add time and sample windowing mode examples using same dataset

Author: omeryagmurlu

example/__init__.py examples/legacy-example/__init__.pyexample/__init__.py renamed to examples/legacy-example/__init__.py

@@ -1,7 +1,7 @@
 from datetime import datetime
 import os
 import sys
-sys.path.insert(1, os.path.join(sys.path[0], '../src')) # add edgeml to path
+sys.path.insert(1, os.path.join(sys.path[0], '../../src')) # add edgeml to path
 
 import time as timelib
 import pandas as pd

example/model_python.py examples/legacy-example/model_python.pyexample/model_python.py renamed to examples/legacy-example/model_python.py

@@ -0,0 +1,10 @@
+### shake-example
+
+This example uses a csv data as input and does inference on the data. Input csv file, as well as the dataset used for training the model can be found in the directory. This dataset is used to create two different models, one with time based windowing, other with sample based windowing.
+
+#### Usage:
+
+```
+❯ python sample_mode_example.py
+❯ python time_mode_example.py
+```

example/prediction_example.py …les/legacy-example/prediction_example.pyexample/prediction_example.py renamed to examples/legacy-example/prediction_example.py example/prediction_example.py renamed to examples/legacy-example/prediction_example.py

@@ -0,0 +1,32 @@
+from datetime import datetime
+import os
+import sys
+sys.path.insert(1, os.path.join(sys.path[0], '../../src')) # add edgeml to path
+
+import time as timelib
+import pandas as pd
+from edgeml.predictor import Predictor, PredictorError
+
+from shakesample_python import score
+
+p = Predictor(
+    predictor = lambda input: score(input),
+    sensors = ['AccelerometerX', 'AccelerometerY', 'AccelerometerZ', 'GyroscopeX', 'GyroscopeY', 'GyroscopeZ'],
+    window_size = 100,
+    labels = ['shake', 'STILL'],
+    scaler = {'scale': [456.89999999999986, 1.7000000000000002, 4.568999999999999, 1.0, 16.37991667923493, 272.6723773333334, 20.55429975152452, 43.50000000000001, 43.50000000000001, 37.7, 553.2333333333333, 4.800000000000001, 5.532333333333334, 1.0, 8.87448778931522, 80.51507597222225, 5.233021007637985, 14.399999999999999, 23.9, 29.300000000000004, 458.1333333333331, 1.1416666666666693, 4.58133333333333, 1.0, 17.97626482222193, 331.8784097777778, 10.897358557974057, 61.400000000000006, 61.400000000000006, 36.7, 27.92177737340528, 0.02923426497090502, 0.27921777373405277, 1.0, 5.674080882891607, 32.662245725674715, 6.131156197061794, 12.037535851004892, 15.02379420116719, 15.063936773963059, 31.333023396428203, 0.019634954084936207, 0.31333023396428206, 1.0, 8.905398086799991, 80.0946344892207, 8.90367312352328, 19.853120241435498, 19.853120241435498, 18.327702475192456, 17.675821999572577, 0.013962634015954637, 0.17675821999572577, 1.0, 5.575594915257246, 31.388752739250815, 5.869938977435963, 12.514010736799342, 12.514010736799342, 12.006119924468992], 'center': [156.8, 1.1, 1.568, 100.0, 10.265825965308393, 105.38718275000001, 11.602474951492033, 25.200000000000003, 25.200000000000003, -11.9, 180.99999999999994, 2.2, 1.8099999999999994, 100.0, 4.214677396379888, 17.763505555555554, 4.944580366421402, 13.200000000000001, 14.0, -7.9, 954.2499999999999, 9.2, 9.542499999999999, 100.0, 6.213705056387677, 38.610130527777784, 13.107873693996973, 29.5, 29.5, -12.200000000000001, -0.029670597283903654, -0.0008726646259971647, -0.0002967059728390365, 100.0, 1.607205387675301, 2.5831091581725145, 1.6081365012688127, 3.286454981505323, 11.082840750163992, -3.2951816277652943, 0.29466975537837603, 0.0017453292519943296, 0.0029466975537837605, 100.0, 3.250629986683856, 10.566595310328285, 3.2519927095577077, 6.485643500410929, 9.889035541799872, -9.321803534901715, -0.061959188445798744, 0.006981317007977318, -0.0006195918844579874, 100.0, 3.978628992873262, 15.829488662931706, 4.101449761530525, 10.10196571054318, 10.567968620825665, -5.46637121724624], 'name': 'RobustScaler'},
+    windowing_mode = "sample"
+)
+
+test = pd.read_csv('./test.csv', index_col=0)
+
+for t, row in test.iterrows():
+    time = int(t) / 1e3
+    for key, valStr in row.iteritems():
+        p.add_datapoint(key, float(valStr), time)
+    
+    try:
+        print(datetime.fromtimestamp(time), p.predict())
+        timelib.sleep(0.01)
+    except PredictorError:
+        pass