SoyVitou2019 commited on
Commit
dddba66
·
1 Parent(s): 1b9abc3

7 pushing

Browse files
Files changed (2) hide show
  1. __pycache__/app.cpython-311.pyc +0 -0
  2. app.py +20 -26
__pycache__/app.cpython-311.pyc ADDED
Binary file (3.18 kB). View file
 
app.py CHANGED
@@ -1,32 +1,27 @@
1
  import gradio as gr
2
- from clients import send_request
3
- import json
4
- import ast
5
  def save_non_abuse_class(top_classes):
6
- top_class_json = json.loads(json.dumps(top_classes['message']))
7
- top_class_json = ast.literal_eval(top_class_json)
8
  dic_predict = {
9
  "Non Abuse Image": 0,
10
  "Abuse Image": 0
11
  }
12
- temp_predict = []
13
- Non_abuse_doc = ["Nice emoji", "Non Abuse"]
14
- temp_predict.append(list(top_class_json.keys())[0])
15
  for k, v in top_class_json.items():
16
- if k in Non_abuse_doc:
17
  dic_predict["Non Abuse Image"] += v
18
  else:
19
  dic_predict["Abuse Image"] += v
 
20
  dic_predict = sorted(dic_predict.items(), key=lambda x: x[1], reverse=True)
21
  dic_predict = {k: v for k, v in dic_predict}
22
  return dic_predict
23
-
24
-
25
-
26
-
27
  def prediction(img):
28
- top_classes = send_request(img_input=img, url='http://18.139.116.122/')
29
- print(top_classes)
30
  save_class = save_non_abuse_class(top_classes)
31
  prediction_str = "\n".join([f"{index+1}. {property} : {round(value*100,2)}%" for index, (property, value) in enumerate(save_class.items())])
32
  return prediction_str
@@ -34,17 +29,16 @@ def prediction(img):
34
  def clear_output():
35
  image_input.value = None
36
  prediction_output.value = None
37
- return None, None
38
 
39
- with gr.Blocks(css="footer{display:none !important}") as demo:
40
- with gr.Row():
41
- image_input = gr.inputs.Image()
42
- prediction_output = gr.Textbox(placeholder="result", label="Prediction")
43
- image_button = gr.Button("Make prediction")
44
- clear_button = gr.Button("Clear")
45
 
46
- image_button.click(prediction, inputs=image_input, outputs=prediction_output)
47
- clear_button.click(clear_output, inputs=[], outputs=[image_input, prediction_output])
48
 
49
- if __name__ == "__main__":
50
- demo.launch()
 
 
 
 
 
1
  import gradio as gr
2
+ import requests
3
+
 
4
  def save_non_abuse_class(top_classes):
 
 
5
  dic_predict = {
6
  "Non Abuse Image": 0,
7
  "Abuse Image": 0
8
  }
9
+
10
+ top_class_json = top_classes['message']
11
+
12
  for k, v in top_class_json.items():
13
+ if k.startswith(("Nice emoji", "Non Abuse")):
14
  dic_predict["Non Abuse Image"] += v
15
  else:
16
  dic_predict["Abuse Image"] += v
17
+
18
  dic_predict = sorted(dic_predict.items(), key=lambda x: x[1], reverse=True)
19
  dic_predict = {k: v for k, v in dic_predict}
20
  return dic_predict
21
+
 
 
 
22
  def prediction(img):
23
+ response = requests.post('http://18.139.116.122/', files={'img_input': img})
24
+ top_classes = response.json()
25
  save_class = save_non_abuse_class(top_classes)
26
  prediction_str = "\n".join([f"{index+1}. {property} : {round(value*100,2)}%" for index, (property, value) in enumerate(save_class.items())])
27
  return prediction_str
 
29
  def clear_output():
30
  image_input.value = None
31
  prediction_output.value = None
 
32
 
33
+ image_input = gr.inputs.Image()
34
+ prediction_output = gr.outputs.Textbox()
 
 
 
 
35
 
36
+ image_button = gr.Button("Make prediction")
37
+ clear_button = gr.Button("Clear")
38
 
39
+ iface = gr.Interface(inputs=image_input, outputs=prediction_output,
40
+ title="Abuse Image Classifier",
41
+ description="Classify an image as abuse or non-abuse.",
42
+ theme="default")
43
+
44
+ iface.launch()