normanlmfung · June 17, 2024 03:48
diff --git a/gistfile1.txt b/gistfile1.txt
 import time
 import datetime
 from datetime import datetime, timedelta
 import enum
 from typing import List, Dict
 import os.path

 import pandas as pd
 from tabulate import tabulate

 from ccxt import okx as ccxt_okx
 from ccxt.base.exchange import Exchange

 from mds.ccxt.exchange_base import ExchangeBase

 EconomicCalendarEventPosNeg : enum = enum.Enum('PosNeg', 'UNDEFINED SMALLER_BETTER BIGGER_BETTER')

 def _fix_column_types(pd_economic_calendar : pd.DataFrame):
    pd_economic_calendar['datetime'] = pd.to_datetime(pd_economic_calendar['datetime'])

    # This is to make it easy to do grouping with Excel pivot table
    pd_economic_calendar['year'] = pd_economic_calendar['datetime'].dt.year
    pd_economic_calendar['month'] = pd_economic_calendar['datetime'].dt.month
    pd_economic_calendar['day'] = pd_economic_calendar['datetime'].dt.day
    pd_economic_calendar['hour'] = pd_economic_calendar['datetime'].dt.hour
    pd_economic_calendar['minute'] = pd_economic_calendar['datetime'].dt.minute
    
 class okx(ccxt_okx, ExchangeBase):
    def fetch_economic_calendar(
            self,
            regions : List[str] = None,
            importance : str = None,
            start_ts : int = -1, # in ms
            end_ts : int = -1, # in ms
            limit : int = 100,
            region_map : Dict[str, str] = None,
            event_name_map : Dict[str, str] = None,
            event_posneg_map : Dict = None,
            rate_limit_winodw_sec : float  = 5 # hit max once per five seconds!
    ) -> pd.DataFrame:
        all_economic_calendars : List[Dict] = []
        need_fetch_using_sliding_window : bool = False
        for region in regions:
            economic_calendars = self._fetch_economic_calendar(
                    region=region, importance=importance, start_ts=start_ts, end_ts=end_ts, limit=limit, 
                    region_map=region_map, event_name_map=event_name_map, event_posneg_map=event_posneg_map, 
                    rate_limit_winodw_sec=rate_limit_winodw_sec
                )

            if len(economic_calendars)<limit:
                all_economic_calendars = all_economic_calendars +  economic_calendars
            else:
                need_fetch_using_sliding_window = True
                break

            time.sleep(rate_limit_winodw_sec)

        if not need_fetch_using_sliding_window:
            pd_economic_calendars = pd.DataFrame(all_economic_calendars)
            _fix_column_types(pd_economic_calendars)
            return pd_economic_calendars

        if end_ts==-1:
            end_ts = (datetime.fromtimestamp(start_ts/1000) + timedelta(days=1)).timestamp() *1000
            end_ts = int(end_ts)

        time.sleep(rate_limit_winodw_sec)

        all_economic_calendars.clear()

        for region in regions:
            this_cutoff = start_ts
            while this_cutoff<=end_ts:
                economic_calendars = self._fetch_economic_calendar(
                        region=region, importance=importance, start_ts=this_cutoff, limit=limit, 
                        region_map=region_map, event_name_map=event_name_map, event_posneg_map=event_posneg_map, 
                        rate_limit_winodw_sec=rate_limit_winodw_sec
                    )
                if economic_calendars:
                    all_economic_calendars = all_economic_calendars + economic_calendars

                    record_ts = max([int(economic_calendar['calendar_item_timestamp_ms']) for economic_calendar in economic_calendars])
                    this_cutoff = record_ts  + 1

                time.sleep(rate_limit_winodw_sec)

        pd_economic_calendars = pd.DataFrame(all_economic_calendars)
        _fix_column_types(pd_economic_calendars)
        return pd_economic_calendars

    def _fetch_economic_calendar(
            self,
            region : str = None,
            importance : str = None,
            start_ts : int = -1, # in ms
            end_ts : int = -1, # in ms
            limit : int = 100,
            region_map : Dict[str, str] = None,
            event_name_map : Dict[str, str] = None,
            event_posneg_map : Dict = None,
            rate_limit_winodw_sec : float  = 5 # hit max once per five seconds!
            
    ) -> List[Dict]:
        '''
        Note, this API requires authentication even in doc it's marked 'public'
        https://www.okx.com/docs-v5/en/?python#public-data-rest-api-get-economic-calendar-data
        https://www.okx.com/docs-v5/en/?python#public-data-websocket-economic-calendar-channel
        '''
        params = {}
        if region:
            params['region'] = region
        if importance:
            if importance=='high':
                importance = 3
            elif importance=='medium':
                importance = 2
            elif importance=='low':
                importance = 1
            else:
                importance = -1
            if importance!=-1:
                params['importance'] = importance
        if start_ts!=-1:
            params['before'] = str(start_ts)
        if end_ts!=-1:
            params['after'] = str(end_ts)
        if limit!=100:
            limit['limit'] = str(limit)

        path = 'public/economic-calendar'
        # Because this API requires authentication, you need specify api=private. Look at okx.py sign() method. If you pass api=public, sign wont add API key headers.
        response = self.request(path = path, method='GET', api = 'private', params=params)        
        economic_calendars = response['data']
        
        for calendar_item in economic_calendars:
            calendar_item_timestamp_ms = int(calendar_item['date'])
            calendar_item_timestamp_sec = int(calendar_item_timestamp_ms/1000)
            calendar_item['calendar_item_timestamp_ms'] = calendar_item_timestamp_ms
            calendar_item['calendar_item_timestamp_sec'] = calendar_item_timestamp_sec
            calendar_item['datetime'] = datetime.fromtimestamp(calendar_item_timestamp_sec)
            calendar_item['source'] = self.name

            if region:
                calendar_item['region'] = region
            else:
                if region_map and calendar_item['region'] in region_map:
                    calendar_item['region'] = region_map[calendar_item['region']]
            
            calendar_item['event_code'] = None
            if event_name_map and calendar_item['event'] in event_name_map:
                calendar_item['event_code'] = event_name_map[calendar_item['event']]

            actual = calendar_item['actual']
            previous = calendar_item['previous']
            forecast = calendar_item['forecast']


            lower_case_words_to_be_removed = [ '$', 'cny' ]

            # OKX source from https://www.tradingview.com/economic-calendar, look there to see what formatting logic you need.
            if actual:
                actual = actual.lower()
                for word in lower_case_words_to_be_removed:
                    actual = actual.replace(word,'')
                actual = actual.strip()
                if '%' in actual:
                    actual = float(actual.replace('%',''))/100
                elif 'k'==actual[-1]:
                    actual = actual.replace('k','000')
                    actual = float(actual)
                elif 'm'==actual[-1]:
                    actual = actual.replace('m','000000')
                    actual = float(actual)
                elif 'b'==actual[-1]:
                    actual = actual.replace('b','000000000')
                    actual = float(actual)
                else:
                    actual = float(actual)
                
            else:
                actual = None

            if previous:
                previous = previous.lower()
                for word in lower_case_words_to_be_removed:
                    previous = previous.replace(word,'')
                previous = previous.strip()
                if '%' in previous:
                    previous = float(previous.replace('%',''))/100
                elif 'k'==previous[-1]:
                    previous = previous.replace('k','000')
                    previous = float(previous)
                elif 'm'==previous[-1]:
                    previous = previous.replace('m','000000')
                    previous = float(previous)
                elif 'b'==previous[-1]:
                    previous = previous.replace('b','000000000')
                    previous = float(previous)
                else:
                    previous = float(previous)
                
            else:
                previous = None

            if forecast:
                forecast = forecast.lower()
                for word in lower_case_words_to_be_removed:
                    forecast = forecast.replace(word,'')
                forecast = forecast.strip()
                if '%' in forecast:
                    forecast = float(forecast.replace('%',''))/100
                elif 'k'==forecast[-1]:
                    forecast = forecast.replace('k','000')
                    forecast = float(forecast)
                elif 'm'==forecast[-1]:
                    forecast = forecast.replace('m','000000')
                    forecast = float(forecast)
                elif 'b'==forecast[-1]:
                    forecast = forecast.replace('b','000000000')
                    forecast = float(forecast)
                else:
                    forecast = float(forecast)
                
            else:
                forecast = None

            calendar_item['actual'] = actual
            calendar_item['previous'] = previous
            calendar_item['forecast'] = forecast

            calendar_item['pos_neg'] = None
            if actual and calendar_item['event_code']:
                calendar_item['pos_neg']  = okx.classify_economic_calendar(calendar_item)

            calendar_item.pop('date')

        economic_calendars.sort(key=lambda x : x['datetime'], reverse=False)
        return economic_calendars

    def classify_economic_calendar(calendar_item : Dict) -> str:
        actual = calendar_item['actual']
        previous = calendar_item['previous']
        forecast = calendar_item['forecast']

        if actual and forecast and previous and calendar_item['event_code']:
            event_pos_neg = event_posneg_map[calendar_item['event_code']]
            if event_pos_neg == EconomicCalendarEventPosNeg.SMALLER_BETTER: # Example Core CPI
                if actual<previous:
                    if actual<forecast:
                        return 'bullish'
                    elif actual==forecast:
                        return 'bullish'
                    else:
                        # actual>forecast
                        return 'bearish'

                elif actual==previous:
                    if actual<forecast:
                        return 'bullish'
                    elif actual==forecast:
                        return 'neutral'
                    else:
                        # actual>forecast
                        return 'bearish'

                else:
                    assert(actual>previous)
                    if actual<forecast:
                        return 'neutral' # neutral or bearish? Debatable
                    elif actual==forecast:
                        return 'bearish'
                    else:
                        # actual>forecast
                        return 'bearish'

            elif event_pos_neg == EconomicCalendarEventPosNeg.BIGGER_BETTER: # Example GDP
                if actual>previous:
                    if actual>forecast:
                        return 'bullish'
                    elif actual==forecast:
                        return 'bullish'
                    else:
                        # actual<forecast
                        return 'bearish'

                elif actual==previous:
                    if actual>forecast:
                        return 'bullish'
                    elif actual==forecast:
                        return 'neutral'
                    else:
                        # actual<forecast
                        return 'bearish'

                else:
                    assert(actual<previous)
                    if actual>forecast:
                        return 'neutral' # neutral or bearish? Debatable
                    elif actual==forecast:
                        return 'bearish'
                    else:
                        # actual>forecast
                        return 'bearish'

        return ""
    
 if __name__ == '__main__':
    # Parameters
    # Fetching economic calendars from OKX requires authentication. Also if your trading volume < usd 5mn (VIP1), regardless of your asset amount, you'd only get calendar items at most 3 months old.
    apiKey : str = None
    secret : str = None
    passphrase : str = None
    subaccount : str = None
    
    # If start_ts older than three months and your account don't have that entitlement, start_ts will be set to today by OKX (i.e. forward looking only)
    start_ts : int = int(datetime(2024,3,17).timestamp()*1000)
    # start_ts : int = int(datetime.now().timestamp()*1000)
    end_ts : int = int(datetime(2024,7,1).timestamp()*1000)

    export_csv_file_name : str = "economic_calanedar.csv"
    
    params = {
        'apiKey' : apiKey,
        'secret' : secret,
        'password' : passphrase,    # Other exchanges dont require this! This is saved in exchange.password!
        'subaccount' : subaccount,
        'rateLimit' : 100,                    # In ms
        'options' : {
            'defaultType': 'swap', # 'funding', 'spot', 'margin', 'future', 'swap', 'option'
        }
    }
    exchange : Exchange = okx(params)


    region_map = {
        'United States' : 'united_states',
        'China' : 'china'
    }
    event_name_map  : Dict[str, str] = {
        'Core Inflation Rate MoM' : 'core_inflation_rate_mom',
        'Core Inflation Rate YoY' : 'core_inflation_rate_yoy',
        'Inflation Rate MoM' : 'inflation_rate_mom',
        'Inflation Rate YoY' : 'inflation_rate_yoy',
        'Fed Interest Rate Decision' : 'fed_interest_rate_decision',
        'Core PCE Price Index MoM' : 'core_pce_price_index_mom',
        'Core PCE Price Index YoY' : 'core_pce_price_index_yoy',
        'Unemployment Rate' : 'unemployment_rate',
        'Non Farm Payrolls' : 'non_farm_payrolls',

        'GDP Growth Rate QoQ Final' : 'gdp_growth_rate_qoq_final',
        'GDP Growth Rate QoQ Adv' : 'gdp_growth_rate_qoq_adv',

        # Careful with spelling here
        'GDP Growth Rate QoQ 1st Est' : 'gdp_growth_rate_qoq_est',
        'GDP Growth Rate QoQ 2nd Est' : 'gdp_growth_rate_qoq_est',
        'GDP Growth Rate QoQ 3rd Est' : 'gdp_growth_rate_qoq_est',
        'GDP Growth Rate QoQ 4th Est' : 'gdp_growth_rate_qoq_est'
    }

    
    event_posneg_map : Dict = {
        'core_inflation_rate_mom' : EconomicCalendarEventPosNeg.SMALLER_BETTER,
        'core_inflation_rate_yoy' : EconomicCalendarEventPosNeg.SMALLER_BETTER,
        'inflation_rate_mom' : EconomicCalendarEventPosNeg.SMALLER_BETTER,
        'inflation_rate_yoy' : EconomicCalendarEventPosNeg.SMALLER_BETTER,
        'fed_interest_rate_decision' : EconomicCalendarEventPosNeg.SMALLER_BETTER,
        'core_pce_price_index_mom' : EconomicCalendarEventPosNeg.SMALLER_BETTER,
        'core_pce_price_index_yoy' : EconomicCalendarEventPosNeg.SMALLER_BETTER,
        'unemployment_rate' : EconomicCalendarEventPosNeg.BIGGER_BETTER, # more people out of job better as FEDs in position to ease.
        'non_farm_payrolls' : EconomicCalendarEventPosNeg.BIGGER_BETTER, # more people out of job better as FEDs in position to ease.
        'gdp_growth_rate_qoq_final' : EconomicCalendarEventPosNeg.BIGGER_BETTER,
        'gdp_growth_rate_qoq' : EconomicCalendarEventPosNeg.BIGGER_BETTER,
        'gdp_growth_rate_qoq_adv' : EconomicCalendarEventPosNeg.BIGGER_BETTER,
        'gdp_growth_rate_qoq_est' : EconomicCalendarEventPosNeg.BIGGER_BETTER
    }

    existing_csv : bool = os.path.isfile(export_csv_file_name)
    if existing_csv:
        pd_old_economic_calendars  = pd.read_csv(export_csv_file_name)
    
    pd_new_economic_calendars = exchange.fetch_economic_calendar(
        regions=['united_states', 'china'], 
        importance='high', 
        start_ts=start_ts, end_ts=end_ts,
        region_map=region_map,
        event_name_map=event_name_map,
        event_posneg_map=event_posneg_map,
        rate_limit_winodw_sec=5
        )
    
    

    min_calendar_item_timestamp_ms = pd_new_economic_calendars['calendar_item_timestamp_ms'].min()
    
    if existing_csv:
        pd_old_economic_calendars = pd_old_economic_calendars[pd_old_economic_calendars['calendar_item_timestamp_ms']<min_calendar_item_timestamp_ms]
        pd_economic_calendars = pd.concat([pd_old_economic_calendars, pd_new_economic_calendars], axis=0)
    else:
        pd_economic_calendars = pd_new_economic_calendars

    print(f"{tabulate(pd_economic_calendars, headers='keys', tablefmt='psql')}")
    pd_economic_calendars.to_csv(export_csv_file_name)
	import time
	import datetime
	from datetime import datetime, timedelta
	import enum
	from typing import List, Dict
	import os.path

	import pandas as pd
	from tabulate import tabulate

	from ccxt import okx as ccxt_okx
	from ccxt.base.exchange import Exchange

	from mds.ccxt.exchange_base import ExchangeBase

	EconomicCalendarEventPosNeg : enum = enum.Enum('PosNeg', 'UNDEFINED SMALLER_BETTER BIGGER_BETTER')

	def _fix_column_types(pd_economic_calendar : pd.DataFrame):
	pd_economic_calendar['datetime'] = pd.to_datetime(pd_economic_calendar['datetime'])

	# This is to make it easy to do grouping with Excel pivot table
	pd_economic_calendar['year'] = pd_economic_calendar['datetime'].dt.year
	pd_economic_calendar['month'] = pd_economic_calendar['datetime'].dt.month
	pd_economic_calendar['day'] = pd_economic_calendar['datetime'].dt.day
	pd_economic_calendar['hour'] = pd_economic_calendar['datetime'].dt.hour
	pd_economic_calendar['minute'] = pd_economic_calendar['datetime'].dt.minute

	class okx(ccxt_okx, ExchangeBase):
	def fetch_economic_calendar(
	self,
	regions : List[str] = None,
	importance : str = None,
	start_ts : int = -1, # in ms
	end_ts : int = -1, # in ms
	limit : int = 100,
	region_map : Dict[str, str] = None,
	event_name_map : Dict[str, str] = None,
	event_posneg_map : Dict = None,
	rate_limit_winodw_sec : float = 5 # hit max once per five seconds!
	) -> pd.DataFrame:
	all_economic_calendars : List[Dict] = []
	need_fetch_using_sliding_window : bool = False
	for region in regions:
	economic_calendars = self._fetch_economic_calendar(
	region=region, importance=importance, start_ts=start_ts, end_ts=end_ts, limit=limit,
	region_map=region_map, event_name_map=event_name_map, event_posneg_map=event_posneg_map,
	rate_limit_winodw_sec=rate_limit_winodw_sec
	)

	if len(economic_calendars)<limit:
	all_economic_calendars = all_economic_calendars + economic_calendars
	else:
	need_fetch_using_sliding_window = True
	break

	time.sleep(rate_limit_winodw_sec)

	if not need_fetch_using_sliding_window:
	pd_economic_calendars = pd.DataFrame(all_economic_calendars)
	_fix_column_types(pd_economic_calendars)
	return pd_economic_calendars

	if end_ts==-1:
	end_ts = (datetime.fromtimestamp(start_ts/1000) + timedelta(days=1)).timestamp() *1000
	end_ts = int(end_ts)

	time.sleep(rate_limit_winodw_sec)

	all_economic_calendars.clear()

	for region in regions:
	this_cutoff = start_ts
	while this_cutoff<=end_ts:
	economic_calendars = self._fetch_economic_calendar(
	region=region, importance=importance, start_ts=this_cutoff, limit=limit,
	region_map=region_map, event_name_map=event_name_map, event_posneg_map=event_posneg_map,
	rate_limit_winodw_sec=rate_limit_winodw_sec
	)
	if economic_calendars:
	all_economic_calendars = all_economic_calendars + economic_calendars

	record_ts = max([int(economic_calendar['calendar_item_timestamp_ms']) for economic_calendar in economic_calendars])
	this_cutoff = record_ts + 1

	time.sleep(rate_limit_winodw_sec)

	pd_economic_calendars = pd.DataFrame(all_economic_calendars)
	_fix_column_types(pd_economic_calendars)
	return pd_economic_calendars

	def _fetch_economic_calendar(
	self,
	region : str = None,
	importance : str = None,
	start_ts : int = -1, # in ms
	end_ts : int = -1, # in ms
	limit : int = 100,
	region_map : Dict[str, str] = None,
	event_name_map : Dict[str, str] = None,
	event_posneg_map : Dict = None,
	rate_limit_winodw_sec : float = 5 # hit max once per five seconds!

	) -> List[Dict]:
	'''
	Note, this API requires authentication even in doc it's marked 'public'
	https://www.okx.com/docs-v5/en/?python#public-data-rest-api-get-economic-calendar-data
	https://www.okx.com/docs-v5/en/?python#public-data-websocket-economic-calendar-channel
	'''
	params = {}
	if region:
	params['region'] = region
	if importance:
	if importance=='high':
	importance = 3
	elif importance=='medium':
	importance = 2
	elif importance=='low':
	importance = 1
	else:
	importance = -1
	if importance!=-1:
	params['importance'] = importance
	if start_ts!=-1:
	params['before'] = str(start_ts)
	if end_ts!=-1:
	params['after'] = str(end_ts)
	if limit!=100:
	limit['limit'] = str(limit)

	path = 'public/economic-calendar'
	# Because this API requires authentication, you need specify api=private. Look at okx.py sign() method. If you pass api=public, sign wont add API key headers.
	response = self.request(path = path, method='GET', api = 'private', params=params)
	economic_calendars = response['data']

	for calendar_item in economic_calendars:
	calendar_item_timestamp_ms = int(calendar_item['date'])
	calendar_item_timestamp_sec = int(calendar_item_timestamp_ms/1000)
	calendar_item['calendar_item_timestamp_ms'] = calendar_item_timestamp_ms
	calendar_item['calendar_item_timestamp_sec'] = calendar_item_timestamp_sec
	calendar_item['datetime'] = datetime.fromtimestamp(calendar_item_timestamp_sec)
	calendar_item['source'] = self.name

	if region:
	calendar_item['region'] = region
	else:
	if region_map and calendar_item['region'] in region_map:
	calendar_item['region'] = region_map[calendar_item['region']]

	calendar_item['event_code'] = None
	if event_name_map and calendar_item['event'] in event_name_map:
	calendar_item['event_code'] = event_name_map[calendar_item['event']]

	actual = calendar_item['actual']
	previous = calendar_item['previous']
	forecast = calendar_item['forecast']


	lower_case_words_to_be_removed = [ '$', 'cny' ]

	# OKX source from https://www.tradingview.com/economic-calendar, look there to see what formatting logic you need.
	if actual:
	actual = actual.lower()
	for word in lower_case_words_to_be_removed:
	actual = actual.replace(word,'')
	actual = actual.strip()
	if '%' in actual:
	actual = float(actual.replace('%',''))/100
	elif 'k'==actual[-1]:
	actual = actual.replace('k','000')
	actual = float(actual)
	elif 'm'==actual[-1]:
	actual = actual.replace('m','000000')
	actual = float(actual)
	elif 'b'==actual[-1]:
	actual = actual.replace('b','000000000')
	actual = float(actual)
	else:
	actual = float(actual)

	else:
	actual = None

	if previous:
	previous = previous.lower()
	for word in lower_case_words_to_be_removed:
	previous = previous.replace(word,'')
	previous = previous.strip()
	if '%' in previous:
	previous = float(previous.replace('%',''))/100
	elif 'k'==previous[-1]:
	previous = previous.replace('k','000')
	previous = float(previous)
	elif 'm'==previous[-1]:
	previous = previous.replace('m','000000')
	previous = float(previous)
	elif 'b'==previous[-1]:
	previous = previous.replace('b','000000000')
	previous = float(previous)
	else:
	previous = float(previous)

	else:
	previous = None

	if forecast:
	forecast = forecast.lower()
	for word in lower_case_words_to_be_removed:
	forecast = forecast.replace(word,'')
	forecast = forecast.strip()
	if '%' in forecast:
	forecast = float(forecast.replace('%',''))/100
	elif 'k'==forecast[-1]:
	forecast = forecast.replace('k','000')
	forecast = float(forecast)
	elif 'm'==forecast[-1]:
	forecast = forecast.replace('m','000000')
	forecast = float(forecast)
	elif 'b'==forecast[-1]:
	forecast = forecast.replace('b','000000000')
	forecast = float(forecast)
	else:
	forecast = float(forecast)

	else:
	forecast = None

	calendar_item['actual'] = actual
	calendar_item['previous'] = previous
	calendar_item['forecast'] = forecast

	calendar_item['pos_neg'] = None
	if actual and calendar_item['event_code']:
	calendar_item['pos_neg'] = okx.classify_economic_calendar(calendar_item)

	calendar_item.pop('date')

	economic_calendars.sort(key=lambda x : x['datetime'], reverse=False)
	return economic_calendars

	def classify_economic_calendar(calendar_item : Dict) -> str:
	actual = calendar_item['actual']
	previous = calendar_item['previous']
	forecast = calendar_item['forecast']

	if actual and forecast and previous and calendar_item['event_code']:
	event_pos_neg = event_posneg_map[calendar_item['event_code']]
	if event_pos_neg == EconomicCalendarEventPosNeg.SMALLER_BETTER: # Example Core CPI
	if actual<previous:
	if actual<forecast:
	return 'bullish'
	elif actual==forecast:
	return 'bullish'
	else:
	# actual>forecast
	return 'bearish'

	elif actual==previous:
	if actual<forecast:
	return 'bullish'
	elif actual==forecast:
	return 'neutral'
	else:
	# actual>forecast
	return 'bearish'

	else:
	assert(actual>previous)
	if actual<forecast:
	return 'neutral' # neutral or bearish? Debatable
	elif actual==forecast:
	return 'bearish'
	else:
	# actual>forecast
	return 'bearish'

	elif event_pos_neg == EconomicCalendarEventPosNeg.BIGGER_BETTER: # Example GDP
	if actual>previous:
	if actual>forecast:
	return 'bullish'
	elif actual==forecast:
	return 'bullish'
	else:
	# actual<forecast
	return 'bearish'

	elif actual==previous:
	if actual>forecast:
	return 'bullish'
	elif actual==forecast:
	return 'neutral'
	else:
	# actual<forecast
	return 'bearish'

	else:
	assert(actual<previous)
	if actual>forecast:
	return 'neutral' # neutral or bearish? Debatable
	elif actual==forecast:
	return 'bearish'
	else:
	# actual>forecast
	return 'bearish'

	return ""

	if __name__ == '__main__':
	# Parameters
	# Fetching economic calendars from OKX requires authentication. Also if your trading volume < usd 5mn (VIP1), regardless of your asset amount, you'd only get calendar items at most 3 months old.
	apiKey : str = None
	secret : str = None
	passphrase : str = None
	subaccount : str = None

	# If start_ts older than three months and your account don't have that entitlement, start_ts will be set to today by OKX (i.e. forward looking only)
	start_ts : int = int(datetime(2024,3,17).timestamp()*1000)
	# start_ts : int = int(datetime.now().timestamp()*1000)
	end_ts : int = int(datetime(2024,7,1).timestamp()*1000)

	export_csv_file_name : str = "economic_calanedar.csv"

	params = {
	'apiKey' : apiKey,
	'secret' : secret,
	'password' : passphrase, # Other exchanges dont require this! This is saved in exchange.password!
	'subaccount' : subaccount,
	'rateLimit' : 100, # In ms
	'options' : {
	'defaultType': 'swap', # 'funding', 'spot', 'margin', 'future', 'swap', 'option'
	}
	}
	exchange : Exchange = okx(params)


	region_map = {
	'United States' : 'united_states',
	'China' : 'china'
	}
	event_name_map : Dict[str, str] = {
	'Core Inflation Rate MoM' : 'core_inflation_rate_mom',
	'Core Inflation Rate YoY' : 'core_inflation_rate_yoy',
	'Inflation Rate MoM' : 'inflation_rate_mom',
	'Inflation Rate YoY' : 'inflation_rate_yoy',
	'Fed Interest Rate Decision' : 'fed_interest_rate_decision',
	'Core PCE Price Index MoM' : 'core_pce_price_index_mom',
	'Core PCE Price Index YoY' : 'core_pce_price_index_yoy',
	'Unemployment Rate' : 'unemployment_rate',
	'Non Farm Payrolls' : 'non_farm_payrolls',

	'GDP Growth Rate QoQ Final' : 'gdp_growth_rate_qoq_final',
	'GDP Growth Rate QoQ Adv' : 'gdp_growth_rate_qoq_adv',

	# Careful with spelling here
	'GDP Growth Rate QoQ 1st Est' : 'gdp_growth_rate_qoq_est',
	'GDP Growth Rate QoQ 2nd Est' : 'gdp_growth_rate_qoq_est',
	'GDP Growth Rate QoQ 3rd Est' : 'gdp_growth_rate_qoq_est',
	'GDP Growth Rate QoQ 4th Est' : 'gdp_growth_rate_qoq_est'
	}


	event_posneg_map : Dict = {
	'core_inflation_rate_mom' : EconomicCalendarEventPosNeg.SMALLER_BETTER,
	'core_inflation_rate_yoy' : EconomicCalendarEventPosNeg.SMALLER_BETTER,
	'inflation_rate_mom' : EconomicCalendarEventPosNeg.SMALLER_BETTER,
	'inflation_rate_yoy' : EconomicCalendarEventPosNeg.SMALLER_BETTER,
	'fed_interest_rate_decision' : EconomicCalendarEventPosNeg.SMALLER_BETTER,
	'core_pce_price_index_mom' : EconomicCalendarEventPosNeg.SMALLER_BETTER,
	'core_pce_price_index_yoy' : EconomicCalendarEventPosNeg.SMALLER_BETTER,
	'unemployment_rate' : EconomicCalendarEventPosNeg.BIGGER_BETTER, # more people out of job better as FEDs in position to ease.
	'non_farm_payrolls' : EconomicCalendarEventPosNeg.BIGGER_BETTER, # more people out of job better as FEDs in position to ease.
	'gdp_growth_rate_qoq_final' : EconomicCalendarEventPosNeg.BIGGER_BETTER,
	'gdp_growth_rate_qoq' : EconomicCalendarEventPosNeg.BIGGER_BETTER,
	'gdp_growth_rate_qoq_adv' : EconomicCalendarEventPosNeg.BIGGER_BETTER,
	'gdp_growth_rate_qoq_est' : EconomicCalendarEventPosNeg.BIGGER_BETTER
	}

	existing_csv : bool = os.path.isfile(export_csv_file_name)
	if existing_csv:
	pd_old_economic_calendars = pd.read_csv(export_csv_file_name)

	pd_new_economic_calendars = exchange.fetch_economic_calendar(
	regions=['united_states', 'china'],
	importance='high',
	start_ts=start_ts, end_ts=end_ts,
	region_map=region_map,
	event_name_map=event_name_map,
	event_posneg_map=event_posneg_map,
	rate_limit_winodw_sec=5
	)



	min_calendar_item_timestamp_ms = pd_new_economic_calendars['calendar_item_timestamp_ms'].min()

	if existing_csv:
	pd_old_economic_calendars = pd_old_economic_calendars[pd_old_economic_calendars['calendar_item_timestamp_ms']<min_calendar_item_timestamp_ms]
	pd_economic_calendars = pd.concat([pd_old_economic_calendars, pd_new_economic_calendars], axis=0)
	else:
	pd_economic_calendars = pd_new_economic_calendars

	print(f"{tabulate(pd_economic_calendars, headers='keys', tablefmt='psql')}")
	pd_economic_calendars.to_csv(export_csv_file_name)