataset_field,
ae_consts.get_status(status=iex_res['status']),
iex_res['err']))
# end of if/else succcess
# end idx, ft_type in enumerate(iex_datasets):
# end of if get_iex_data
if get_td_data:
num_td_ds = len(td_datasets)
log.debug('{} td datasets={}'.format(label, num_td_ds))
for idx, ft_type in enumerate(td_datasets):
dataset_field = td_consts.get_ft_str_td(ft_type=ft_type)
log.info('{} td={}/{} field={} ticker={}'.format(
label, idx, num_td_ds, dataset_field, ticker))
td_label = '{}-{}'.format(label, dataset_field)
td_req = copy.deepcopy(work_dict)
td_req['label'] = td_label
td_req['ft_type'] = ft_type
td_req['field'] = dataset_field
td_req['ticker'] = ticker
td_res = td_data.get_data_from_td(work_dict=td_req)
if td_res['status'] == ae_consts.SUCCESS:
td_rec = td_res['rec']
log.info('{} td ticker={} field={} '
'status={} err={}'.format(
label, ticker, dataset_field,
ae_consts.get_status(status=td_res['status']),
td_res['err']))
if dataset_field == 'tdcalls':
rec['tdcalls'] = td_rec['data']
if dataset_field == 'tdputs':
rec['tdputs'] = td_rec['data']
else:
rec[dataset_field] = td_rec['data']
else:
log.critical(
'{} failed TD ticker={} field={} '
'status={} err={}'.format(
label, ticker, dataset_field,
ae_consts.get_status(status=td_res['status']),
td_res['err']))
# end of if/else succcess
# end idx, ft_type in enumerate(td_datasets):
# end of if get_td_data
update_req = {'data': rec}
update_req['ticker'] = ticker
update_req['ticker_id'] = ticker_id
update_req['strike'] = cur_strike
update_req['contract'] = contract_type
update_req['s3_enabled'] = work_dict.get('s3_enabled',
ae_consts.ENABLED_S3_UPLOAD)
update_req['redis_enabled'] = work_dict.get(
'redis_enabled', ae_consts.ENABLED_REDIS_PUBLISH)
update_req['s3_bucket'] = s3_bucket
update_req['s3_key'] = s3_key
update_req['s3_access_key'] = work_dict.get('s3_access_key',
ae_consts.S3_ACCESS_KEY)
update_req['s3_secret_key'] = work_dict.get('s3_secret_key',
ae_consts.S3_SECRET_KEY)
update_req['s3_region_name'] = work_dict.get('s3_region_name',
ae_consts.S3_REGION_NAME)
update_req['s3_address'] = work_dict.get('s3_address',
ae_consts.S3_ADDRESS)
update_req['s3_secure'] = work_dict.get('s3_secure',
ae_consts.S3_SECURE)
update_req['redis_key'] = redis_key
update_req['redis_address'] = work_dict.get('redis_address',
ae_consts.REDIS_ADDRESS)
update_req['redis_password'] = work_dict.get('redis_password',
ae_consts.REDIS_PASSWORD)
update_req['redis_db'] = int(
work_dict.get('redis_db', ae_consts.REDIS_DB))
update_req['redis_expire'] = work_dict.get('redis_expire',
ae_consts.REDIS_EXPIRE)
update_req['updated'] = rec['updated']
update_req['label'] = label
update_req['celery_disabled'] = True
update_status = ae_consts.NOT_SET
try:
update_res = publisher.run_publish_pricing_update(
work_dict=update_req)
update_status = update_res.get('status', ae_consts.NOT_SET)
if ae_consts.ev('DEBUG_RESULTS', '0') == '1':
log.info('{} update_res status={} data={}'.format(
label, ae_consts.get_status(status=update_status),
ae_consts.ppj(update_res)))
else:
log.info('{} run_publish_pricing_update status={}'.format(
label, ae_consts.get_status(status=update_status)))
# end of if/else
rec['publish_pricing_update'] = update_res
res = build_result.build_result(status=ae_consts.SUCCESS,
err=None,
rec=rec)
except Exception as f:
err = ('{} publisher.run_publish_pricing_update failed '
'with ex={}'.format(label, f))
log.error(err)
res = build_result.build_result(status=ae_consts.ERR,
err=err,
rec=rec)
# end of trying to publish results to connected services
except Exception as e:
res = build_result.build_result(status=ae_consts.ERR,
err=('failed - get_new_pricing_data '
'dict={} with ex={}').format(
work_dict, e),
rec=rec)
log.error('{} - {}'.format(label, res['err']))
# end of try/ex
if ae_consts.ev('DATASET_COLLECTION_SLACK_ALERTS', '0') == '1':
env_name = 'DEV'
if ae_consts.ev('PROD_SLACK_ALERTS', '1') == '1':
env_name = 'PROD'
done_msg = ('Dataset collected ticker=*{}* on env=*{}* '
'redis_key={} s3_key={} iex={} yahoo={}'.format(
ticker, env_name, redis_key, s3_key, get_iex_data,
get_yahoo_data))
log.debug('{} sending slack msg={}'.format(label, done_msg))
if res['status'] == ae_consts.SUCCESS:
slack_utils.post_success(msg=done_msg, block=False, jupyter=True)
else:
slack_utils.post_failure(msg=done_msg, block=False, jupyter=True)
# end of if/else success
# end of publishing to slack
log.info('task - get_new_pricing_data done - '
'{} - status={}'.format(label,
ae_consts.get_status(res['status'])))
return get_task_results.get_task_results(work_dict=work_dict, result=res)
def get_new_pricing_data(self, work_dict):
"""get_new_pricing_data
Get Ticker information on:
- prices - turn off with ``work_dict.get_pricing = False``
- news - turn off with ``work_dict.get_news = False``
- options - turn off with ``work_dict.get_options = False``
:param work_dict: dictionary for key/values
"""
label = 'get_new_pricing_data'
log.debug(f'task - {label} - start ' f'work_dict={work_dict}')
num_success = 0
ticker = ae_consts.TICKER
ticker_id = ae_consts.TICKER_ID
rec = {
'pricing': None,
'options': None,
'calls': None,
'puts': None,
'news': None,
'daily': None,
'minute': None,
'quote': None,
'stats': None,
'peers': None,
'iex_news': None,
'financials': None,
'earnings': None,
'dividends': None,
'company': None,
'exp_date': None,
'publish_pricing_update': None,
'num_success': num_success,
'date': ae_utils.utc_now_str(),
'updated': None,
'version': ae_consts.DATASET_COLLECTION_VERSION
}
res = {'status': ae_consts.NOT_RUN, 'err': None, 'rec': rec}
try:
ticker = work_dict.get('ticker', ticker)
ticker_id = work_dict.get('ticker_id', ae_consts.TICKER_ID)
s3_bucket = work_dict.get('s3_bucket', ae_consts.S3_BUCKET)
s3_key = work_dict.get('s3_key', ae_consts.S3_KEY)
redis_key = work_dict.get('redis_key', ae_consts.REDIS_KEY)
exp_date = work_dict.get('exp_date', None)
cur_date = ae_utils.last_close()
cur_strike = work_dict.get('strike', None)
contract_type = str(work_dict.get('contract', 'C')).upper()
label = work_dict.get('label', label)
iex_datasets = work_dict.get('iex_datasets',
iex_consts.DEFAULT_FETCH_DATASETS)
td_datasets = work_dict.get('td_datasets',
td_consts.DEFAULT_FETCH_DATASETS_TD)
fetch_mode = work_dict.get('fetch_mode', ae_consts.FETCH_MODE_ALL)
iex_token = work_dict.get('iex_token', iex_consts.IEX_TOKEN)
td_token = work_dict.get('td_token', td_consts.TD_TOKEN)
str_fetch_mode = str(fetch_mode).lower()
# control flags to deal with feed issues:
get_iex_data = True
get_td_data = True
if (fetch_mode == ae_consts.FETCH_MODE_ALL
or str_fetch_mode == 'initial'):
get_iex_data = True
get_td_data = True
iex_datasets = ae_consts.IEX_INITIAL_DATASETS
elif (fetch_mode == ae_consts.FETCH_MODE_ALL
or str_fetch_mode == 'all'):
get_iex_data = True
get_td_data = True
iex_datasets = ae_consts.IEX_DATASETS_DEFAULT
elif (fetch_mode == ae_consts.FETCH_MODE_YHO
or str_fetch_mode == 'yahoo'):
get_iex_data = False
get_td_data = False
elif (fetch_mode == ae_consts.FETCH_MODE_IEX
or str_fetch_mode == 'iex-all'):
get_iex_data = True
get_td_data = False
iex_datasets = ae_consts.IEX_DATASETS_DEFAULT
elif (fetch_mode == ae_consts.FETCH_MODE_IEX
or str_fetch_mode == 'iex'):
get_iex_data = True
get_td_data = False
iex_datasets = ae_consts.IEX_INTRADAY_DATASETS
elif (fetch_mode == ae_consts.FETCH_MODE_INTRADAY
or str_fetch_mode == 'intra'):
get_iex_data = True
get_td_data = True
iex_datasets = ae_consts.IEX_INTRADAY_DATASETS
elif (fetch_mode == ae_consts.FETCH_MODE_DAILY
or str_fetch_mode == 'daily'):
get_iex_data = True
get_td_data = False
iex_datasets = ae_consts.IEX_DAILY_DATASETS
elif (fetch_mode == ae_consts.FETCH_MODE_WEEKLY
or str_fetch_mode == 'weekly'):
get_iex_data = True
get_td_data = False
iex_datasets = ae_consts.IEX_WEEKLY_DATASETS
elif (fetch_mode == ae_consts.FETCH_MODE_TD or str_fetch_mode == 'td'):
get_iex_data = False
get_td_data = True
else:
get_iex_data = False
get_td_data = False
fetch_arr = str_fetch_mode.split(',')
found_fetch = False
iex_datasets = []
for fetch_name in fetch_arr:
if fetch_name not in iex_datasets:
if fetch_name == 'iex_min':
iex_datasets.append('minute')
elif fetch_name == 'iex_day':
iex_datasets.append('daily')
elif fetch_name == 'iex_quote':
iex_datasets.append('quote')
elif fetch_name == 'iex_stats':
iex_datasets.append('stats')
elif fetch_name == 'iex_peers':
iex_datasets.append('peers')
elif fetch_name == 'iex_news':
iex_datasets.append('news')
elif fetch_name == 'iex_fin':
iex_datasets.append('financials')
elif fetch_name == 'iex_earn':
iex_datasets.append('earnings')
elif fetch_name == 'iex_div':
iex_datasets.append('dividends')
elif fetch_name == 'iex_comp':
iex_datasets.append('company')
elif fetch_name == 'td':
get_td_data = True
else:
log.warn('unsupported IEX dataset ' f'{fetch_name}')
found_fetch = (len(iex_datasets) != 0)
if not found_fetch:
log.error(f'{label} - unsupported '
f'fetch_mode={fetch_mode} value')
else:
get_iex_data = True
log.debug(f'{label} - '
f'fetching={len(iex_datasets)} '
f'{iex_datasets} '
f'fetch_mode={fetch_mode}')
# end of screening custom fetch_mode settings
num_tokens = 0
if get_iex_data:
if not iex_token:
log.warn(f'{label} - '
'please set a valid IEX Cloud Account token ('
'https://iexcloud.io/cloud-login/#/register'
') to fetch data from IEX Cloud. It must be '
'set as an environment variable like: '
'export IEX_TOKEN=<token>')
get_iex_data = False
else:
num_tokens += 1
# sanity check - disable IEX fetch if the token is not set
if get_td_data:
missing_td_token = [
'MISSING_TD_TOKEN', 'SETYOURTDTOKEN', 'SETYOURTRADIERTOKENHERE'
]
if td_token in missing_td_token:
log.warn(f'{label} - '
'please set a valid Tradier Account token ('
'https://developer.tradier.com/user/sign_up'
') to fetch pricing data from Tradier. It must be '
'set as an environment variable like: '
'export TD_TOKEN=<token>')
get_td_data = False
else:
num_tokens += 1
# sanity check - disable Tradier fetch if the token is not set
"""
as of Thursday, Jan. 3, 2019:
https://developer.yahoo.com/yql/
Important EOL Notice: As of Thursday, Jan. 3, 2019
the YQL service at query.yahooapis.com will be retired
"""
get_yahoo_data = False
if (not get_iex_data and not get_td_data and not get_yahoo_data):
err = None
if num_tokens == 0:
res['status'] = ae_consts.MISSING_TOKEN
err = (f'Please set a valid IEX_TOKEN or TD_TOKEN '
f'environment variable')
else:
err = (f'Please set at least one supported datafeed from '
f'either: '
f'IEX Cloud (fetch -t TICKER -g iex) or '
f'Tradier (fetch -t TICKER -g td) '
f'for '
f'ticker={ticker} '
f'cur_date={cur_date} '
f'IEX enabled={get_iex_data} '
f'TD enabled={get_td_data} '
f'YHO enabled={get_yahoo_data}')
res['status'] = ae_consts.ERR
res['err'] = err
return get_task_results.get_task_results(work_dict=work_dict,
result=res)
# end of checking that there is at least 1 feed on
if not exp_date:
exp_date = opt_dates.option_expiration(date=exp_date)
else:
exp_date = datetime.datetime.strptime(exp_date, '%Y-%m-%d')
rec['updated'] = cur_date.strftime('%Y-%m-%d %H:%M:%S')
log.debug(f'{label} getting pricing for ticker={ticker} '
f'cur_date={cur_date} exp_date={exp_date} '
f'IEX={get_iex_data} '
f'TD={get_td_data} '
f'YHO={get_yahoo_data}')
yahoo_rec = {
'ticker': ticker,
'pricing': None,
'options': None,
'calls': None,
'puts': None,
'news': None,
'exp_date': None,
'publish_pricing_update': None,
'date': None,
'updated': None
}
# disabled on 2019-01-03
if get_yahoo_data:
log.debug(f'{label} YHO ticker={ticker}')
yahoo_res = yahoo_data.get_data_from_yahoo(work_dict=work_dict)
status_str = ae_consts.get_status(status=yahoo_res['status'])
if yahoo_res['status'] == ae_consts.SUCCESS:
yahoo_rec = yahoo_res['rec']
msg = (f'{label} YHO ticker={ticker} '
f'status={status_str} err={yahoo_res["err"]}')
if ae_consts.ev('SHOW_SUCCESS', '0') == '1':
log.info(msg)
else:
log.debug(msg)
rec['pricing'] = yahoo_rec.get('pricing', '{}')
rec['news'] = yahoo_rec.get('news', '{}')
rec['options'] = yahoo_rec.get('options', '{}')
rec['calls'] = rec['options'].get('calls',
ae_consts.EMPTY_DF_STR)
rec['puts'] = rec['options'].get('puts',
ae_consts.EMPTY_DF_STR)
num_success += 1
else:
log.error(f'{label} failed YHO ticker={ticker} '
f'status={status_str} err={yahoo_res["err"]}')
# end of get from yahoo
if get_iex_data:
num_iex_ds = len(iex_datasets)
log.debug(f'{label} IEX datasets={num_iex_ds}')
for idx, ft_type in enumerate(iex_datasets):
dataset_field = iex_consts.get_ft_str(ft_type=ft_type)
log.debug(f'{label} IEX={idx}/{num_iex_ds} '
f'field={dataset_field} ticker={ticker}')
iex_label = f'{label}-{dataset_field}'
iex_req = copy.deepcopy(work_dict)
iex_req['label'] = iex_label
iex_req['ft_type'] = ft_type
iex_req['field'] = dataset_field
iex_req['ticker'] = ticker
iex_res = iex_data.get_data_from_iex(work_dict=iex_req)
status_str = (ae_consts.get_status(status=iex_res['status']))
if iex_res['status'] == ae_consts.SUCCESS:
iex_rec = iex_res['rec']
msg = (f'{label} IEX ticker={ticker} '
f'field={dataset_field} '
f'status={status_str} '
f'err={iex_res["err"]}')
if ae_consts.ev('SHOW_SUCCESS', '0') == '1':
log.info(msg)
else:
log.debug(msg)
if dataset_field == 'news':
rec['iex_news'] = iex_rec['data']
else:
rec[dataset_field] = iex_rec['data']
num_success += 1
else:
log.debug(f'{label} failed IEX ticker={ticker} '
f'field={dataset_field} '
f'status={status_str} err={iex_res["err"]}')
# end of if/else succcess
# end idx, ft_type in enumerate(iex_datasets):
# end of if get_iex_data
if get_td_data:
num_td_ds = len(td_datasets)
log.debug(f'{label} TD datasets={num_td_ds}')
for idx, ft_type in enumerate(td_datasets):
dataset_field = td_consts.get_ft_str_td(ft_type=ft_type)
log.debug(f'{label} TD={idx}/{num_td_ds} '
f'field={dataset_field} ticker={ticker}')
td_label = (f'{label}-{dataset_field}')
td_req = copy.deepcopy(work_dict)
td_req['label'] = td_label
td_req['ft_type'] = ft_type
td_req['field'] = dataset_field
td_req['ticker'] = ticker
td_res = td_data.get_data_from_td(work_dict=td_req)
status_str = (ae_consts.get_status(status=td_res['status']))
if td_res['status'] == ae_consts.SUCCESS:
td_rec = td_res['rec']
msg = (f'{label} TD ticker={ticker} '
f'field={dataset_field} '
f'status={status_str} '
f'err={td_res["err"]}')
if ae_consts.ev('SHOW_SUCCESS', '0') == '1':
log.info(msg)
else:
log.debug(msg)
if dataset_field == 'tdcalls':
rec['tdcalls'] = td_rec['data']
if dataset_field == 'tdputs':
rec['tdputs'] = td_rec['data']
else:
rec[dataset_field] = td_rec['data']
num_success += 1
else:
log.critical(f'{label} failed TD ticker={ticker} '
f'field={dataset_field} '
f'status={status_str} err={td_res["err"]}')
# end of if/else succcess
# end idx, ft_type in enumerate(td_datasets):
# end of if get_td_data
rec['num_success'] = num_success
update_req = {'data': rec}
update_req['ticker'] = ticker
update_req['ticker_id'] = ticker_id
update_req['strike'] = cur_strike
update_req['contract'] = contract_type
update_req['s3_enabled'] = work_dict.get('s3_enabled',
ae_consts.ENABLED_S3_UPLOAD)
update_req['redis_enabled'] = work_dict.get(
'redis_enabled', ae_consts.ENABLED_REDIS_PUBLISH)
update_req['s3_bucket'] = s3_bucket
update_req['s3_key'] = s3_key
update_req['s3_access_key'] = work_dict.get('s3_access_key',
ae_consts.S3_ACCESS_KEY)
update_req['s3_secret_key'] = work_dict.get('s3_secret_key',
ae_consts.S3_SECRET_KEY)
update_req['s3_region_name'] = work_dict.get('s3_region_name',
ae_consts.S3_REGION_NAME)
update_req['s3_address'] = work_dict.get('s3_address',
ae_consts.S3_ADDRESS)
update_req['s3_secure'] = work_dict.get('s3_secure',
ae_consts.S3_SECURE)
update_req['redis_key'] = redis_key
update_req['redis_address'] = work_dict.get('redis_address',
ae_consts.REDIS_ADDRESS)
update_req['redis_password'] = work_dict.get('redis_password',
ae_consts.REDIS_PASSWORD)
update_req['redis_db'] = int(
work_dict.get('redis_db', ae_consts.REDIS_DB))
update_req['redis_expire'] = work_dict.get('redis_expire',
ae_consts.REDIS_EXPIRE)
update_req['updated'] = rec['updated']
update_req['label'] = label
update_req['celery_disabled'] = True
update_status = ae_consts.NOT_SET
try:
update_res = publisher.run_publish_pricing_update(
work_dict=update_req)
update_status = update_res.get('status', ae_consts.NOT_SET)
status_str = ae_consts.get_status(status=update_status)
if ae_consts.ev('DEBUG_RESULTS', '0') == '1':
log.debug(f'{label} update_res '
f'status={status_str} '
f'data={ae_consts.ppj(update_res)}')
else:
log.debug(f'{label} run_publish_pricing_update '
f'status={status_str}')
# end of if/else
rec['publish_pricing_update'] = update_res
res = build_result.build_result(status=ae_consts.SUCCESS,
err=None,
rec=rec)
except Exception as f:
err = (f'{label} publisher.run_publish_pricing_update failed '
f'with ex={f}')
log.error(err)
res = build_result.build_result(status=ae_consts.ERR,
err=err,
rec=rec)
# end of trying to publish results to connected services
except Exception as e:
res = build_result.build_result(status=ae_consts.ERR,
err=('failed - get_new_pricing_data '
f'dict={work_dict} with ex={e}'),
rec=rec)
log.error(f'{label} - {res["err"]}')
# end of try/ex
if ae_consts.ev('DATASET_COLLECTION_SLACK_ALERTS', '0') == '1':
env_name = 'DEV'
if ae_consts.ev('PROD_SLACK_ALERTS', '1') == '1':
env_name = 'PROD'
done_msg = (f'Dataset collected ticker=*{ticker}* on '
f'env=*{env_name}* '
f'redis_key={redis_key} s3_key={s3_key} '
f'IEX={get_iex_data} '
f'TD={get_td_data} '
f'YHO={get_yahoo_data}')
log.debug(f'{label} sending slack msg={done_msg}')
if res['status'] == ae_consts.SUCCESS:
slack_utils.post_success(msg=done_msg, block=False, jupyter=True)
else:
slack_utils.post_failure(msg=done_msg, block=False, jupyter=True)
# end of if/else success
# end of publishing to slack
log.debug('task - get_new_pricing_data done - '
f'{label} - status={ae_consts.get_status(res["status"])}')
return get_task_results.get_task_results(work_dict=work_dict, result=res)
def publish(data,
label=None,
convert_to_json=False,
is_df=False,
output_file=None,
df_compress=False,
compress=False,
redis_enabled=True,
redis_key=None,
redis_address=None,
redis_db=None,
redis_password=None,
redis_expire=None,
redis_serializer='json',
redis_encoding='utf-8',
s3_enabled=True,
s3_key=None,
s3_address=None,
s3_bucket=None,
s3_access_key=None,
s3_secret_key=None,
s3_region_name=None,
s3_secure=False,
slack_enabled=False,
slack_code_block=False,
slack_full_width=False,
verbose=False,
silent=False,
**kwargs):
"""publish
Publish ``data`` to multiple optional endpoints:
- a local file path (``output_file``)
- minio (``s3_bucket`` and ``s3_key``)
- redis (``redis_key``)
- slack
:return: status value
:param data: data to publish
:param convert_to_json: convert ``data`` to a
json-serialized string. this function will throw if
``json.dumps(data)`` fails
:param is_df: convert ``pd.DataFrame`` using
``pd.DataFrame.to_json()`` to a
json-serialized string. this function will throw if
``to_json()`` fails
:param label: log tracking label
:param output_file: path to save the data
to a file
:param df_compress: optional - compress data that is a
``pandas.DataFrame`` before publishing
:param compress: optional - compress before publishing
(default is ``False``)
:param verbose: optional - boolean to log output
(default is ``False``)
:param silent: optional - boolean no log output
(default is ``False``)
:param kwargs: optional - future argument support
**(Optional) Redis connectivity arguments**
:param redis_enabled: bool - toggle for auto-caching all
datasets in Redis
(default is ``True``)
:param redis_key: string - key to save the data in redis
(default is ``None``)
:param redis_address: Redis connection string format: ``host:port``
(default is ``localhost:6379``)
:param redis_db: Redis db to use
(default is ``0``)
:param redis_password: optional - Redis password
(default is ``None``)
:param redis_expire: optional - Redis expire value
(default is ``None``)
:param redis_serializer: not used yet - support for future
pickle objects in redis
:param redis_encoding: format of the encoded key in redis
**(Optional) Minio (S3) connectivity arguments**
:param s3_enabled: bool - toggle for auto-archiving on Minio (S3)
(default is ``True``)
:param s3_key: string - key to save the data in redis
(default is ``None``)
:param s3_address: Minio S3 connection string format: ``host:port``
(default is ``localhost:9000``)
:param s3_bucket: S3 Bucket for storing the artifacts
(default is ``dev``) which should be viewable on a browser:
http://localhost:9000/minio/dev/
:param s3_access_key: S3 Access key
(default is ``trexaccesskey``)
:param s3_secret_key: S3 Secret key
(default is ``trex123321``)
:param s3_region_name: S3 region name
(default is ``us-east-1``)
:param s3_secure: Transmit using tls encryption
(default is ``False``)
**(Optional) Slack arguments**
:param slack_enabled: optional - boolean for
publishing to slack
:param slack_code_block: optional - boolean for
publishing as a code black in slack
:param slack_full_width: optional - boolean for
publishing as a to slack using the full
width allowed
"""
status = ae_consts.NOT_RUN
use_data = data
if (not df_compress and not is_df and not use_data):
log.info('missing data')
return ae_consts.INVALID
if convert_to_json and not is_df:
if verbose:
log.debug('start convert to json')
use_data = json.dumps(data)
if verbose:
log.debug('done convert to json')
if is_df:
if verbose:
log.debug('start df to_json')
use_data = data.to_json(orient='records', date_format='iso')
if verbose:
log.debug('done df to_json')
already_compressed = False
if df_compress:
use_data = compress_data.compress_data(data=data)
already_compressed = True
elif compress and not df_compress:
if verbose:
log.debug('compress start')
use_data = zlib.compress(use_data.encode(redis_encoding), 9)
already_compressed = True
if verbose:
log.debug('compress end')
num_bytes = len(use_data)
num_mb = ae_consts.get_mb(num_bytes)
if verbose:
log.debug(f'start - file={output_file} s3_key={s3_key} '
f'redis_key={redis_key} slack={slack_enabled} '
f'compress={compress} size={num_mb}MB')
if s3_enabled and s3_address and s3_bucket and s3_key:
endpoint_url = f'http{"s" if s3_secure else ""}://{s3_address}'
if verbose:
log.debug(f's3 start - {label} endpoint_url={endpoint_url} '
f'region={s3_region_name}')
s3 = boto3.resource(
's3',
endpoint_url=endpoint_url,
aws_access_key_id=s3_access_key,
aws_secret_access_key=s3_secret_key,
region_name=s3_region_name,
config=boto3.session.Config(signature_version='s3v4'))
if s3.Bucket(s3_bucket) not in s3.buckets.all():
if verbose:
log.debug(f's3 creating bucket={s3_bucket} {label}')
s3.create_bucket(Bucket=s3_bucket)
if verbose:
log.debug(f's3 upload start - bytes={num_mb} to '
f'{s3_bucket}:{s3_key} {label}')
s3.Bucket(s3_bucket).put_object(Key=s3_key, Body=use_data)
if verbose:
log.debug(f's3 upload done - bytes={num_mb} to '
f'{s3_bucket}:{s3_key} {label}')
# end of s3_enabled
if redis_enabled and redis_address and redis_key:
redis_split = redis_address.split(':')
redis_host = redis_split[0]
redis_port = int(redis_split[1])
log.debug(f'{label if label else ""} '
f'redis={redis_host}:{redis_port}@{redis_db} connect '
f'key={redis_key} expire={redis_expire}')
rc = redis.Redis(host=redis_host,
port=redis_port,
password=redis_password,
db=redis_db)
redis_res = redis_utils.set_data_in_redis_key(
label=label,
client=rc,
key=redis_key,
data=use_data,
already_compressed=already_compressed,
serializer=redis_serializer,
encoding=redis_encoding,
expire=redis_expire,
px=None,
nx=False,
xx=False)
if redis_res['status'] != ae_consts.SUCCESS:
log.error(f'redis failed - '
f'{ae_consts.get_status(status=redis_res["status"])} '
f'{redis_res["err"]}')
return ae_consts.REDIS_FAILED
# end of redis_enabled
if output_file:
if verbose:
log.debug(f'file start - output_file={output_file}')
file_exists = file_utils.write_to_file(output_file=output_file,
data=data)
if not file_exists:
log.error(f'file failed - did not find '
f'output_file={output_file}')
return ae_consts.FILE_FAILED
if verbose:
log.debug(f'file done - output_file={output_file}')
# end of writing to file
if slack_enabled:
if verbose:
log.debug('slack start')
slack_utils.post_success(msg=use_data,
block=slack_code_block,
full_width=slack_full_width)
if verbose:
log.debug('slack end')
# end of sending to slack
status = ae_consts.SUCCESS
if verbose:
log.debug(
f'end - {ae_consts.get_status(status=status)} file={output_file} '
f's3_key={s3_key} redis_key={redis_key} slack={slack_enabled} '
f'compress={compress} size={num_mb}MB')
return status