search

nayash / stock_cnn_blog_pub

This project is a loose implementation of paper "Algorithmic Financial Trading with Deep Convolutional Neural Networks: Time Series to Image Conversion Approach"
Apache License 2.0
165 stars 95 forks source link

I run into a index error when running stock_keras.ipynb first cell. #2

Open louis78100 opened 4 years ago

louis78100 commented 4 years ago

When I run the first cell of the notebook stock_keras.ipynb, i got the following error. This is appearing after doing all the labels for every technical indicators except the IBR.

Here is the context : running stock_cnn.py ['./train.py', 'WMT', 'original'] ../stock_history/WMT ../outputs/fresh_rolling_train ../stock_history/WMT/WMT.csv 1 ) 28-07-2020 16_40_55 MainThread INFO\ Initialized logging at path ../outputs/fresh_rolling_train/logs/log_WMT_original_fresh_rolling_train_28-07-2020_16_40_55.log Tensorflow devices path to company data: ../stock_history/WMT/WMT.csv 2 ) 28-07-2020 16_40_55 MainThread DEBUG\ Data for WMT ready to use 3 ) 28-07-2020 16_40_55 MainThread DEBUG\ Technical indicators already calculated. Loading... 4 ) 28-07-2020 16_40_55 MainThread DEBUG\ Dropped 26 nan rows before label calculation 5 ) 28-07-2020 16_40_55 MainThread DEBUG\ creating label with original paper strategy Calculating labels HBox(children=(IntProgress(value=0, max=5007), HTML(value='')))

IndexError Traceback (most recent call last) ~/Downloads/stock_cnn_blog_pub-master/src/train.py in 71 72 start_time = time.time() ---> 73 data_gen = DataGenerator(company_code, PATH_TO_COMPANY_DATA, OUTPUT_PATH, strategy_type, False, logger) 74 # exit here, since training is done with stock_keras.ipynb. 75 # comment sys.exit() if you want to try out rolling window training.

~/Downloads/stock_cnn_blog_pub-master/src/data_generator.py in init(self, company_code, data_path, output_path, strategy_type, update, logger) 40 self.update = update 41 self.download_stock_data() ---> 42 self.df = self.create_features() 43 self.feat_idx = self.feature_selection() 44 self.one_hot_enc = OneHotEncoder(sparse=False, categories='auto')

~/Downloads/stock_cnn_blog_pub-master/src/data_generator.py in create_features(self) 266 short = self.strategytype.split('')[0] 267 long = self.strategytype.split('')[1] --> 268 df['labels'] = self.create_label_short_long_ma_crossover(df, 'close', short, long) 269 else: 270 df['labels'] = self.create_labels(df, 'close')

~/Downloads/stock_cnn_blog_pub-master/src/data_generator.py in create_labels(self, df, col_name, window_size) 142 143 labels[window_middle] = 1 --> 144 else: 145 146 labels[window_middle] = 2

IndexError: only integers, slices (:), ellipsis (...), numpy.newaxis (None) and integer or boolean arrays are valid indices

catalinmut commented 4 years ago

In your while loop, window_middle is divided by 2 and it becomes a float. Cast it back as an integer or you can try to use int(window_middle) instead of window_middle