phaeaco-analysis/analyze at master · fargonauts/phaeaco-analysis · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
#!/usr/bin/env python3

import sys
import numpy as np
import pandas  as ps
import seaborn as sns
import matplotlib.pyplot as plt

from read import read

from collections import OrderedDict
from pprint import pprint
from copy import deepcopy

def subject_indices(d, subject):
    for i, s in enumerate(d['subjects'].values()):
        if s == subject:
            yield i

def add_percent_stat(d, k):
    pdict = dict()
    for subject in ['human', 'phaeaco']:
        m = max(d[k][i] for i in subject_indices(d, subject))
        for i in subject_indices(d, subject):
            pdict[i] = d[k][i] / max(m, 1)
    d['p' + k] = pdict

def extend(df):
    d = df.to_dict()
    for t in ['ctime', 'itime', 'ntime']:
        add_percent_stat(d, t)

    for a in ['correct', 'incorrect', 'no answer']:
        add_percent_stat(d, a)

    return ps.DataFrame(d)

def subject(df, s):
    return df[(df.subjects == s)]

grid_vars = ['bongard problems', 'pctime', 'pcorrect', 'sdev']

def plot_grid(df):
    # Compare too many things using phaeaco's sparse data
    pal = dict(human="#41dff4", phaeaco="#f4aa42")

    g = sns.pairplot(df, hue='subjects', palette=pal, vars=grid_vars,
                                                      kind='reg', diag_kind='kde')
    for i, ax in enumerate(g.axes.flat): # set every-other axis for testing purposes
        if i < 4:
            ax.set_ylim(0, 200)
        elif i < 12:
            ax.set_ylim(0, 1)

def plot_subject_grids(df):
    for subject in ['human', 'phaeaco']:
        g = sns.pairplot(df, vars=grid_vars, kind='reg', diag_kind='kde')
        for i, ax in enumerate(g.axes.flat): # set every-other axis for testing purposes
            if i < 4:
                ax.set_ylim(0, 200)
            elif i < 12:
                ax.set_ylim(0, 1)
        g.fig.suptitle(subject)

def plot_pcorrect(df):
    # Compare pcorrect in a bar graph
    sns.barplot(x='bongard problems', y='pcorrect', hue='subjects', data=df)

def plot_kde(df):
    # Show shorter problems take less time
    sns.jointplot(x='pcorrect', y='pctime', data=df, kind='kde', xlim=(0,1), ylim=(0,1))
    #sns.jointplot(x='pcorrect', y='pctime', data=df, kind='reg', xlim=(0,1), ylim=(0,1))
    #g.set(ylim=(0, 1), xlim=(0, 1))

def plot_all(df):
    #plt.style.use("dark_background")
    df = extend(df)
    print(df)

    #plot_pcorrect(df)
    #plot_grid(df)
    #plot_kde(df)
    plot_subject_grids(df)

    plt.show()

def main(args):
    data = read(includeNone=False)
    plot_all(data)
    return 0

if __name__ == '__main__':
    sys.exit(main(sys.argv[1:]))