import psycopg2
connection = psycopg2.connect("host=localhost port=5432 dbname=postgres user=kimishi password=********")
connection.get_backend_pid()

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import math
from matplotlib.patches import Ellipse
from scipy import stats

df_pre = pd.read_sql(sql="SELECT t1.*,t2.artist_genres FROM tracks as t1 left join artists as t2 on t1.artist_id=t2.artist_id where valence>=0.5 and energy>=0.5;", con=connection, index_col='index' )
df = df_pre[df_pre['artist_genres'].str.contains('pop')]

def corrfunc(x, y, **kws):
    r, _ = stats.spearmanr(x, y)
    ax = plt.gca()
    ax.axis('off')
    ellcolor = plt.cm.RdBu(0.5*(r+1))
    txtcolor = 'black' if math.fabs(r) < 0.5 else 'white'
    ax.add_artist(Ellipse(xy=[.5, .5], width=math.sqrt(1+r), height=math.sqrt(1-r), angle=45,
        facecolor=ellcolor, edgecolor='none', transform=ax.transAxes))
    ax.text(.5, .5, '{:.0f}'.format(r*100), color=txtcolor, fontsize=28,
        horizontalalignment='center', verticalalignment='center', transform=ax.transAxes)

d2 = pd.DataFrame({'valence':df['valence'], 'energy':df['energy'], 'acousticness':df['acousticness'], 'danceability':df['danceability'],
                  'instrumentalness':df['instrumentalness'], 'liveness':df['liveness'], 'loudness':df['loudness'] , 'speechiness':df['speechiness'] , 'tempo':df['tempo'] , 'popularity':df['popularity']},
                  columns=['valence', 'energy', 'acousticness', 'danceability', 'instrumentalness', 'liveness', 'loudness','speechiness','tempo','popularity'])

sns.set(font_scale=2)
g = sns.PairGrid(d2)
g = g.map_lower(sns.kdeplot, cmap='Blues_d')
g = g.map_diag(sns.distplot, kde=False)
g = g.map_upper(corrfunc)
g.fig.subplots_adjust(wspace=0.05, hspace=0.05)
for ax in g.axes.flatten():
    for t in ax.get_xticklabels():
        _ = t.set(rotation=40)
g.savefig('/Users/kimishi/Desktop/fig1_1_1.png')


df_pre = pd.read_sql(sql="SELECT t1.*,t2.artist_genres FROM tracks as t1 left join artists as t2 on t1.artist_id=t2.artist_id where valence<0.5 and energy>=0.5;", con=connection, index_col='index' )
df = df_pre[df_pre['artist_genres'].str.contains('pop')]

def corrfunc(x, y, **kws):
    r, _ = stats.spearmanr(x, y)
    ax = plt.gca()
    ax.axis('off')
    ellcolor = plt.cm.RdBu(0.5*(r+1))
    txtcolor = 'black' if math.fabs(r) < 0.5 else 'white'
    ax.add_artist(Ellipse(xy=[.5, .5], width=math.sqrt(1+r), height=math.sqrt(1-r), angle=45,
        facecolor=ellcolor, edgecolor='none', transform=ax.transAxes))
    ax.text(.5, .5, '{:.0f}'.format(r*100), color=txtcolor, fontsize=28,
        horizontalalignment='center', verticalalignment='center', transform=ax.transAxes)

d3 = pd.DataFrame({'valence':df['valence'], 'energy':df['energy'], 'acousticness':df['acousticness'], 'danceability':df['danceability'],
                  'instrumentalness':df['instrumentalness'], 'liveness':df['liveness'], 'loudness':df['loudness'] , 'speechiness':df['speechiness'] , 'tempo':df['tempo'] , 'popularity':df['popularity']},
                  columns=['valence', 'energy', 'acousticness', 'danceability', 'instrumentalness', 'liveness', 'loudness','speechiness','tempo','popularity'])

sns.set(font_scale=2)
g = sns.PairGrid(d3)
g = g.map_lower(sns.kdeplot, cmap='Blues_d')
g = g.map_diag(sns.distplot, kde=False)
g = g.map_upper(corrfunc)
g.fig.subplots_adjust(wspace=0.05, hspace=0.05)
for ax in g.axes.flatten():
    for t in ax.get_xticklabels():
        _ = t.set(rotation=40)
g.savefig('/Users/kimishi/Desktop/fig1_1_2.png')


df_pre = pd.read_sql(sql="SELECT t1.*,t2.artist_genres FROM tracks as t1 left join artists as t2 on t1.artist_id=t2.artist_id where valence<0.5 and energy<0.5;", con=connection, index_col='index' )
df = df_pre[df_pre['artist_genres'].str.contains('pop')]

def corrfunc(x, y, **kws):
    r, _ = stats.spearmanr(x, y)
    ax = plt.gca()
    ax.axis('off')
    ellcolor = plt.cm.RdBu(0.5*(r+1))
    txtcolor = 'black' if math.fabs(r) < 0.5 else 'white'
    ax.add_artist(Ellipse(xy=[.5, .5], width=math.sqrt(1+r), height=math.sqrt(1-r), angle=45,
        facecolor=ellcolor, edgecolor='none', transform=ax.transAxes))
    ax.text(.5, .5, '{:.0f}'.format(r*100), color=txtcolor, fontsize=28,
        horizontalalignment='center', verticalalignment='center', transform=ax.transAxes)

d4 = pd.DataFrame({'valence':df['valence'], 'energy':df['energy'], 'acousticness':df['acousticness'], 'danceability':df['danceability'],
                  'instrumentalness':df['instrumentalness'], 'liveness':df['liveness'], 'loudness':df['loudness'] , 'speechiness':df['speechiness'] , 'tempo':df['tempo'] , 'popularity':df['popularity']},
                  columns=['valence', 'energy', 'acousticness', 'danceability', 'instrumentalness', 'liveness', 'loudness','speechiness','tempo','popularity'])

sns.set(font_scale=2)
g = sns.PairGrid(d4)
g = g.map_lower(sns.kdeplot, cmap='Blues_d')
g = g.map_diag(sns.distplot, kde=False)
g = g.map_upper(corrfunc)
g.fig.subplots_adjust(wspace=0.05, hspace=0.05)
for ax in g.axes.flatten():
    for t in ax.get_xticklabels():
        _ = t.set(rotation=40)
g.savefig('/Users/kimishi/Desktop/fig1_1_3.png')


df_pre = pd.read_sql(sql="SELECT t1.*,t2.artist_genres FROM tracks as t1 left join artists as t2 on t1.artist_id=t2.artist_id where valence>=0.5 and energy<0.5;", con=connection, index_col='index' )
df = df_pre[df_pre['artist_genres'].str.contains('pop')]

def corrfunc(x, y, **kws):
    r, _ = stats.spearmanr(x, y)
    ax = plt.gca()
    ax.axis('off')
    ellcolor = plt.cm.RdBu(0.5*(r+1))
    txtcolor = 'black' if math.fabs(r) < 0.5 else 'white'
    ax.add_artist(Ellipse(xy=[.5, .5], width=math.sqrt(1+r), height=math.sqrt(1-r), angle=45,
        facecolor=ellcolor, edgecolor='none', transform=ax.transAxes))
    ax.text(.5, .5, '{:.0f}'.format(r*100), color=txtcolor, fontsize=28,
        horizontalalignment='center', verticalalignment='center', transform=ax.transAxes)

d5 = pd.DataFrame({'valence':df['valence'], 'energy':df['energy'], 'acousticness':df['acousticness'], 'danceability':df['danceability'],
                  'instrumentalness':df['instrumentalness'], 'liveness':df['liveness'], 'loudness':df['loudness'] , 'speechiness':df['speechiness'] , 'tempo':df['tempo'] , 'popularity':df['popularity']},
                  columns=['valence', 'energy', 'acousticness', 'danceability', 'instrumentalness', 'liveness', 'loudness','speechiness','tempo','popularity'])

sns.set(font_scale=2)
g = sns.PairGrid(d5)
g = g.map_lower(sns.kdeplot, cmap='Blues_d')
g = g.map_diag(sns.distplot, kde=False)
g = g.map_upper(corrfunc)
g.fig.subplots_adjust(wspace=0.05, hspace=0.05)
for ax in g.axes.flatten():
    for t in ax.get_xticklabels():
        _ = t.set(rotation=40)
g.savefig('/Users/kimishi/Desktop/fig1_1_4.png')