#!/usr/bin/env python3

# -*- coding: utf-8 -*-

"""

Created on Mon Feb 8 16:49:04 2021

@author: umreenimam

"""

import os

import math

import numpy as np

import pandas as pd

from scipy.io import loadmat

import matplotlib.pyplot as plt

import seaborn as sns

from pyeeg import bin_power

from docx import Document

from sklearn.preprocessing import MinMaxScaler

from sklearn.model_selection import train_test_split

from sklearn.metrics import confusion_matrix, accuracy_score

from sklearn.neighbors import KNeighborsClassifier

"""

Step 1: Load the Matlab data files and transpose it

"""

# Setting file names

os.chdir('/Users/umreenimam/documents/masters/masters_classes/cs_5310/week_3/lab_chapter3')

filename = 'Acquisition-15-data.mat'

prefile = 'Pre.mat'

medfile = 'Med.mat'

postfile = 'Post.mat'

# Loading data

data = loadmat(filename)

predata = loadmat(prefile)

meddata = loadmat(medfile)

postdata = loadmat(postfile)

eeg = data['data']

pre = predata['data']

med = meddata['data']

post = postdata['data']

# Transpose data and create dataframes

eeg_transposed = pd.DataFrame(eeg.T)

pre_transposed = pd.DataFrame(pre.T)

med_transposed = pd.DataFrame(med.T)

post_transposed = pd.DataFrame(post.T)

"""

Step 2 through 4: Run EEG calculations to build a data frame with alpha values

"""

# Computing Alpha PSI for 3 dataframes

# Function to compute alpha PSI

def getPSI(data):

return alpha_psi

# Running function for each dataset

prePSI = getPSI(pre_transposed)

medPSI = getPSI(med_transposed)

postPSI = getPSI(post_transposed)

# Brain state labels

pre_label = ["Pre"] * prePSI.shape[0]

med_label = ["Med"] * medPSI.shape[0]

post_label = ["Post"] * postPSI.shape[0]

# Combine dataframes

frames = [prePSI, medPSI, postPSI]

combinedPSI = pd.concat(frames, ignore_index=True)

# Combine label lists

combinedLabels = pre_label + med_label + post_label

"""

Step 5 through 7: Remove any co-linear columns

"""

# Create correlation coefficient matrix

corrMat = combinedPSI.corr()

rows = corrMat.shape[0]

cols = corrMat.shape[1]

# Initialize an empty list to keep track of which columns need to be removed

columns_to_remove = []

for i in range(rows):

for j in range(cols):

if i == j:

# If the row number equals the column number, then exit this for

# loop because we don't want to accidentally remove the same

# column number twice.

break

if corrMat.iloc[i, j] > 0.9 or corrMat.iloc[i, j] < -0.9:

# If any values in the correlation matrix are greater than 0.9

# or less than -0.9, add their index number to the list of columns

# that need to be removed.

columns_to_remove.append(i)

break

# Remove any columns that had a correlation factor that was too high

alpha_df = combinedPSI.drop(columns = columns_to_remove, axis = 1)

# Create a new correlation matrix and print it

alpha_corr_2 = alpha_df.corr()

print(alpha_corr_2)

sns.set_theme(style = "white")

plt.figure(figsize=(6,5))

color_map = sns.diverging_palette(230, 20, as_cmap = True)

sns.heatmap(corrMat, annot = False, cmap = color_map, vmax = 1,

center = 0, square = True, linewidths = 0.1,

cbar_kws = {"shrink": 0.75})

plt.title('Heat Map of Correlation Coefficient Matrix', fontsize = 18)

plt.xlabel('Column Number from the Data Frame', fontsize = 12)

plt.ylabel('Column Number from the Data Frame', fontsize = 12)

plt.savefig('fig1.png')

plt.show

plt.figure(figsize=(6,5))

color_map = sns.diverging_palette(230, 20, as_cmap = True)

sns.heatmap(alpha_corr_2, annot = True, cmap = color_map, vmax = 1,

center = 0, square = True, linewidths = 0.5,

cbar_kws = {"shrink": 0.75})

plt.title('Heat Map of Correlation Coefficient Matrix', fontsize = 18)

plt.xlabel('Column Number from the Data Frame', fontsize = 12)

plt.ylabel('Column Number from the Data Frame', fontsize = 12)

plt.savefig('fig2.png')

plt.show

"""

Step 8 and 9: Normalize the data and split data into train-test data sets

"""

# Normalize data in every remaining channel using the min-max normalization

scaler = MinMaxScaler()

normalized_corr = scaler.fit_transform(alpha_df)

normalized_corr = pd.DataFrame(normalized_corr)

print(normalized_corr)

X = normalized_corr

y = combinedLabels

X_train, X_test, labels_train, labels_test = train_test_split(

X,

y,

random_state = 42,

test_size = 0.2,

stratify = combinedLabels)

"""

Step 10 through 12: Build KNN model and generate confusion matrix

"""

# Train the kNN model using the training dataset

n_neighbors = round(math.sqrt(X_train.shape[1]))

classifier = KNeighborsClassifier(n_neighbors = n_neighbors)

classifier.fit(X_train, labels_train)

# Use the kNN model to predict the target feature of the testing dataset

prediction = classifier.predict(X_test)

pd.crosstab(labels_test, prediction)

# Create a confusion matrix to compare the predicted state activities to

# the actual activities and compute the accuracy

confusion_mat = confusion_matrix(labels_test, prediction)

print(confusion_mat)

# Compute the accuracy

accuracy_rate = accuracy_score(labels_test, prediction)

print(accuracy_rate)

"""

Generate a Word document with the results from this lab

"""

doc = Document()

doc.add_heading('Correlation Coefficient Matrix Before Removing Co-linearity', level = 1)

doc.add_picture('fig1.png')

doc.add_paragraph()

doc.add_page_break()

doc.add_heading('Correlation Coefficient Matrix After Removing Co-linearity', level = 1)

doc.add_picture('fig2.png')

doc.add_paragraph()

doc.add_heading('Confusion Matrix:', level = 1)

table = doc.add_table(rows = confusion_mat.shape[0] + 1, cols = confusion_mat.shape[1] + 1)

table.style = 'Medium Grid 3 Accent 3'

row = table.rows[0]

row.cells[1].text = 'Predicted Med'

row.cells[2].text = 'Predicted Post'

row.cells[3].text = 'Predicted Pre'

col = table.columns[0]

col.cells[1].text = 'Actual Med'

col.cells[2].text = 'Actual Post'

col.cells[3].text = 'Actual Pre'

for i in range(confusion_mat.shape[0]):

for j in range(confusion_mat.shape[1]):

table.cell(i + 1, j + 1).text = str(confusion_mat[i, j])

doc.add_paragraph()

doc.add_paragraph('Accuracy Rate: {}%'.format(round(accuracy_rate * 100, 1)))

doc.save('Chapter3-Lab-Group4.docx')