def indicator_macd():
check_result = checkParameters(
args=request.args,
parametersList=['timestamp', 'symbol'],
)
if check_result:
return jsonify(check_result)
check_result = checkSymbol(request.args['symbol'])
if check_result:
return jsonify(check_result)
check_result = checkTimestamp(request.args['timestamp'])
if check_result:
return jsonify(check_result)
check_result = checkDate(request.args['symbol'], request.args['timestamp'])
if check_result:
return jsonify(check_result)
# timestamp = arrow.get(request.args.get('timestamp', 0))
# data_list = []
period_first = 12
period_second = 26
r_first = getDailyData(request.args['symbol'], request.args['timestamp'],
period_first)
r_second = getDailyData(request.args['symbol'], request.args['timestamp'],
period_second)
res = {
'type': 'result',
'time': arrow.utcnow().isoformat(),
'result': {
'symbol':
request.args.get('symbol', 'ERROR'),
'indicator':
'MACD',
'timestamp':
arrow.get(request.args['timestamp']).isoformat(),
'data':
idut.get_macd(request.args['symbol'], request.args['timestamp'])
}
}
return jsonify(res)
def indicator_vr():
check_result = checkParameters(
args=request.args,
parametersList=['timestamp', 'symbol'],
)
if check_result:
return jsonify(check_result)
check_result = checkSymbol(request.args['symbol'])
if check_result:
return jsonify(check_result)
check_result = checkTimestamp(request.args['timestamp'])
if check_result:
return jsonify(check_result)
check_result = checkDate(request.args['symbol'], request.args['timestamp'])
if check_result:
return jsonify(check_result)
# timestamp = arrow.get(request.args.get('timestamp', 0))
# data_list = []
period = int(request.args.get('period', 24))
r = getDailyData(request.args['symbol'], request.args['timestamp'],
period + 1)
res = {
'type': 'result',
'time': arrow.utcnow().isoformat(),
'result': {
'symbol':
request.args.get('symbol', 'ERROR'),
'indicator':
'VR',
'timestamp':
arrow.get(request.args['timestamp']).isoformat(),
'data':
Indicator.VR(price=np.float_(r['close'][-1]),
historical_price=np.array(r['close'][:-1]),
historical_volume=np.array(r['volume'][:-1]))
}
}
return jsonify(res)
def daily_data():
check_result = checkParameters(
args=request.args,
parametersList=[
'symbol',
],
)
if check_result:
return jsonify(check_result)
check_result = checkSymbol(request.args['symbol'])
if check_result:
return jsonify(check_result)
return jsonify(
getDailyData(request.args['symbol'],
request.args.get('timestamp', None)))
def indicator_ema():
check_result = checkParameters(
args=request.args,
parametersList=['timestamp', 'symbol'],
)
if check_result:
return jsonify(check_result)
check_result = checkSymbol(request.args['symbol'])
if check_result:
return jsonify(check_result)
check_result = checkTimestamp(request.args['timestamp'])
if check_result:
return jsonify(check_result)
check_result = checkDate(request.args['symbol'], request.args['timestamp'])
if check_result:
return jsonify(check_result)
# timestamp = arrow.get(request.args.get('timestamp', 0))
# data_list = []
period = int(request.args.get('period', 10))
r = getDailyData(request.args['symbol'], request.args['timestamp'],
period + 1)
res = {
'type': 'result',
'time': arrow.utcnow().isoformat(),
'result': {
'symbol': request.args.get('symbol', 'ERROR'),
'indicator': 'EMA',
'timestamp': arrow.get(request.args['timestamp']).isoformat(),
'data': idut.get_ema(request.args['symbol'],
request.args['timestamp'])
}
}
return jsonify(res)
def predict():
# check parameters
check_result = checkParameters(
args=request.args,
parametersList=['symbol', 'term', 'timestamp'],
parameterOptions={'term': ['short', 'long']})
if check_result:
return jsonify(check_result)
check_result = checkSymbol(request.args['symbol'])
if check_result:
return jsonify(check_result)
check_result = checkTimestamp(request.args['timestamp'])
if check_result:
return jsonify(check_result)
if request.args['term'] == 'short':
r = getDailyData(request.args['symbol'], request.args['timestamp'], 50)
time = np.array(r['timestamp']).reshape(-1, 1)
price = np.array(r['open'])
else:
r = getDailyData(request.args['symbol'], request.args['timestamp'],
252)
time = np.array(r['timestamp']).reshape(-1, 1)
price = np.array(r['open'])
predict_time = arrow.get(request.args['timestamp']).timestamp
bayes = pool.apply_async(
Predictor.bayesian_ridge,
[time, price, np.array(predict_time).reshape(-1, 1)]).get()
svr = pool.apply_async(
Predictor.SVR,
[time, price, np.array(predict_time).reshape(-1, 1)]).get()
dnn = pool.apply_async(
Predictor.DNN,
[time, price, np.array(predict_time).reshape(-1, 1)]).get()
#ans = (bayes[0] + svr[0]) / 2 if bayes[0] < 0 else (bayes[0] + svr[0] + dnn[0]) / 3
bayes[0] = (svr[0] + dnn[0]) if bayes[0] < 0 else bayes[0]
if request.args['term'] == 'short':
ans = 0.1 * bayes[0] + 0.7 * svr[0] + 0.2 * dnn[0]
elif request.args['term'] == 'long':
ans = 0.1 * bayes[0] + 0.2 * svr[0] + 0.7 * dnn[0]
else:
ans = 0
res = {
'type': 'result',
'time': arrow.utcnow().isoformat(),
'result': {
'symbol':
request.args.get('symbol'),
'predictPrice':
ans,
'predictor': [{
'name': 'bayes',
'price': bayes[0]
}, {
'name': 'Support Vector Regression',
'price': svr[0]
}, {
'name': 'Deep Neural Network',
'price': dnn[0]
}],
'note':
'ONLY FOR TESTING!',
'timestamp':
arrow.get(request.args['timestamp']).isoformat()
}
}
return jsonify(res)