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
| import sys
import subprocess
import os
from PIL import Image
from datetime import datetime
from ASR_metrics import utils as metrics
from werkzeug.wrappers import Request, Response
from flask import Flask, render_template, request, jsonify
sys.path.append('/home/nvidia/7th_CV')
sys.path.append('/home/nvidia/7th_ASR')
pathASR = "/home/nvidia/7th_ASR"
pathSky = '/home/nvidia'
app = Flask(__name__, static_folder='')
uploadPath = 'uploads/'
try_model_1 = None
@app.route('/')
def index():
return render_template('sky7.html', template_folder='templates')
@app.route('/asr/load')
def asrLoad():
global try_model_1
if try_model_1 == None:
import nemo.collections.asr as nemo_asr
print('Loading Nemo')
try_model_1 = nemo_asr.models.EncDecCTCModel.restore_from("/home/nvidia/7th_ASR/7th_asr_model.nemo")
print('Done loading Nemo')
return 'ok'
@app.route('/asr/upload', methods=['POST'])
def asrUpload():
if request.method == 'POST':
f = request.files['file']
if(f.headers.get('Content-Type') != 'audio/wav'):
return '音频格式有错误', 400
else:
fileName = f'{uploadPath}audio.wav'
f.save(fileName)
dt = datetime.now()
ts = str(int(datetime.timestamp(dt)))
return jsonify(f'/{uploadPath}audio.wav?t={ts}')
@app.route('/asr/identify', methods=['GET', 'POST'])
def asrIdentify():
global try_model_1
if try_model_1 == None:
return '模型无效，请重新加载', 500
try:
asr_result = try_model_1.transcribe(paths2audio_files=["uploads/audio.wav"])
s1 = request.form.get('defaultText')
s2 = " ".join(asr_result)
result = {
"asr_result": asr_result,
"word_error_rate": metrics.calculate_cer(s1,s2),
"word_accuracy_rate":1-metrics.calculate_cer(s1,s2)
}
return jsonify(result)
except Exception as e:
return '无法识别', 400
@app.route("/cv/upload", methods=['POST'])
def cvUpload():
if request.method == 'POST':
f = request.files['file']
print('image', f, f.filename)
if not 'image' in f.headers.get('Content-Type'):
return '图片有错误', 400
original = f'{uploadPath}original.jpg'
try:
im = Image.open(f)
rgb_im = im.convert('RGB')
rgb_im.save(original)
dt = datetime.now()
ts = str(int(datetime.timestamp(dt)))
return jsonify(original+'?t='+ts)
except Exception as e:
return '有错误', 400
@app.route("/api/detect/image")
def detectImage():
cv_results = subprocess.Popen('python3 /home/nvidia/7th_CV/detection_image.py /home/nvidia/uploads/original.jpg', shell = True, stdout = subprocess.PIPE, stderr = subprocess.PIPE)
print('code', cv_results.returncode)
cv_results = str(cv_results.stdout.read()).split('\\n')[-2]
dt = datetime.now()
ts = str(int(datetime.timestamp(dt)))
result = {
"detection_result_image_path": f'/uploads/result.jpg?t={ts}'
}
return jsonify(result)
@app.route("/api/detect/fps")
def detectFPS():
fps_results = subprocess.Popen('python3 /home/nvidia/7th_CV/cv_fps.py', shell = True, stdout = subprocess.PIPE, stderr = subprocess.STDOUT)
fps_results = str(fps_results.stdout.read()).split('\\n')[-2]
fps_results = fps_results.split(" ")[-1]
result = {
"detection_FPS": fps_results,
}
return jsonify(result)
@app.route("/api/detect/map")
def detectMAP():
map_results = subprocess.Popen('python3 /home/nvidia/7th_CV/cv_map.py', shell = True, stdout = subprocess.PIPE, stderr = subprocess.STDOUT)
map_results = str(map_results.stdout.read())
bytes(map_results, encoding="utf-8").decode()
map_results = map_results[-9:-3]
result = {
"detection_mAP": map_results,
}
return jsonify(result)
if __name__ == "__main__":
app.run(debug=True)
|
