doc: update readme

on three sessions after testing on one

Handle_emg_data.py

@@ -9,6 +9,7 @@ import sys
 sys.path.insert(0, '/Users/Markus/Prosjekter git/Slovakia 2021/psf_lib/python_speech_features/python_speech_features')
 from psf_lib.python_speech_features.python_speech_features import mfcc
 import json
+import os
 
 
 # Global variables for MFCC
@@ -21,41 +22,39 @@ class Data_container:
     
     # Initiates personal data container for each subject. Dict for each session with keys 'left' and 'right',
     # and values equal to lists of EMG data indexed 0-7  
-    # NB! More sessions has to be added here in the future   
-    def __init__(self, subject_nr:int, subject_name:str):
+    def __init__(self, subject_nr:int, subject_name:str, nr_sessions:int):
         self.subject_nr = subject_nr
         self.subject_name = subject_name
-        self.data_dict_round1 = {'left': [None]*8, 'right': [None]*8}
-        self.data_dict_round2 = {'left': [None]*8, 'right': [None]*8}
-        self.data_dict_round3 = {'left': [None]*8, 'right': [None]*8}
-        self.data_dict_round4 = {'left': [None]*8, 'right': [None]*8}
-        self.dict_list =   [self.data_dict_round1, 
-                            self.data_dict_round2, 
-                            self.data_dict_round3, 
-                            self.data_dict_round4
-                            ]
+        self.dict_list =  [{'left': [None]*8, 'right': [None]*8} for i in range(nr_sessions)]
 
+    def __str__(self) -> str:
+        return 'Name: {}, \tID: {}'.format(self.subject_name, self.subject_nr)
 
 class CSV_handler:
 
     # Initiates object to store all datapoints in the experiment
-    def __init__(self):
+    def __init__(self, nr_subjects:int, nr_sessions:int):
         self.working_dir = str(Path.cwd())
-        self.data_container_dict = {}   # Dict with keys equal subject numbers and values equal to its respective datacontainer
-        self.data_type = None   # String describing which type of data is stored in the object
+        self.nr_subjects = nr_subjects
+        self.nr_sessions = nr_sessions
+        # Dict with keys equal subject numbers and values equal to its respective datacontainer 
+        self.data_container_dict = {i: None for i in range(nr_subjects)}  
+        # String describing which type of data is stored in the object 
+        self.data_type = None   
 
     # Makes dataframe from the csv files in the working directory
     # Input: filename of a csv-file
     # Output: DataFrame
     def make_df(self, filename):
-        filepath = self.working_dir + str(filename)
+        #filepath = self.working_dir + str(filename)
+        filepath = str(filename)
         df = pd.read_csv(filepath)
         return df
 
     # Extracts out the timestamp and the selected emg signal into a new dataframe
     # Input: filename of a csv-file, EMG nr 
     # Output: DataFrame(timestamp/EMG)
-    def get_time_emg_table(self, filename:str, emg_nr:int):
+    def get_emg_table_from_file(self, filename:str, emg_nr:int):
         tot_data_frame = self.make_df(filename)
         emg_str = 'emg' + str(emg_nr)
         filtered_df = tot_data_frame[["timestamp", emg_str]]
@@ -65,34 +64,14 @@ class CSV_handler:
     # Input: filename of a csv-file, EMG nr, left/right arm, subject's data_container, session nr 
     # Output: None -> stores EMG data in data container
     def store_df_in_container(self, filename:str, emg_nr:int, which_arm:str, data_container:Data_container, session:int):
-        df = self.get_time_emg_table(filename, emg_nr+1)
+        df = self.get_emg_table_from_file(filename, emg_nr+1)
 
         if df.isnull().values.any():
             print('NaN in: subject', data_container.subject_nr, 'arm:', which_arm, 'session:', session, 'emg nr:', emg_nr)
 
         # Places the data correctly:
-        if session == 1:
-            if which_arm == 'left':
-                data_container.data_dict_round1['left'][emg_nr] = df    # Zero indexed emg_nr in the dict
-            else:
-                data_container.data_dict_round1['right'][emg_nr] = df
-        elif session == 2:
-            if which_arm == 'left':
-                data_container.data_dict_round2['left'][emg_nr] = df
-            else:
-                data_container.data_dict_round2['right'][emg_nr] = df
-        elif session == 3:
-            if which_arm == 'left':
-                data_container.data_dict_round3['left'][emg_nr] = df
-            else:
-                data_container.data_dict_round3['right'][emg_nr] = df
-        elif session == 4:
-            if which_arm == 'left':
-                data_container.data_dict_round4['left'][emg_nr] = df
-            else:
-                data_container.data_dict_round4['right'][emg_nr] = df
-        else:
-            raise IndexError('Not a valid index')
+        data_container.dict_list[session-1][which_arm][emg_nr] = df
+
     
     # Links the data container for a subject to the csv_handler object
     # Input: the subject's data_container
@@ -110,9 +89,55 @@ class CSV_handler:
         df = container.dict_list[session - 1].get(which_arm)[emg_nr - 1]
         return df
 
-    # Loads the data from the csv files into the storing system of the CSV_handler object
-    # Input: None(CSV_handler)
+
+    # Loads data the to the CSV_handler(general load func). Choose data_type: hard, hardPP, soft og softPP as str. 
+    # Input: String(datatype you want), direction name of that type
     # Output: None -> load and stores data
+    def load_data(self, type:str, type_dir_name:str):
+
+        data_path = self.working_dir + '/data/' + type_dir_name
+        subject_id = 100
+        subject_name = 'bruh'
+        nr_sessions = 101
+        container = None
+        session_count = 0
+
+        for i, (path, subject_dir, session_dir) in enumerate(os.walk(data_path)):
+            
+            if path is not data_path:
+                
+                if subject_dir:
+                    session_count = 0
+                    subject_id = int(path[-1])
+                    subject_name = subject_dir[0].split('_')[0]
+                    nr_sessions = len(subject_dir)
+                    container = Data_container(subject_id, subject_name, nr_sessions)
+                    continue
+                else:
+                    session_count += 1
+
+                for f in session_dir:
+                    spes_path = os.path.join(path, f)
+                    if f == 'myoLeftEmg.csv':
+                        for emg_nr in range(8):
+                            self.store_df_in_container(spes_path, emg_nr, 'left', container, session_count)
+                    elif f == 'myoRightEmg.csv':
+                        for emg_nr in range(8):
+                            self.store_df_in_container(spes_path, emg_nr, 'right', container, session_count)
+                self.link_container_to_handler(container)
+        self.data_type = type
+        return self.data_container_dict
+
+    # Retrieved data. Send in loaded csv_handler and data detailes you want. 
+    # Input: Experiment detailes
+    # Output: DataFrame,  samplerate:int
+    def get_data(self, subject_nr, which_arm, session, emg_nr):
+        data_frame = self.get_df_from_data_dict(subject_nr, which_arm, session, emg_nr)
+        samplerate = get_samplerate(data_frame)
+        return data_frame, samplerate
+
+
+    # OBSOLETE
     def load_hard_PP_emg_data(self):
 
         # CSV data from subject 1
@@ -477,11 +502,7 @@ class CSV_handler:
             self.link_container_to_handler(data_container)
         self.data_type = 'soft'
         return self.data_container_dict
-
-    # Loads data the to the CSV_handler(general load func). Choose data_type: hard, hardPP, soft og softPP as str. 
-    # Input: String(datatype you want)
-    # Output: None -> load and stores data
-    def load_data(self, data_type):
+    def load_data_OLD(self, data_type):
         if data_type == 'hard': 
             self.load_hard_original_emg_data()
         elif data_type == 'hardPP':
@@ -493,13 +514,6 @@ class CSV_handler:
         else:
             raise Exception('Wrong input')
 
-    # Retrieved data. Send in loaded csv_handler and data detailes you want. 
-    # Input: Experiment detailes
-    # Output: DataFrame,  samplerate:int
-    def get_data(self, subject_nr, which_arm, session, emg_nr):
-        data_frame = self.get_df_from_data_dict(subject_nr, which_arm, session, emg_nr)
-        samplerate = get_samplerate(data_frame)
-        return data_frame, samplerate
 
     # NOT IMPLEMENTED
     def get_keyboard_data(self, filename:str, pres_or_release:str='pressed'):
@@ -514,31 +528,13 @@ class CSV_handler:
 class NN_handler:
 
     # Paths for data storage in json to later use in Neural_Network_Analysis.py
-    JSON_PATH_REG = "reg_data.json"
     JSON_PATH_MFCC = "mfcc_data.json"
 
     # Class to manipulate data from the CSV_handler and store it for further analysis
-    # NB! More subject needs to be added manually 
     def __init__(self, csv_handler:CSV_handler) -> None:
         self.csv_handler = csv_handler
-        # Should med 4 sessions * split nr of samples per person. Each sample is structured like this: [sample_df, samplerate]
-        self.reg_samples_per_subject = {1: [],  
-                                        2: [], 
-                                        3: [],
-                                        4: [],
-                                        5: []
-                                        }
         # Should med 4 sessions * (~150, 208) of mfcc samples per person. One [DataFrame, session_length_list] per subject
-        self.mfcc_samples_per_subject = {1: [],  
-                                         2: [], 
-                                         3: [],
-                                         4: [],
-                                         5: []
-                                         }
-
-    # GET method for reg_samples_dict
-    def get_reg_samples_dict(self) -> dict:
-        return self.reg_samples_per_subject
+        self.mfcc_samples_per_subject = {k+1:[] for k in range(csv_handler.nr_subjects)}
 
     # GET method for mfcc_samples_dict  
     def get_mfcc_samples_dict(self) -> dict:
@@ -592,39 +588,6 @@ class NN_handler:
 
         return tot_session_df
     
-    # Takes in all EMG session Dataframe and merges the EMG data into one column, creating one signal 
-    # Input: DataFrame(shape[1]=16, EMG data)
-    # Output: DataFrame(signal), samplerate of it
-    def reshape_session_df_to_signal(self, df:DataFrame):
-            main_df = df[['timestamp', 1]].rename(columns={1: 'emg'})
-            for i in range(2, 17):
-                adding_df = df[['timestamp', i]].rename(columns={i: 'emg'})
-                main_df = pd.concat([main_df, adding_df], ignore_index=True)
-            samplerate = get_samplerate(main_df)
-            return main_df, samplerate
-
-    # Stores split, merged signals in the NN-handler's reg_samples_per_subject
-    # Input: Split_nr:int(how many times to split this merged signal)
-    # Output: None -> stores in NN_handler
-    def store_samples(self, split_nr) -> None:
-        for subject_nr in range(5):
-            subj_samples = []
-            for session_nr in range(4):
-                list_of_emg = self.get_emg_list(subject_nr+1, session_nr+1)
-                tot_session_df = self.make_subj_sample(list_of_emg)
-
-                # TESTING FOR NAN
-                if tot_session_df.isnull().values.any():
-                    print('NaN in: subject', subject_nr+1, 'session:', session_nr+1, 'where? HERE')
-
-                samples = np.array_split(tot_session_df.to_numpy(), split_nr)
-                for array in samples:
-                    df = DataFrame(array).rename(columns={0:'timestamp'})
-                    df_finished, samplerate = self.reshape_session_df_to_signal(df)
-                    subj_samples.append([df_finished, samplerate])
-            
-            self.reg_samples_per_subject[subject_nr+1] = subj_samples
-
     # Takes in all EMG session Dataframe and creates DataFrame of MFCC samples
     # Input: DataFrame(shape[1]=16, EMG data)
     # Output: DataFrame(merged MFCC data, shape: (n, 13*16)), length of session datapoints
@@ -673,11 +636,75 @@ class NN_handler:
             result_df = pd.concat(subj_samples, axis=0, ignore_index=True)
             self.mfcc_samples_per_subject[subject_nr+1] = [result_df, session_length_list]
 
-
-    # Makes MFCC data from reg_samples_per_subject and stores it in a json file
+    # Stores MFCC data from mfcc_samples_per_subject in a json file
     # Input: Path to the json file
     # Output: None -> stores in json
-    def save_json_reg(self, json_path=JSON_PATH_REG):
+    def save_json_mfcc(self, json_path=JSON_PATH_MFCC):
+            
+            # dictionary to store mapping, labels, and MFCCs
+            data = {
+                "mapping": [],
+                "labels": [],
+                "mfcc": [],
+
+                "session_lengths": []
+            }
+
+            raw_data_dict = self.get_mfcc_samples_dict()
+        
+            # loop through all subjects to get samples
+            for key, value in raw_data_dict.items():
+                
+                # save subject label in the mapping
+                subject_label = 'Subject ' + str(key)
+                print("\nProcessing: {}".format(subject_label))
+                data["mapping"].append(subject_label)       # Subject label
+                data["session_lengths"].append(value[1])   # List[subject][session_length_list]
+
+                # process all samples per subject
+                for i, sample in enumerate(value[0]):
+
+                    data["labels"].append(key-1)    # Subject nr
+                    data["mfcc"].append(sample)  # MFCC sample on same index
+                     
+                    print("sample:{} is done".format(i+1))
+                    #print(np.array(mfcc_data).shape)
+
+            # save MFCCs to json file
+            with open(json_path, "w") as fp:
+                json.dump(data, fp, indent=4)
+
+
+
+    # OBSOLETE
+    def get_reg_samples_dict(self) -> dict:
+        return self.reg_samples_per_subject
+    def reshape_session_df_to_signal(self, df:DataFrame):
+            main_df = df[['timestamp', 1]].rename(columns={1: 'emg'})
+            for i in range(2, 17):
+                adding_df = df[['timestamp', i]].rename(columns={i: 'emg'})
+                main_df = pd.concat([main_df, adding_df], ignore_index=True)
+            samplerate = get_samplerate(main_df)
+            return main_df, samplerate
+    def store_samples(self, split_nr) -> None:
+        for subject_nr in range(5):
+            subj_samples = []
+            for session_nr in range(4):
+                list_of_emg = self.get_emg_list(subject_nr+1, session_nr+1)
+                tot_session_df = self.make_subj_sample(list_of_emg)
+
+                # TESTING FOR NAN
+                if tot_session_df.isnull().values.any():
+                    print('NaN in: subject', subject_nr+1, 'session:', session_nr+1, 'where? HERE')
+
+                samples = np.array_split(tot_session_df.to_numpy(), split_nr)
+                for array in samples:
+                    df = DataFrame(array).rename(columns={0:'timestamp'})
+                    df_finished, samplerate = self.reshape_session_df_to_signal(df)
+                    subj_samples.append([df_finished, samplerate])
+            
+            self.reg_samples_per_subject[subject_nr+1] = subj_samples
+    def save_json_reg(self, json_path):
         
         # Dictionary to store mapping, labels, and MFCCs
         data = {
@@ -730,43 +757,6 @@ class NN_handler:
         with open(json_path, "w") as fp:
             json.dump(data, fp, indent=4)
 
-    # Stores MFCC data from mfcc_samples_per_subject in a json file
-    # Input: Path to the json file
-    # Output: None -> stores in json
-    def save_json_mfcc(self, json_path=JSON_PATH_MFCC):
-            
-            # dictionary to store mapping, labels, and MFCCs
-            data = {
-                "mapping": [],
-                "labels": [],
-                "mfcc": [],
-
-                "session_lengths": []
-            }
-
-            raw_data_dict = self.get_mfcc_samples_dict()
-        
-            # loop through all subjects to get samples
-            for key, value in raw_data_dict.items():
-                
-                # save subject label in the mapping
-                subject_label = 'Subject ' + str(key)
-                print("\nProcessing: {}".format(subject_label))
-                data["mapping"].append(subject_label)       # Subject label
-                data["session_lengths"].append(value[1])   # List[subject][session_length_list]
-
-                # process all samples per subject
-                for i, sample in enumerate(value[0]):
-
-                    data["labels"].append(key-1)    # Subject nr
-                    data["mfcc"].append(sample)  # MFCC sample on same index
-                     
-                    print("sample:{} is done".format(i+1))
-                    #print(np.array(mfcc_data).shape)
-
-            # save MFCCs to json file
-            with open(json_path, "w") as fp:
-                json.dump(data, fp, indent=4)
 
 
 # HELP FUNCTIONS: ------------------------------------------------------------------------: 

Neural_Network_Analysis.py

@@ -12,6 +12,7 @@ from keras.callbacks import Callback, CSVLogger, ModelCheckpoint
 from pathlib import Path
 import pandas as pd
 import matplotlib.pyplot as plt
+#from matplotlib.legend import _get_legend_handles_
 import statistics
 import csv
 
@@ -66,6 +67,185 @@ def plot_train_history(history, val_data=False):
 
     plt.show()
 
+# Plots the training history of four networks inverse cross-validated
+# Input: data, nr of sessions in total, batch_size and epochs
+# Ouput: None -> plot
+def plot_4_x_inverse_cross_val(X, y, session_lengths, nr_sessions, batch_size=64, epochs=30):
+
+    history_dict = {'GRU': [],
+                    'LSTM': [],
+                    'FFN': [],
+                    'CNN_1D': []}
+    
+    for i in range(nr_sessions):
+
+        X_test_session, X_train_session, y_test_session, y_train_session = prepare_datasets_sessions(X, y, session_lengths, i)
+
+        model_GRU = GRU(input_shape=(1, 208))
+        GRU_h = train(model_GRU, X_train_session, y_train_session, 1, batch_size=batch_size, epochs=epochs)
+        history_dict['GRU'].append(GRU_h)
+        del model_GRU
+        K.clear_session()
+
+        model_LSTM = LSTM(input_shape=(1, 208))
+        LSTM_h = train(model_LSTM, X_train_session, y_train_session, 1, batch_size=batch_size, epochs=epochs)
+        history_dict['LSTM'].append(LSTM_h)
+        del model_LSTM
+        K.clear_session()
+
+        model_FFN = FFN(input_shape=(1, 208)) 
+        FFN_h = train(model_FFN, X_train_session, y_train_session, 1, batch_size=batch_size, epochs=epochs)
+        history_dict['FFN'].append(FFN_h)
+        del model_FFN
+        K.clear_session()
+
+        model_CNN_1D = CNN_1D(input_shape=(208, 1))
+        X_train_session = np.reshape(X_train_session, (X_train_session.shape[0], 208, 1))
+        X_test_session = np.reshape(X_test_session, (X_test_session.shape[0], 208, 1))
+        CNN_1D_h = train(model_CNN_1D, X_train_session, y_train_session, 1, batch_size=batch_size, epochs=epochs)
+        history_dict['CNN_1D'].append(CNN_1D_h)
+        del model_CNN_1D
+        K.clear_session()
+    
+
+    fig, axs = plt.subplots(2, 2, sharey=True)
+    plt.ylim(0, 1)
+
+    # GRU plot:
+    axs[0, 0].plot(history_dict['GRU'][0].history["accuracy"])
+    axs[0, 0].plot(history_dict['GRU'][1].history["accuracy"], 'tab:orange')
+    axs[0, 0].plot(history_dict['GRU'][2].history["accuracy"], 'tab:green')
+    axs[0, 0].plot(history_dict['GRU'][3].history["accuracy"], 'tab:red')
+    axs[0, 0].set_title('GRU')
+    # LSTM plot:
+    axs[0, 1].plot(history_dict['LSTM'][0].history["accuracy"])
+    axs[0, 1].plot(history_dict['LSTM'][1].history["accuracy"], 'tab:orange')
+    axs[0, 1].plot(history_dict['LSTM'][2].history["accuracy"], 'tab:green')
+    axs[0, 1].plot(history_dict['LSTM'][3].history["accuracy"], 'tab:red')
+    axs[0, 1].set_title('LSTM')
+    # FFN plot: 
+    axs[1, 0].plot(history_dict['FFN'][0].history["accuracy"])
+    axs[1, 0].plot(history_dict['FFN'][1].history["accuracy"], 'tab:orange')
+    axs[1, 0].plot(history_dict['FFN'][2].history["accuracy"], 'tab:green')
+    axs[1, 0].plot(history_dict['FFN'][3].history["accuracy"], 'tab:red')
+    axs[1, 0].set_title('FFN')
+    # CNN_1D plot:
+    axs[1, 1].plot(history_dict['CNN_1D'][0].history["accuracy"])
+    axs[1, 1].plot(history_dict['CNN_1D'][1].history["accuracy"], 'tab:orange')
+    axs[1, 1].plot(history_dict['CNN_1D'][2].history["accuracy"], 'tab:green')
+    axs[1, 1].plot(history_dict['CNN_1D'][3].history["accuracy"], 'tab:red')
+    axs[1, 1].set_title('CNN_1D')
+
+    for ax in axs.flat:
+        ax.set(xlabel='Epochs', ylabel='Accuracy')
+
+    # Hide x labels and tick labels for top plots and y ticks for right plots.
+    for ax in axs.flat:
+        ax.label_outer()
+    
+    plt.show()
+    
+
+# Plots the average training history of four networks inverse cross-validated
+# Input: data, nr of sessions in total, batch_size and epochs
+# Ouput: None -> plot
+def plot_4_x_average_val(X, y, session_lengths, nr_sessions, batch_size=64, epochs=30):
+
+    history_dict = {'GRU_train': [],
+                    'LSTM_train': [],
+                    'FFN_train': [], 
+                    'CNN_1D_train': []}
+    history_dict_val = {'GRU_val': [],  
+                        'LSTM_val': [],
+                        'FFN_val': [], 
+                        'CNN_1D_val': []}
+    
+    for i in range(nr_sessions):
+        # Prepare data
+        X_val_session, X_train_session, y_val_session, y_train_session = prepare_datasets_sessions(X, y, session_lengths, i)
+        
+        # GRU
+        model_GRU = GRU(input_shape=(1, 208))
+        GRU_h = train(model_GRU, X_train_session, y_train_session, 1, batch_size=batch_size, epochs=epochs, 
+                                                            X_validation=X_val_session, y_validation=y_val_session)
+        history_dict['GRU_train'].append(GRU_h.history['accuracy'])
+        history_dict_val['GRU_val'].append(GRU_h.history['val_accuracy'])
+        del model_GRU
+        K.clear_session()
+        
+        # LSTM
+        model_LSTM = LSTM(input_shape=(1, 208))
+        LSTM_h = train(model_LSTM, X_train_session, y_train_session, 1, batch_size=batch_size, epochs=epochs, 
+                                                            X_validation=X_val_session, y_validation=y_val_session)
+        history_dict['LSTM_train'].append(LSTM_h.history['accuracy'])
+        history_dict_val['LSTM_val'].append(LSTM_h.history['val_accuracy'])
+        del model_LSTM
+        K.clear_session()
+        
+        # FFN
+        model_FFN = FFN(input_shape=(1, 208)) 
+        FFN_h = train(model_FFN, X_train_session, y_train_session, 1, batch_size=batch_size, epochs=epochs, 
+                                                            X_validation=X_val_session, y_validation=y_val_session)
+        history_dict['FFN_train'].append(FFN_h.history['accuracy'])
+        history_dict_val['FFN_val'].append(FFN_h.history['val_accuracy'])
+        del model_FFN
+        K.clear_session()
+        
+        # CNN_1D
+        model_CNN_1D = CNN_1D(input_shape=(208, 1))
+        X_train_session = np.reshape(X_train_session, (X_train_session.shape[0], 208, 1))
+        X_val_session = np.reshape(X_val_session, (X_val_session.shape[0], 208, 1))
+        CNN_1D_h = train(model_CNN_1D, X_train_session, y_train_session, 1, batch_size=batch_size, epochs=epochs, 
+                                                            X_validation=X_val_session, y_validation=y_val_session)
+        history_dict['CNN_1D_train'].append(CNN_1D_h.history['accuracy'])
+        history_dict_val['CNN_1D_val'].append(CNN_1D_h.history['val_accuracy'])
+        del model_CNN_1D
+        K.clear_session()
+    
+    # Averaging out session training for each network
+    for key in history_dict:
+        history_dict[key] = list(np.average([x, y, z, c]) for x, y, z, c in list(zip(*history_dict[key])))
+    for key in history_dict_val:
+        history_dict_val[key] = list(np.average([x, y, z, c]) for x, y, z, c in list(zip(*history_dict_val[key])))
+
+    '''
+    history_dict = {'GRU_train': [0.5, 0.8],
+                    'LSTM_train': [0.5, 0.9],
+                    'FFN_train': [0.75, 0.8], 
+                    'CNN_1D_train': [0.8, 0.95]}
+    history_dict_val = {'GRU_val': [0.5, 0.8],  
+                        'LSTM_val': [0.5, 0.9],
+                        'FFN_val': [0.75, 0.8], 
+                        'CNN_1D_val': [0.8, 0.95]}
+    '''
+
+    # Plot:
+    fig, axs = plt.subplots(2, sharey=True)
+    plt.ylim(0, 1)
+    plt.subplots_adjust(hspace=1.0, top=0.85, bottom=0.15, right=0.75)
+    fig.suptitle('Avarage accuracy with cross-session-training', fontsize=16)
+
+    axs[0].plot(history_dict['GRU_train'], label='GRU')
+    axs[0].plot(history_dict['LSTM_train'], 'tab:orange', label='LSTM')
+    axs[0].plot(history_dict['FFN_train'], 'tab:green', label='FFN')
+    axs[0].plot(history_dict['CNN_1D_train'], 'tab:red', label='CNN_1D')
+    axs[0].set_title('Training accuracy')
+    
+
+    axs[1].plot(history_dict_val['GRU_val'], label='GRU')
+    axs[1].plot(history_dict_val['LSTM_val'], 'tab:orange', label='LSTM')
+    axs[1].plot(history_dict_val['FFN_val'], 'tab:green', label='FFN')
+    axs[1].plot(history_dict_val['CNN_1D_val'], 'tab:red', label='CNN_1D')
+    axs[1].set_title('Validation accuracy')
+    
+    for ax in axs.flat:
+        ax.set(xlabel='Epochs', ylabel='Accuracy')
+    
+    plt.legend(bbox_to_anchor=(1.05, 1.5), title='Networks', loc='center left') 
+    plt.show()
+
+        
+
 # Takes in data and labels, and splits it into train, validation and test sets by percentage
 # Input: Data, labels, whether to shuffle, % validatiion, % test
 # Ouput: X_train, X_validation, X_test, y_train, y_validation, y_test
@@ -173,7 +353,7 @@ def train( model, X_train, y_train, verbose, batch_size=64, epochs=30,
     #csv_path = str(Path.cwd()) + '/logs/{}/{}_train_log.csv'.format(MODEL_NAME, MODEL_NAME)
     #csv_logger = CSVLogger(csv_path, append=False)
 
-    if X_validation != None:
+    if X_validation.any():
         history = model.fit(X_train, 
                             y_train, 
                             validation_data=(X_validation, y_validation), 
@@ -285,12 +465,57 @@ def session_cross_validation(model_name:str, X, y, session_lengths, nr_sessions,
 
     return average_result, session_training_results
 
+# Retrieves data sets for each session as train set and evalutes on the others.
+# the average of networks trained om them
+# Input: raw data, session_lengths list, total nr of sessions, batch_size, and nr of epochs 
+# Ouput: tuple(cross validation average, list(result for each dataset(len=nr_sessions)))
+def inverse_session_cross_validation(model_name:str, X, y, session_lengths, nr_sessions, log_to_csv=True, batch_size=64, epochs=30):
+    session_training_results = []
+    for i in range(nr_sessions):
+
+        X_test_session, X_train_session, y_test_session, y_train_session = prepare_datasets_sessions(X, y, session_lengths, i)
+    
+        # Model:
+        if model_name == 'LSTM':
+            model = LSTM(input_shape=(1, 208))
+
+        elif model_name == 'GRU':
+            model = GRU(input_shape=(1, 208))
+
+        elif model_name == 'CNN_1D':
+            X_train_session = np.reshape(X_train_session, (X_train_session.shape[0], 208, 1))
+            X_test_session = np.reshape(X_test_session, (X_test_session.shape[0], 208, 1))
+            model = CNN_1D(input_shape=(208, 1))
+
+        elif model_name == 'FFN':
+            model = FFN(input_shape=(1, 208))
+
+        else:
+            raise Exception('Model not found')
+
+        train(model, X_train_session, y_train_session, verbose=1, batch_size=batch_size, epochs=epochs)
+        test_loss, test_acc = model.evaluate(X_test_session, y_test_session, verbose=0)
+        session_training_results.append(test_acc)
+        if log_to_csv:
+            custom_path = '/{}_train_session{}_log.csv'
+            prediction_csv_logger(X_test_session, y_test_session, model_name, model, i, custom_path)
+        del model
+        K.clear_session()
+        #print('Session', i, 'as test data gives accuracy:', test_acc)
+
+    average_result = statistics.mean((session_training_results))
+
+    return average_result, session_training_results
+
 # Takes in test data and logs input data and the prediction from a model
 # Input: raw data, session_lengths list, total nr of sessions, batch_size, and nr of epochs 
 # Ouput: tuple(cross validation average, list(result for each dataset(len=nr_sessions)))
-def prediction_csv_logger(X, y, model_name, model, session_nr):
+def prediction_csv_logger(X, y, model_name, model, session_nr, custom_path=None):
     
     csv_path = str(Path.cwd()) + '/logs/{}/{}_session{}_log.csv'.format(model_name, model_name, session_nr+1)
+    if custom_path:
+        path = str(Path.cwd()) + '/logs/{}' + custom_path
+        csv_path = path.format(model_name, model_name, session_nr+1)
 
     layerOutput = model.predict(X, verbose=0)
 
@@ -377,12 +602,12 @@ if __name__ == "__main__":
     NR_SUBJECTS = 5
     NR_SESSIONS = 4
     BATCH_SIZE = 64
-    EPOCHS = 5
+    EPOCHS = 10
 
     TEST_SESSION_NR = 4
     VERBOSE = 1
     MODEL_NAME = 'CNN_1D'
-    LOG = True
+    LOG = False
     
     # ----- Get prepared data: train, validation, and test ------
         # X_train.shape = (2806-X_test, 1, 208)
@@ -430,8 +655,9 @@ if __name__ == "__main__":
     '''
 
 
-    #'''
+    '''
     # ----- Cross validation ------
+    # Trained on three sessions, tested on one
     average_GRU = session_cross_validation('GRU', X, y, session_lengths, nr_sessions=NR_SESSIONS, 
                                                                         log_to_csv=LOG,
                                                                         batch_size=BATCH_SIZE, 
@@ -450,10 +676,44 @@ if __name__ == "__main__":
                                                                         epochs=EPOCHS)
 
     print('\n')
-    print('Crossvalidated GRU:', average_GRU)
-    print('Crossvalidated LSTM:', average_LSTM)
-    print('Crossvalidated FFN:', average_FFN)
+    print('Cross-validated GRU:', average_GRU)
+    print('Cross-validated LSTM:', average_LSTM)
+    print('Cross-validated FFN:', average_FFN)
     print('Cross-validated CNN_1D:', average_CNN)
     print('\n')
-    #'''
+    '''
+
+    '''
+    # ----- Inverse cross-validation ------
+    # Trained on one session, tested on three
+    average_GRU = inverse_session_cross_validation('GRU', X, y, session_lengths, nr_sessions=NR_SESSIONS,
+                                                                        log_to_csv=LOG, 
+                                                                        batch_size=BATCH_SIZE, 
+                                                                        epochs=EPOCHS)
+    average_LSTM = inverse_session_cross_validation('LSTM', X, y, session_lengths, nr_sessions=NR_SESSIONS,
+                                                                        log_to_csv=LOG,
+                                                                        batch_size=BATCH_SIZE, 
+                                                                        epochs=EPOCHS)
+    average_FFN = inverse_session_cross_validation('FFN', X, y, session_lengths, nr_sessions=NR_SESSIONS,
+                                                                        log_to_csv=LOG,
+                                                                        batch_size=BATCH_SIZE, 
+                                                                        epochs=EPOCHS)
+    average_CNN = inverse_session_cross_validation('CNN_1D', X, y, session_lengths, nr_sessions=NR_SESSIONS,
+                                                                        log_to_csv=LOG,
+                                                                        batch_size=BATCH_SIZE, 
+                                                                        epochs=EPOCHS)
+
+    print('\n')
+    print('Cross-validated one-session-train GRU:', average_GRU)
+    print('Cross-validated one-session-train LSTM:', average_LSTM)
+    print('Cross-validated one-session-train FFN:', average_FFN)
+    print('Cross-validated one-session-train CNN_1D:', average_CNN)
+    print('\n')
+    '''
+
+    # ----- PLOTTING ------
+
+    #plot_4xinverse_cross_val(X, y, session_lengths, NR_SESSIONS, epochs=30)
+    plot_4_x_average_val(X, y, session_lengths, NR_SESSIONS, epochs=30)
+
 

Present_data.py

@@ -125,8 +125,9 @@ def pretty(dict):
 
 # DATA FUNCTIONS: --------------------------------------------------------------: 
 
-# The CSV_handler takes in data_type(soft, hard, softPP, hardPP)
-# E.g. handler = CSV_handler('soft')
+# The CSV_handler takes in nr of subjects and nr of sessions in the experiment
+# E.g. handler = CSV_handler(nr_subjects=5, nr_sessions=4)
+# Needs to load data: handler.load_data(<type>, <type_directory_name>)
     
 # Denoices one set of EMG data 
 # Input: CSV_handler and detailes for ID 
@@ -238,15 +239,25 @@ def mfcc_all_emg_plots(csv_handler:CSV_handler):
 
     plot_all_emg_mfcc(feat_list, label_list)
 
+
 # MAIN: ------------------------------------------------------------------------: 
 
 if __name__ == "__main__":
 
-    csv_handler = CSV_handler()
-    csv_handler.load_data('soft')
+    NR_SUBJECTS = 5
+    NR_SESSIONS = 4
+
+    soft_dir_name = 'Exp20201205_2myo_softType'
+    hard_dir_name = 'Exp20201205_2myo_hardType'
+    JSON_TEST_NAME = 'TEST_mfcc.json'
+
+    csv_handler = CSV_handler(NR_SUBJECTS, NR_SESSIONS)
+    dict = csv_handler.load_data('soft', soft_dir_name)
+    
     nn_handler = NN_handler(csv_handler)
     nn_handler.store_mfcc_samples()
-    nn_handler.save_json_mfcc()
+    nn_handler.save_json_mfcc(JSON_TEST_NAME)
+
 
 
 

README.md

@@ -11,12 +11,10 @@ Scripts to handle CSV files composed by 2 * 8 EMG sensors(left & right) devided
 * Community libs: Python_speech_features, Pywt
 
 #### Challanges in the module
-* The CSV handlig is for the moment hard-coded to fit the current project due to a very specific file structure and respective naming convention.
+* The CSV handlig requires a specific file structure. Se "How to use it"
 * Preprocessing is still limited in Signal_prep.py
-* Neural_Network_Analysis.py lacks a more general way to access multiple types of networks
 
 #### Credits for insporational code
-
 * Kapre: Keunwoochoi
 * Audio-Classification: seth814
 * DeepLearningForAudioWithPyhton: musikalkemist
@@ -35,7 +33,8 @@ Scripts to handle CSV files composed by 2 * 8 EMG sensors(left & right) devided
 
 1. Clone the repo
 2. Place the data files in the working directory 
-3. (For now) Add the session filenames in the desired load_data() function 
+3. Place the data files within the `data`-folder 
+(format: /`data`/<datatype>/<subject-folder+ID>/<session-folder>/<left/right-CSV-files>)
 4. Assuming NN analysis:
     1. Create a `CSV_handler` object 
     2. Load data with `load_data(CSV_handler, <datatype>)`