-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathprogram2.py
203 lines (169 loc) · 8.35 KB
/
program2.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
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
import subprocess
import pandas as pd
import os
import json
from pathlib import Path
import numpy as np
import math
diagonal = 735
in_col = ["participant_id", "behavior_1", "behavior_2", "behavior_3", "behavior_4"]
fin_col = ['participant_id', 'behavior']
pose = ['gaze_angle_x', 'gaze_angle_y', 'pose_Rx', 'pose_Ry', 'pose_Rz', 'p_rx', 'p_ry', 'p_rz']
au = ['AU01_r', 'AU02_r', 'AU04_r', 'AU05_r', 'AU06_r', 'AU07_r', 'AU09_r', 'AU10_r', 'AU12_r', 'AU14_r', 'AU15_r', 'AU17_r', 'AU20_r', 'AU23_r', 'AU25_r', 'AU26_r', 'AU45_r']
pose_keypoints = ['Neck', 'RShoulder', 'RElbow', 'RWrist',
'LShoulder', 'LElbow', 'LWrist',
'MidHip', 'RHip', 'RKnee', 'RAnkle',
'LHip', 'LKnee', 'LAnkle',
'REye', 'LEye', 'REar', 'LEar',
'LBigToe', 'LSmallToe', 'LHeel',
'RBigToe', 'RSmallToe', 'RHeel']
hand_keypoints = ['Wrist_L', 'Thumb1_L', 'Thumb2_L', 'Thumb3_L', 'Thumb4_L',
'Index1_L', 'Index2_L', 'Index3_L', 'Index4_L',
'Middle1_L', 'Middle2_L', 'Middle3_L', 'Middle4_L',
'Ring1_L', 'Ring2_L', 'Ring3_L', 'Ring4_L',
'Pinky1_L', 'Pinky2_L', 'Pinky3_L', 'Pinky4_L',
'Wrist_R', 'Thumb1_R', 'Thumb2_R', 'Thumb3_R', 'Thumb4_R',
'Index1_R', 'Index2_R', 'Index3_R', 'Index4_R',
'Middle1_R', 'Middle2_R', 'Middle3_R', 'Middle4_R',
'Ring1_R', 'Ring2_R', 'Ring3_R', 'Ring4_R',
'Pinky1_R', 'Pinky2_R', 'Pinky3_R', 'Pinky4_R']
def save_csv(folder, filename, df):
"""
Salva un DataFrame come file CSV nella cartella specificata (se non esiste, la crea).
Parameters
----------
folder : str
Il percorso della cartella in cui salvare il file CSV.
filename : str
Il nome con cui salvare il DataFrame come file CSV.
df : pandas.DataFrame
Il DataFrame da salvare come file CSV.
Returns
-------
None.
"""
if not os.path.isdir(folder):
os.mkdir(folder)
df.to_csv(os.path.join(folder + '/' + filename), index=False)
def concat_csv(folder, filename, df_list):
"""
Concatena tutti i DataFrame contenuti nella lista e li salva in un file CSV.
Parameters
----------
folder : str
Il percorso della cartella in cui salvare il file CSV.
filename : str
Il nome con cui salvare il DataFrame come file CSV.
df_list : list
Lista di DataFrame da concatenare in un unico DataFrame.
Returns
-------
None.
"""
df_completo = pd.concat(df_list)
df_completo.to_csv(folder + '/' + filename, index=False)
def csv_intersec():
#subprocess.run(["python", "ELANtoCSV.py"])
pd.options.display.max_rows = None
pd.options.display.max_columns = None
folder1 = "csv_from_elan"
folder2 = "csv"
folder3 = "json"
dataframes = {}
lista_dataframe = []
files1 = [f for f in os.listdir(folder1) if f.endswith('.csv')] #elenco dei file CSV nella cartella csv_from_elan
for file in files1:
file_path = os.path.join(folder1, file)
df = pd.read_csv(file_path) #carica ciascun file CSV in un DataFrame
dataframes[file] = df #salva il DataFrame nel dizionario con il nome del file come chiave (es 9.csv)
files2 = [f for f in os.listdir(folder2) if f.endswith('.csv')] #elenco dei file CSV nella cartella csv
for file in files2:
new_filename = file[3:]
number = new_filename.split(".")[0]
file_path2 = os.path.join(folder2, file)
keepCol = pose + au
df2 = pd.read_csv(file_path2, usecols=keepCol) #carica ciascun file CSV in un DataFrame
list_df2 = df2.values.tolist() #trasforma il DataFrame in una lista di liste
df_col = fin_col + pose + au + pose_keypoints + hand_keypoints
bn1 = np.zeros((len(df2), len(in_col)+1))
bn1[:, 0] = number
bn2 = np.zeros((len(df2), (len(pose_keypoints) + len(hand_keypoints))))
temp = []
temprh = []
templh = []
nose = []
num = '0' + number if len(number) == 1 else number
for entry in Path(folder3).glob(f'**/*{num}'):
if not entry.is_dir():
continue
subfolder_path = entry.resolve()
for file_entry in subfolder_path.glob('*.json'):
with file_entry.open() as f:
data = json.load(f)
people = data.get('people', [])
if not people:
continue
frame = int(file_entry.stem.split('_')[1].lstrip('0') or 0) #riga
el_p = people[0]['pose_keypoints_2d']
el_lh = people[0]['hand_left_keypoints_2d']
el_rh = people[0]['hand_right_keypoints_2d']
nose = el_p[0:2]
temp = [el_p[i:i+2] for i in range(3, len(el_p), 3)]
temprh = [el_rh[i:i+2] for i in range(0, len(el_rh), 3)]
templh = [el_lh[i:i+2] for i in range(0, len(el_lh), 3)]
hands = temprh + templh
dist = []
if nose[0] != 0 and nose[1] != 0:
for l in temp:
#if l[0] or l[1] < 1: le metto a 0 senza dist euclidea
#elif l[0] and l[1] != 0: distanza euclidea
if l[0] < 1 or l[1] < 1:
dist.append(0)
else:
#il naso ha coordinate (xn, yn) e un punto sulla faccia (xp, yp)
#d((xn,yn), (xp,yp)) = sqrt((xp - xn)^2 + (yp - yn)^2)
dist.append(math.sqrt((l[0] - nose[0])**2 + (l[1] - nose[1])**2))
for l in hands:
if l[0] < 1 or l[1] < 1:
dist.append(diagonal)
else:
dist.append(math.sqrt((l[0] - nose[0])**2 + (l[1] - nose[1])**2))
else:
dist.append(0)
#bn2
bn2[frame] = dist
if new_filename in dataframes:
df1 = dataframes[new_filename] #prende il DataFrame corrispondente nel dizionario utilizzando il nome del file
for index, row in df1.iterrows():
start = row['inizio'] - 1
stop = row['fine'] - 1
action = row['classe']
#popolo sempre le 4 colonne di behavior con 1
bn1[start:stop, action] = 1
#non serve gestire la sovrapposizione in alcun modo perchè comunque
#eliminiamo ogni riga che ha più di un 1
bn1[start:stop, 5] = action
#elimino le righe con comportamenti sovrapposti o senza comportamenti
index_to_delete = []
for index, row in enumerate(bn1[:, 1:5]):
if sum(row) > 1 or sum(row) == 0:
index_to_delete.append(index)
index_to_delete.reverse() #faccio il reverse degli indici per cancellare dalla fine
bn3 = np.dstack((bn1[:, 0], bn1[:, -1]))
bn4 = bn3[0]
concat_list = np.concatenate((bn4, list_df2), axis=1).tolist() #concatena i numpy array e la lista di liste
concat_list2 = np.concatenate((concat_list, bn2), axis=1).tolist()
for x in index_to_delete:
del concat_list2[x]
new_df = pd.DataFrame(concat_list2, columns=(tuple(df_col)))
new_df[fin_col] = new_df[fin_col].astype(int)
new_df[pose] = new_df[pose].apply(lambda x: x * 1000).applymap('{:.0f}'.format)
new_df[au] = new_df[au].applymap('{:.2f}'.format)
lista_dataframe.append(new_df) #aggiunge il Dataframe corrente alla lista dei Dataframe
output_folder = "csv_output"
save_csv(output_folder, new_filename, new_df) #salva il DataFrame come csv
name = 'final.csv'
concat_csv(output_folder, name, lista_dataframe) #concatena tutti i DataFrame e li salva come csv
print("Programma terminato correttamente")
if __name__ == '__main__':
csv_intersec()