-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathbulk_run.py
185 lines (145 loc) · 7.75 KB
/
bulk_run.py
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
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
from mongo_connection import Mongo_connection
import numpy as np
import pandas as pd
import pair_transition_analysis
import granger_causation_test
from matplotlib import pyplot as plt
from collections import defaultdict
import roi_config
import fixation
import hypothesis_testing
mongo = Mongo_connection()
mongo.connect()
def get_basic_metrics(df_fixation):
d = {}
d["fixation_mean_duration"] = df_fixation["duration"].mean()
d["fixation_rate"] = len(df_fixation)/df_fixation["end"].values[-1]*1000/60
d["saccade_amplitude"] = np.sqrt(np.diff(df_fixation["x"])**2 + np.diff(df_fixation["y"])**2).mean()
d["saccade_mean_duration"] = np.mean(df_fixation["start"].values[1:] - df_fixation["end"].values[:-1])
return d
def create_transition_matrix(transitions):
m = pd.crosstab(pd.Series(list(transitions)[1:], name = "t+1"),
pd.Series(list(transitions)[:-1], name = "t"),normalize=1)
return m
def calculate_entropy(transitions):
if len(transitions) == 0:
return 0, 0
# transitions = replace_repeated_character(transitions)
trans_matrix = create_transition_matrix(transitions)
m = {}
for c in trans_matrix.columns:
m[c] = trans_matrix[c].tolist()
Hs = 0
Ht = 0
pA = {c:len(np.where(np.array(list(transitions))==c)[0])/len(transitions) for c in list(set(transitions))}
for k,v in pA.items():
Hs += -1 * np.nan_to_num(v*np.log2(v))
Ht += -sum(pA[k]*(np.nan_to_num(m[k]*np.log2(m[k]))))
return Hs, Ht
def get_advanced_metrics(df_fixation):
d = {}
transitions, L = pair_transition_analysis.encode_transition(df_fixation["roi"])
Hs, Ht = calculate_entropy(transitions)
d["Hs"] = Hs
d["Ht"] = Ht
return d
def get_dwell_stat(df_data):
agg_sum = df_data.groupby(["roi"]).agg({'duration': 'sum'})
agg_sum_percent = agg_sum/sum(agg_sum['duration'])
agg_mean = df_data.groupby(["roi"]).agg({'duration': 'mean'})
agg_var = df_data.groupby(["roi"]).agg({'duration': 'var'})
agg_fixrate = df_data.groupby(["roi"]).agg({'duration': 'count'})/df_data.iloc[-1]["end"]*1000/60
list_roi = list(set(df_data["roi"]))
d = {}
for roi in list_roi:
d["duration_{}".format(roi)] = agg_sum.loc[roi][0]
d["duration_percentage_{}".format(roi)] = agg_sum_percent.loc[roi][0]
d["duration_average_{}".format(roi)] = agg_mean.loc[roi][0]
d["duration_var_{}".format(roi)] = agg_var.loc[roi][0]
d["fix_rate_{}".format(roi)] = agg_fixrate.loc[roi][0]
# df_data = fixation.merge_consecutive_fixations_in_same_roi(df_data)
# df_runway = df_data[df_data["roi"]=="runway"]
# duration_in_between_runway = df_runway["start"].values[1:] - df_runway["end"].values[:-1]
# n_transition_in_between_runway = np.diff(df_runway.index) - 1
# d["between_runway_mean_duration_all"] = np.mean(duration_in_between_runway/n_transition_in_between_runway)
# n_trans = list(set(n_transition_in_between_runway))
# for n in range(1,4):
# d["between_runway_n_trans_{}".format(n)] = sum(n_transition_in_between_runway==n)
# d["between_runway_mean_duration_{}".format(n)] = np.mean(duration_in_between_runway[np.where(n_transition_in_between_runway == n)])
return d
def run_dwell_stats(cmd):
documents = mongo.find(cmd)
d = defaultdict(list)
for document in documents:
print("trial: {}, group: {}, pID: {}".format(document["trial"], document["group"], document["pID"]))
if document["trial"] == 4:
continue
d['pID'].append(document["pID"])
d['group'].append(document["group"])
d['trial'].append(document["trial"])
d['rating'].append(document["rating"])
d["null_percent"].append(document["null_percent"])
# d["calibration"].append(document["calibration"])
d_data = document["data"]
df_data = pd.DataFrame(d_data)
d_dwell = get_dwell_stat(df_data)
for k in roi_config.encode_table.keys():
d["duration_{}".format(k)].append(d_dwell.get("duration_{}".format(k), 0))
d["duration_percentage_{}".format(k)].append(d_dwell.get("duration_percentage_{}".format(k), 0))
d["duration_average_{}".format(k)].append(d_dwell.get("duration_average_{}".format(k), 0))
d["fix_rate_{}".format(k)].append(d_dwell.get("fix_rate_{}".format(k), 0))
# for k in d_dwell.keys():
# if "between_runway" in str(k):
# d[k].append(d_dwell.get(k, 0))
df_dwell = pd.DataFrame(d).sort_values(["pID","trial", "group","rating"]).dropna().reset_index().drop(columns=["index"])
return df_dwell
def run_basic_metrics(cmd):
documents = mongo.find(cmd)
d = defaultdict(list)
for document in documents:
print("trial: {}, group: {}, pID: {}".format(document["trial"], document["group"], document["pID"]))
if document["trial"] == 4:
continue
d['pID'].append(document["pID"])
d['group'].append(document["group"])
d['trial'].append(document["trial"])
d["null_percent"].append(document["null_percent"])
# d["calibration"].append(document["calibration"])
d_data = document["data"]
df_data = pd.DataFrame(d_data)
transitions, L = pair_transition_analysis.encode_transition(df_data["roi"])
basic_metrics = get_basic_metrics(df_data)
advance_metrics = get_advanced_metrics(df_data)
d["fixation_mean_duration"].append(basic_metrics["fixation_mean_duration"])
d["fixation_rate"].append(basic_metrics["fixation_rate"])
d["saccade_amplitude"].append(basic_metrics["saccade_amplitude"])
d["saccade_mean_duration"].append(basic_metrics["saccade_mean_duration"])
# d["Hs"].append(advance_metrics["Hs"])
# d["Ht"].append(advance_metrics["Ht"])
# for ngram_length in range(3,7):
# subseqcount = defaultdict(dict)
# for i in range(len(transitions)-ngram_length + 1):
# substring = transitions[i:i+ngram_length]
# if subseqcount[substring].get("count"):
# subseqcount[substring]["count"] += 1
# subseqcount[substring]["duration"] += df_data.iloc[i:i+ngram_length]["duration"].sum()
# else:
# subseqcount[substring]["count"] = 1
# subseqcount[substring]["duration"] = df_data.iloc[i:i+ngram_length]["duration"].sum()
# sorted_subseqcount = {k: v for k, v in sorted(subseqcount.items(), key=lambda item: item[1]["count"], reverse=True)}
# more_than_2_time_seq = {k: v for k, v in subseqcount.items() if v["count"] >= 2}
# count_of_most_seq = 0
# N_unique_seq = 0
# print(sorted_subseqcount)
# if len(subseqcount.values())>0:
# count_of_most_seq = max(subseqcount.values())/sum(subseqcount.values())
# N_unique_seq = len(subseqcount.keys())
# N_unique_seq_more_than_2_times = 0
# if len(more_than_2_time_seq.values())>0:
# N_unique_seq_more_than_2_times = len(more_than_2_time_seq.keys())
# d["N_unique_seq_{}".format(ngram_length)].append(N_unique_seq)
# d["count_of_most_seq_{}".format(ngram_length)].append(count_of_most_seq)
# d["N_unique_seq_more_than_2_times_{}".format(ngram_length)].append(N_unique_seq_more_than_2_times)
# d["mean_repetition_{}".format(ngram_length)].append(np.mean(list(more_than_2_time_seq.values())))
df_res = pd.DataFrame(d).sort_values(["pID","trial", "group","rating"]).dropna().reset_index().drop(columns=["index"])
return df_res