Spaces:
Running
on
CPU Upgrade
Running
on
CPU Upgrade
chore: update
Browse files
app.py
CHANGED
@@ -1,13 +1,14 @@
|
|
1 |
import os
|
2 |
import shutil
|
3 |
import subprocess
|
|
|
4 |
from pathlib import Path
|
5 |
-
from time import time
|
6 |
from typing import List, Tuple, Union
|
7 |
|
8 |
import gradio as gr
|
9 |
import numpy as np
|
10 |
import pandas as pd
|
|
|
11 |
from preprocessing import pretty_print
|
12 |
from symptoms_categories import SYMPTOMS_LIST
|
13 |
|
@@ -16,16 +17,21 @@ from concrete.ml.deployment import FHEModelClient, FHEModelDev, FHEModelServer
|
|
16 |
from concrete.ml.sklearn import XGBClassifier as ConcreteXGBoostClassifier
|
17 |
|
18 |
INPUT_BROWSER_LIMIT = 635
|
19 |
-
|
20 |
# This repository's main necessary folders
|
21 |
REPO_DIR = Path(__file__).parent
|
22 |
MODEL_PATH = REPO_DIR / "client_folder"
|
23 |
KEYS_PATH = REPO_DIR / ".fhe_keys"
|
24 |
-
|
25 |
-
|
|
|
|
|
|
|
|
|
|
|
26 |
|
27 |
-
|
28 |
-
|
29 |
|
30 |
|
31 |
def clean_directory():
|
@@ -169,8 +175,8 @@ def key_gen_fn(user_symptoms):
|
|
169 |
|
170 |
# np.save(f".fhe_keys/{user_id}/eval_key.npy", serialized_evaluation_keys)
|
171 |
evaluation_key_path = KEYS_PATH / f"{user_id}/evaluation_key"
|
172 |
-
with evaluation_key_path.open("wb") as
|
173 |
-
|
174 |
|
175 |
serialized_evaluation_keys_shorten_hex = serialized_evaluation_keys.hex()[:INPUT_BROWSER_LIMIT]
|
176 |
|
@@ -200,7 +206,7 @@ def encrypt_fn(user_symptoms, user_id):
|
|
200 |
|
201 |
quant_user_symptoms = client.model.quantize_input(user_symptoms)
|
202 |
encrypted_quantized_user_symptoms = client.quantize_encrypt_serialize(user_symptoms)
|
203 |
-
|
204 |
encrypted_input_path = KEYS_PATH / f"{user_id}/encrypted_symptoms"
|
205 |
|
206 |
with encrypted_input_path.open("wb") as f:
|
@@ -227,46 +233,69 @@ def encrypt_fn(user_symptoms, user_id):
|
|
227 |
}
|
228 |
|
229 |
|
230 |
-
|
231 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
232 |
|
233 |
-
|
234 |
-
|
235 |
-
# filter_name (str): The current filter to consider.
|
236 |
-
# """
|
237 |
-
# # Get the evaluation key path
|
238 |
|
|
|
|
|
239 |
|
240 |
-
|
|
|
|
|
241 |
|
242 |
-
|
243 |
-
|
244 |
|
245 |
-
|
246 |
-
|
|
|
|
|
|
|
|
|
247 |
|
248 |
-
#
|
249 |
-
|
|
|
|
|
|
|
250 |
|
251 |
-
|
252 |
-
|
253 |
-
|
254 |
-
|
255 |
-
# }
|
256 |
|
257 |
-
#
|
258 |
-
|
259 |
-
|
260 |
-
|
|
|
|
|
|
|
|
|
261 |
|
262 |
-
|
263 |
-
# url = SERVER_URL + "send_input"
|
264 |
-
# with requests.post(
|
265 |
-
# url=url,
|
266 |
-
# data=data,
|
267 |
-
# files=files,
|
268 |
-
# ) as response:
|
269 |
-
# return response.ok
|
270 |
|
271 |
|
272 |
# def decrypt_prediction(encrypted_quantized_vect, user_id):
|
@@ -277,11 +306,13 @@ def encrypt_fn(user_symptoms, user_id):
|
|
277 |
# return predictions
|
278 |
|
279 |
|
|
|
280 |
|
|
|
281 |
|
282 |
-
def clear_all_btn():
|
283 |
return {
|
284 |
box_default: None,
|
|
|
285 |
user_id_textbox: None,
|
286 |
eval_key_textbox: None,
|
287 |
quant_vect_textbox: None,
|
@@ -291,13 +322,14 @@ def clear_all_btn():
|
|
291 |
error_box_1: gr.update(visible=False),
|
292 |
error_box_2: gr.update(visible=False),
|
293 |
error_box_3: gr.update(visible=False),
|
|
|
|
|
294 |
**{box: None for box in check_boxes},
|
295 |
}
|
296 |
|
297 |
|
298 |
if __name__ == "__main__":
|
299 |
print("Starting demo ...")
|
300 |
-
|
301 |
|
302 |
(df_train, X_train, X_test), (df_test, y_train, y_test) = load_data()
|
303 |
|
@@ -423,7 +455,7 @@ if __name__ == "__main__":
|
|
423 |
gr.Markdown("# Step 3: Encode the message with the private key")
|
424 |
gr.Markdown("Client side")
|
425 |
|
426 |
-
encrypt_btn = gr.Button("Encode the message with the private key
|
427 |
|
428 |
error_box_3 = gr.Textbox(label="Error", visible=False)
|
429 |
|
@@ -452,12 +484,25 @@ if __name__ == "__main__":
|
|
452 |
outputs=[vect_textbox, quant_vect_textbox, encrypted_vect_textbox, error_box_3],
|
453 |
)
|
454 |
|
455 |
-
gr.Markdown("# Step 4:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
456 |
gr.Markdown("Server side")
|
457 |
|
458 |
run_fhe = gr.Button("Run the FHE evaluation")
|
459 |
|
460 |
-
gr.Markdown("# Step
|
461 |
gr.Markdown("Server side")
|
462 |
|
463 |
decrypt_target_botton = gr.Button("Decrypt the sentiment")
|
@@ -478,10 +523,13 @@ if __name__ == "__main__":
|
|
478 |
error_box_1,
|
479 |
error_box_2,
|
480 |
error_box_3,
|
|
|
|
|
481 |
user_id_textbox,
|
482 |
eval_key_textbox,
|
483 |
quant_vect_textbox,
|
484 |
user_vector_textbox,
|
|
|
485 |
eval_key_len_textbox,
|
486 |
encrypted_vect_textbox,
|
487 |
*check_boxes,
|
|
|
1 |
import os
|
2 |
import shutil
|
3 |
import subprocess
|
4 |
+
import time
|
5 |
from pathlib import Path
|
|
|
6 |
from typing import List, Tuple, Union
|
7 |
|
8 |
import gradio as gr
|
9 |
import numpy as np
|
10 |
import pandas as pd
|
11 |
+
import requests
|
12 |
from preprocessing import pretty_print
|
13 |
from symptoms_categories import SYMPTOMS_LIST
|
14 |
|
|
|
17 |
from concrete.ml.sklearn import XGBClassifier as ConcreteXGBoostClassifier
|
18 |
|
19 |
INPUT_BROWSER_LIMIT = 635
|
20 |
+
SERVER_URL = "http://localhost:8000/"
|
21 |
# This repository's main necessary folders
|
22 |
REPO_DIR = Path(__file__).parent
|
23 |
MODEL_PATH = REPO_DIR / "client_folder"
|
24 |
KEYS_PATH = REPO_DIR / ".fhe_keys"
|
25 |
+
CLIENT_TMP_PATH = REPO_DIR / "client_tmp"
|
26 |
+
SERVER_TMP_PATH = REPO_DIR / "server_tmp"
|
27 |
+
|
28 |
+
# Create the necessary folders
|
29 |
+
KEYS_PATH.mkdir(exist_ok=True)
|
30 |
+
CLIENT_TMP_PATH.mkdir(exist_ok=True)
|
31 |
+
SERVER_TMP_PATH.mkdir(exist_ok=True)
|
32 |
|
33 |
+
subprocess.Popen(["uvicorn", "server:app"], cwd=REPO_DIR)
|
34 |
+
time.sleep(3)
|
35 |
|
36 |
|
37 |
def clean_directory():
|
|
|
175 |
|
176 |
# np.save(f".fhe_keys/{user_id}/eval_key.npy", serialized_evaluation_keys)
|
177 |
evaluation_key_path = KEYS_PATH / f"{user_id}/evaluation_key"
|
178 |
+
with evaluation_key_path.open("wb") as f:
|
179 |
+
f.write(serialized_evaluation_keys)
|
180 |
|
181 |
serialized_evaluation_keys_shorten_hex = serialized_evaluation_keys.hex()[:INPUT_BROWSER_LIMIT]
|
182 |
|
|
|
206 |
|
207 |
quant_user_symptoms = client.model.quantize_input(user_symptoms)
|
208 |
encrypted_quantized_user_symptoms = client.quantize_encrypt_serialize(user_symptoms)
|
209 |
+
assert isinstance(encrypted_quantized_user_symptoms, bytes)
|
210 |
encrypted_input_path = KEYS_PATH / f"{user_id}/encrypted_symptoms"
|
211 |
|
212 |
with encrypted_input_path.open("wb") as f:
|
|
|
233 |
}
|
234 |
|
235 |
|
236 |
+
def is_nan(input):
|
237 |
+
return input is None or (input is not None and len(input) < 1)
|
238 |
+
|
239 |
+
|
240 |
+
def send_input_fn(user_id, user_symptoms):
|
241 |
+
"""Send the encrypted input image as well as the evaluation key to the server.
|
242 |
+
|
243 |
+
Args:
|
244 |
+
user_id (int): The current user's ID.
|
245 |
+
filter_name (str): The current filter to consider.
|
246 |
+
"""
|
247 |
+
# Get the evaluation key path
|
248 |
+
|
249 |
+
if is_nan(user_id) or is_nan(user_symptoms):
|
250 |
+
return {
|
251 |
+
error_box_4: gr.update(
|
252 |
+
visible=True,
|
253 |
+
value="Please ensure that the evaluation key has been generated "
|
254 |
+
"and the symptoms have been submitted before sending the data to the server",
|
255 |
+
)
|
256 |
+
}
|
257 |
|
258 |
+
evaluation_key_path = KEYS_PATH / f"{user_id}/evaluation_key"
|
259 |
+
encrypted_input_path = KEYS_PATH / f"{user_id}/encrypted_symptoms"
|
|
|
|
|
|
|
260 |
|
261 |
+
if not evaluation_key_path.is_file():
|
262 |
+
print(f"Please generate the private key, first.{evaluation_key_path.is_file()=}")
|
263 |
|
264 |
+
return {
|
265 |
+
error_box_4: gr.update(visible=True, value="Please generate the private key first.")
|
266 |
+
}
|
267 |
|
268 |
+
if not encrypted_input_path.is_file():
|
269 |
+
print(f"Please submit your symptoms, first.{encrypted_input_path.is_file()=}")
|
270 |
|
271 |
+
return {
|
272 |
+
error_box_4: gr.update(
|
273 |
+
visible=True,
|
274 |
+
value="Please generate the private key and then encrypt an image first.",
|
275 |
+
)
|
276 |
+
}
|
277 |
|
278 |
+
# Define the data and files to post
|
279 |
+
data = {
|
280 |
+
"user_id": user_id,
|
281 |
+
"filter": user_symptoms,
|
282 |
+
}
|
283 |
|
284 |
+
files = [
|
285 |
+
("files", open(encrypted_input_path, "rb")),
|
286 |
+
("files", open(evaluation_key_path, "rb")),
|
287 |
+
]
|
|
|
288 |
|
289 |
+
# Send the encrypted input image and evaluation key to the server
|
290 |
+
url = SERVER_URL + "send_input"
|
291 |
+
with requests.post(
|
292 |
+
url=url,
|
293 |
+
data=data,
|
294 |
+
files=files,
|
295 |
+
) as response:
|
296 |
+
print(f"response.ok: {response.ok}")
|
297 |
|
298 |
+
return {error_box_4: gr.update(visible=False), server_response_box: gr.update(visible=True)}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
299 |
|
300 |
|
301 |
# def decrypt_prediction(encrypted_quantized_vect, user_id):
|
|
|
306 |
# return predictions
|
307 |
|
308 |
|
309 |
+
def clear_all_btn():
|
310 |
|
311 |
+
clean_directory()
|
312 |
|
|
|
313 |
return {
|
314 |
box_default: None,
|
315 |
+
vect_textbox: None,
|
316 |
user_id_textbox: None,
|
317 |
eval_key_textbox: None,
|
318 |
quant_vect_textbox: None,
|
|
|
322 |
error_box_1: gr.update(visible=False),
|
323 |
error_box_2: gr.update(visible=False),
|
324 |
error_box_3: gr.update(visible=False),
|
325 |
+
error_box_4: gr.update(visible=False),
|
326 |
+
server_response_box: gr.update(visible=False),
|
327 |
**{box: None for box in check_boxes},
|
328 |
}
|
329 |
|
330 |
|
331 |
if __name__ == "__main__":
|
332 |
print("Starting demo ...")
|
|
|
333 |
|
334 |
(df_train, X_train, X_test), (df_test, y_train, y_test) = load_data()
|
335 |
|
|
|
455 |
gr.Markdown("# Step 3: Encode the message with the private key")
|
456 |
gr.Markdown("Client side")
|
457 |
|
458 |
+
encrypt_btn = gr.Button("Encode the message with the private key")
|
459 |
|
460 |
error_box_3 = gr.Textbox(label="Error", visible=False)
|
461 |
|
|
|
484 |
outputs=[vect_textbox, quant_vect_textbox, encrypted_vect_textbox, error_box_3],
|
485 |
)
|
486 |
|
487 |
+
gr.Markdown("# Step 4: Send the encrypted data to the server.")
|
488 |
+
gr.Markdown("Client side")
|
489 |
+
|
490 |
+
send_input_btn = gr.Button("Send the encrypted data to the server..")
|
491 |
+
error_box_4 = gr.Textbox(label="Error", visible=False)
|
492 |
+
server_response_box = gr.Textbox(value="Data sent", visible=False, show_label=False)
|
493 |
+
|
494 |
+
send_input_btn.click(
|
495 |
+
send_input_fn,
|
496 |
+
inputs=[user_id_textbox, user_vector_textbox],
|
497 |
+
outputs=[error_box_4, server_response_box],
|
498 |
+
)
|
499 |
+
|
500 |
+
gr.Markdown("# Step 5: Run the FHE evaluation")
|
501 |
gr.Markdown("Server side")
|
502 |
|
503 |
run_fhe = gr.Button("Run the FHE evaluation")
|
504 |
|
505 |
+
gr.Markdown("# Step 6: Decrypt the sentiment")
|
506 |
gr.Markdown("Server side")
|
507 |
|
508 |
decrypt_target_botton = gr.Button("Decrypt the sentiment")
|
|
|
523 |
error_box_1,
|
524 |
error_box_2,
|
525 |
error_box_3,
|
526 |
+
error_box_4,
|
527 |
+
vect_textbox,
|
528 |
user_id_textbox,
|
529 |
eval_key_textbox,
|
530 |
quant_vect_textbox,
|
531 |
user_vector_textbox,
|
532 |
+
server_response_box,
|
533 |
eval_key_len_textbox,
|
534 |
encrypted_vect_textbox,
|
535 |
*check_boxes,
|
server.py
CHANGED
@@ -9,6 +9,11 @@ from fastapi.responses import JSONResponse, Response
|
|
9 |
|
10 |
from concrete.ml.deployment import FHEModelServer
|
11 |
|
|
|
|
|
|
|
|
|
|
|
12 |
# Initialize an instance of FastAPI
|
13 |
app = FastAPI()
|
14 |
|
@@ -29,65 +34,65 @@ def send_input(
|
|
29 |
filter: str = Form(),
|
30 |
files: List[UploadFile] = File(),
|
31 |
):
|
|
|
32 |
"""Send the inputs to the server."""
|
33 |
# Retrieve the encrypted input image and the evaluation key paths
|
34 |
-
|
35 |
-
|
36 |
|
37 |
-
# Write the files using the above paths
|
38 |
-
with
|
39 |
"wb"
|
40 |
) as evaluation_key:
|
41 |
-
|
42 |
evaluation_key.write(files[1].file.read())
|
43 |
|
44 |
|
45 |
-
@app.post("/run_fhe")
|
46 |
-
def run_fhe(
|
47 |
-
user_id: str = Form(),
|
48 |
-
filter: str = Form(),
|
49 |
-
):
|
50 |
-
"""Execute the filter on the encrypted input image using FHE."""
|
51 |
-
# Retrieve the encrypted input image and the evaluation key paths
|
52 |
-
encrypted_image_path = get_server_file_path("encrypted_image", user_id, filter)
|
53 |
-
evaluation_key_path = get_server_file_path("evaluation_key", user_id, filter)
|
54 |
|
55 |
-
|
56 |
-
|
57 |
-
|
58 |
-
|
59 |
-
|
60 |
-
|
|
|
|
|
|
|
61 |
|
62 |
-
|
63 |
-
|
|
|
|
|
|
|
|
|
64 |
|
65 |
-
|
66 |
-
|
67 |
-
encrypted_output_image = fhe_server.run(encrypted_image, evaluation_key)
|
68 |
-
fhe_execution_time = round(time.time() - start, 2)
|
69 |
|
70 |
-
|
71 |
-
|
|
|
|
|
72 |
|
73 |
-
|
74 |
-
|
75 |
-
encrypted_output.write(encrypted_output_image)
|
76 |
|
77 |
-
|
|
|
|
|
78 |
|
|
|
79 |
|
80 |
-
@app.post("/get_output")
|
81 |
-
def get_output(
|
82 |
-
user_id: str = Form(),
|
83 |
-
filter: str = Form(),
|
84 |
-
):
|
85 |
-
"""Retrieve the encrypted output image."""
|
86 |
-
# Retrieve the encrypted output image path
|
87 |
-
encrypted_output_path = get_server_file_path("encrypted_output", user_id, filter)
|
88 |
|
89 |
-
|
90 |
-
|
91 |
-
|
|
|
|
|
|
|
|
|
|
|
92 |
|
93 |
-
|
|
|
|
|
|
9 |
|
10 |
from concrete.ml.deployment import FHEModelServer
|
11 |
|
12 |
+
REPO_DIR = Path(__file__).parent
|
13 |
+
KEYS_PATH = REPO_DIR / ".fhe_keys"
|
14 |
+
MODEL_PATH = REPO_DIR / "client_folder"
|
15 |
+
|
16 |
+
SERVER_TMP_PATH = REPO_DIR / "server_tmp"
|
17 |
# Initialize an instance of FastAPI
|
18 |
app = FastAPI()
|
19 |
|
|
|
34 |
filter: str = Form(),
|
35 |
files: List[UploadFile] = File(),
|
36 |
):
|
37 |
+
|
38 |
"""Send the inputs to the server."""
|
39 |
# Retrieve the encrypted input image and the evaluation key paths
|
40 |
+
evaluation_key_path = SERVER_TMP_PATH / f"{user_id}_valuation_key"
|
41 |
+
encrypted_input_path = SERVER_TMP_PATH / f"{user_id}_encrypted_symptoms"
|
42 |
|
43 |
+
# # Write the files using the above paths
|
44 |
+
with encrypted_input_path.open("wb") as encrypted_input, evaluation_key_path.open(
|
45 |
"wb"
|
46 |
) as evaluation_key:
|
47 |
+
encrypted_input.write(files[0].file.read())
|
48 |
evaluation_key.write(files[1].file.read())
|
49 |
|
50 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
51 |
|
52 |
+
# @app.post("/run_fhe")
|
53 |
+
# def run_fhe(
|
54 |
+
# user_id: str = Form(),
|
55 |
+
# filter: str = Form(),
|
56 |
+
# ):
|
57 |
+
# """Execute the filter on the encrypted input image using FHE."""
|
58 |
+
# Retrieve the encrypted input image and the evaluation key paths
|
59 |
+
# encrypted_image_path = get_server_file_path("encrypted_image", user_id, filter)
|
60 |
+
# evaluation_key_path = get_server_file_path("evaluation_key", user_id, filter)
|
61 |
|
62 |
+
# Read the files using the above paths
|
63 |
+
# with encrypted_image_path.open("rb") as encrypted_image_file, evaluation_key_path.open(
|
64 |
+
# "rb"
|
65 |
+
# ) as evaluation_key_file:
|
66 |
+
# encrypted_image = encrypted_image_file.read()
|
67 |
+
# evaluation_key = evaluation_key_file.read()
|
68 |
|
69 |
+
# Load the FHE server
|
70 |
+
# fhe_server = FHEServer(FILTERS_PATH / f"{filter}/deployment")
|
|
|
|
|
71 |
|
72 |
+
# Run the FHE execution
|
73 |
+
# start = time.time()
|
74 |
+
# encrypted_output_image = fhe_server.run(encrypted_image, evaluation_key)
|
75 |
+
# fhe_execution_time = round(time.time() - start, 2)
|
76 |
|
77 |
+
# Retrieve the encrypted output image path
|
78 |
+
# encrypted_output_path = get_server_file_path("encrypted_output", user_id, filter)
|
|
|
79 |
|
80 |
+
# Write the file using the above path
|
81 |
+
# with encrypted_output_path.open("wb") as encrypted_output:
|
82 |
+
# encrypted_output.write(encrypted_output_image)
|
83 |
|
84 |
+
# return JSONResponse(content=fhe_execution_time)
|
85 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
86 |
|
87 |
+
# @app.post("/get_output")
|
88 |
+
# def get_output(
|
89 |
+
# user_id: str = Form(),
|
90 |
+
# filter: str = Form(),
|
91 |
+
# ):
|
92 |
+
# """Retrieve the encrypted output image."""
|
93 |
+
# Retrieve the encrypted output image path
|
94 |
+
# encrypted_output_path = get_server_file_path("encrypted_output", user_id, filter)
|
95 |
|
96 |
+
# Read the file using the above path
|
97 |
+
# with encrypted_output_path.open("rb") as encrypted_output_file:
|
98 |
+
# encrypted_output = encrypted_output_file.read()
|