augmen
commented
4 years ago I am still struggling connecting with websocket as the websockets fetches data in real time and in json.
First Approach I have converted the json strings into array like this requests.get('http://localhost:8005/trade?data={}'.format([close,volume])).json()) and try to feed them into app.py flask server using below :
@app.route('/trade', methods = ['GET']) def trade(): data = json.loads(request.args.get('data')) return jsonify(agent.trade(data))
at the same time we are feeding the app.py with .csv training file in below line :
df = pd.read_csv('TWTR.csv') real_trend = df['Close'].tolist() parameters = [df['Close'].tolist(), df['Volume'].tolist()] minmax = MinMaxScaler(feature_range = (100, 200)).fit(np.array(parameters).T) scaled_parameters = minmax.transform(np.array(parameters).T).T.tolist() initial_money = np.max(parameters[0]) * 2
Second approach we have tried to save live data to DB and create a function called def trade_iteration which fetches batches of data from DB , converts it intopandas Dataframe and feeds into app.py via below line 350-355;
df = get_last_n_records_for_symbol('ETHUSDT', n=RECORDS_TO_FEED) print(df.shape) RUN_INTERVAL_TIME_IN_SEC = 6 RECORDS_TO_FEED = 30 real_trend = df['close'].tolist() print(real_trend) parameters = [df['close'].tolist(), df['volume'].tolist()] minmax = MinMaxScaler(feature_range=(100, 200)).fit(np.array(parameters).T) scaled_parameters = minmax.transform(np.array(parameters).T).T.tolist() initial_money = 10000
in both the above approaches we have make sure that son data is in ['close', 'volume'] dictionary .
The problem is Dataframe is static data while json data is dynamic !
Can you guide us what's the right method and format of data to put into app.py from websocket.
we are still not sure for hoe the app.py generates results through agent.trade()
Also what is this what variables or function it is used for ;
@app.route('/queue', methods = ['GET']) def queue(): return jsonify(agent._queue)
It looks like a queue for processing multiple iterations on the neural network. I thought it was a method.
@huseinzol05 @AlconDivino @marvin-hansen Plz guide
marvin-hansen
i am stuck with the code : i want to know we are feeding live data to the agent.trade() function with request.ipynb then why we need the same data again to be feed here in df ?
df = pd.read_csv('TWTR.csv') real_trend = df['Close'].tolist() parameters = [df['Close'].tolist(), df['Volume'].tolist()] minmax = MinMaxScaler(feature_range = (100, 200)).fit(np.array(parameters).T) scaled_parameters = minmax.transform(np.array(parameters).T).T.tolist() initial_money = np.max(parameters[0]) * 2here we are using the same csv file right ? or we are comparing past data with live data ?
@huseinzol05 pax guide or anyone can guide on right method to feed live data .