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
|
import io
import os
import re
import asyncio
import skimage
import skimage.io
import numpy as np
from PIL import Image
from pathlib import Path
from sys import getsizeof
import tensorflow as tf
from ATRI.log import logger as log
from ATRI.utils import request
from ATRI.exceptions import RequestError, WriteFileError
SETU_PATH = Path(".") / "data" / "database" / "setu"
TEMP_PATH = Path(".") / "data" / "temp"
os.makedirs(SETU_PATH, exist_ok=True)
os.makedirs(TEMP_PATH, exist_ok=True)
MODULE_URL = "https://cdn.jsdelivr.net/gh/Kyomotoi/CDN@master/project/ATRI/nsfw.tflite"
VGG_MEAN = [104, 117, 123]
def prepare_image(img):
H, W, _ = img.shape
h, w = (224, 224)
h_off = max((H - h) // 2, 0)
w_off = max((W - w) // 2, 0)
image = img[h_off : h_off + h, w_off : w_off + w, :]
image = image[:, :, ::-1]
image = image.astype(np.float32, copy=False)
image = image * 255.0
image = image - np.array(VGG_MEAN, dtype=np.float32)
image = np.expand_dims(image, axis=0)
return image
async def detect_image(url: str, file_size: int) -> list:
try:
req = await request.get(url)
except RequestError:
raise RequestError("Get info from download image failed!")
img_byte = getsizeof(req.read()) // 1024
if img_byte < file_size:
return [0, 0]
try:
pattern = r"-(.*?)\/"
file_name = re.findall(pattern, url)[0]
path = TEMP_PATH / f"{file_name}.jpg"
with open(path, "wb") as f:
f.write(req.read())
except WriteFileError:
raise WriteFileError("Writing file failed!")
await init_module()
model_path = str((SETU_PATH / "nsfw.tflite").absolute())
try:
interpreter = tf.Interpreter(model_path=model_path) # type: ignore
except Exception:
interpreter = tf.lite.Interpreter(model_path=model_path)
interpreter.allocate_tensors()
input_details = interpreter.get_input_details()
output_details = interpreter.get_output_details()
im = Image.open(path)
if im.mode != "RGB":
im = im.convert("RGB")
imr = im.resize((256, 256), resample=Image.BILINEAR)
fh_im = io.BytesIO()
imr.save(fh_im, format="JPEG")
fh_im.seek(0)
image = skimage.img_as_float32(skimage.io.imread(fh_im))
final = prepare_image(image)
interpreter.set_tensor(input_details[0]["index"], final)
interpreter.invoke()
output_data = interpreter.get_tensor(output_details[0]["index"])
result = np.squeeze(output_data).tolist()
return result
async def init_module():
file_name = "nsfw.tflite"
path = SETU_PATH / file_name
if not path.is_file():
log.warning("缺少模型文件，装载中")
data = await request.get(MODULE_URL)
try:
with open(path, "wb") as w:
w.write(data.read())
log.info("模型装载完成")
except WriteFileError:
raise WriteFileError("NSFW TF module init failed!")
loop = asyncio.get_event_loop()
loop.create_task(init_module())
|
