federated-learning-framework 0.0.61__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -0,0 +1,9 @@
1
+ from .active_learning import query_active_learning
2
+ from .central_server import CentralServer
3
+ from .client_device import ClientDevice
4
+ from .connection import ConnectionServer, ConnectionClient
5
+ from .decorators import federated_learning_decorator, encryption_decorator
6
+ from .encryption import create_context, encrypt_weights, decrypt_weights
7
+ from .models.tensorflow_model import TensorFlowModel
8
+ from .models.pytorch_model import PyTorchModel
9
+ from .utils import setup_logging
@@ -0,0 +1,7 @@
1
+ import numpy as np
2
+
3
+ def query_active_learning(model, unlabeled_data, num_samples):
4
+ predictions = model.predict(unlabeled_data)
5
+ uncertainty = predictions.max(axis=1)
6
+ selected_indices = np.argsort(uncertainty)[:num_samples]
7
+ return selected_indices
@@ -0,0 +1,56 @@
1
+ import asyncio
2
+ import logging
3
+ import numpy as np
4
+ from federated_learning_framework.connection import ConnectionServer
5
+ from websockets.exceptions import ConnectionClosedError
6
+ from federated_learning_framework.encryption import create_context
7
+
8
+ class CentralServer:
9
+ def __init__(self, connection_type='websocket', host='0.0.0.0', port=8089, context=None):
10
+ self.model_weights = None
11
+ self.lock = asyncio.Lock()
12
+ self.clients = set()
13
+ self.logger = logging.getLogger(__name__)
14
+ self.connection = ConnectionServer(connection_type, host, port, self.handle_client)
15
+ self.context = context or create_context()
16
+
17
+ async def run_server(self):
18
+ self.logger.info("Central Server is starting...")
19
+ await self.connection.start()
20
+
21
+ async def handle_client(self, websocket, client_id):
22
+ self.clients.add(client_id)
23
+ self.logger.info(f"Central Server: Client {client_id} connected")
24
+ try:
25
+ while True:
26
+ message = await self.connection.receive(client_id)
27
+ if isinstance(message, dict):
28
+ if 'weights' in message:
29
+ await self.transmit_weights(message['weights'])
30
+ elif 'data_request' in message:
31
+ data = await self.get_data_from_client(client_id)
32
+ await self.send_data_to_client(client_id, {'data': data})
33
+ except ConnectionClosedError:
34
+ self.logger.info(f"Central Server: Client {client_id} disconnected")
35
+ finally:
36
+ self.clients.remove(client_id)
37
+
38
+ async def transmit_weights(self, weights):
39
+ async with self.lock:
40
+ self.model_weights = weights
41
+ await asyncio.gather(*[self.connection.send(client_id, {'weights': self.model_weights}) for client_id in self.clients])
42
+ self.logger.info("Central Server: Transmitted weights to clients")
43
+
44
+ async def send_data_to_client(self, client_id, data):
45
+ self.logger.info(f"Central Server: Sending data to client {client_id}")
46
+ await self.connection.send(client_id, data)
47
+
48
+ async def get_data_from_client(self, client_id):
49
+ self.logger.info(f"Central Server: Requesting data from client {client_id}. Simulating response.")
50
+ await asyncio.sleep(1)
51
+ return np.random.rand(10, 3072)
52
+
53
+ def query_active_learning(self, unlabeled_data, model):
54
+ uncertainty = model.predict(unlabeled_data)
55
+ selected_indices = np.argsort(uncertainty.max(axis=1))[:5]
56
+ return selected_indices
@@ -0,0 +1,59 @@
1
+ import asyncio
2
+ import logging
3
+ import tensorflow as tf
4
+ from federated_learning_framework.encryption import encrypt_weights, decrypt_weights
5
+ from federated_learning_framework.models.tensorflow_model import TensorFlowModel
6
+ import websockets
7
+
8
+ class ClientDevice:
9
+ def __init__(self, client_id, model: TensorFlowModel, context):
10
+ self.client_id = client_id
11
+ self.model = model
12
+ self.context = context
13
+ self.connection = None
14
+ self.logger = logging.getLogger(__name__)
15
+
16
+ async def connect_to_central_server(self, uri):
17
+ try:
18
+ self.connection = await websockets.connect(uri)
19
+ await self.connection.send({'client_id': self.client_id})
20
+ self.logger.info(f"Client {self.client_id}: Connected to central server at {uri}")
21
+ except Exception as e:
22
+ self.logger.error(f"Client {self.client_id}: Error connecting to central server: {e}")
23
+
24
+ async def federated_learning(self, x_train, y_train):
25
+ try:
26
+ while True:
27
+ weights = await self.receive_weights()
28
+ if weights is None:
29
+ break
30
+ self.model.set_weights(decrypt_weights(self.context, weights))
31
+ self.model.train(x_train, y_train, epochs=1)
32
+ new_weights = self.model.get_weights()
33
+ await self.send_weights(encrypt_weights(self.context, new_weights))
34
+ except Exception as e:
35
+ self.logger.error(f"Client {self.client_id}: Error in federated learning loop: {e}")
36
+
37
+ async def receive_weights(self):
38
+ try:
39
+ message = await self.connection.recv()
40
+ self.logger.info(f"Client {self.client_id}: Received weights")
41
+ return message['weights']
42
+ except Exception as e:
43
+ self.logger.error(f"Client {self.client_id}: Error receiving weights: {e}")
44
+
45
+ async def send_weights(self, weights):
46
+ try:
47
+ await self.connection.send({'weights': weights})
48
+ self.logger.info(f"Client {self.client_id}: Sent weights to central server")
49
+ except Exception as e:
50
+ self.logger.error(f"Client {self.client_id}: Error sending weights: {e}")
51
+
52
+ async def request_data(self):
53
+ try:
54
+ await self.connection.send({'data_request': True})
55
+ data = await self.connection.recv()
56
+ self.logger.info(f"Client {self.client_id}: Received data from central server")
57
+ return data['data']
58
+ except Exception as e:
59
+ self.logger.error(f"Client {self.client_id}: Error requesting data: {e}")
@@ -0,0 +1,54 @@
1
+ import asyncio
2
+ import websockets
3
+ import pickle
4
+ from websockets.exceptions import ConnectionClosedError
5
+
6
+ class ConnectionServer:
7
+ def __init__(self, connection_type, host, port, client_handler):
8
+ self.connection_type = connection_type
9
+ self.host = host
10
+ self.port = port
11
+ self.client_handler = client_handler
12
+ self.clients = {}
13
+
14
+ async def start(self):
15
+ if self.connection_type == 'websocket':
16
+ async with websockets.serve(self.handle_client, self.host, self.port):
17
+ await asyncio.Future() # Run forever
18
+ else:
19
+ raise NotImplementedError(f"Connection type {self.connection_type} not supported")
20
+
21
+ async def handle_client(self, websocket, path):
22
+ client_id = len(self.clients) + 1
23
+ self.clients[client_id] = websocket
24
+ await self.client_handler(websocket, client_id)
25
+
26
+ async def send(self, client_id, message):
27
+ client = self.clients[client_id]
28
+ serialized_message = pickle.dumps(message)
29
+ await client.send(serialized_message)
30
+
31
+ async def receive(self, client_id):
32
+ client = self.clients[client_id]
33
+ message = await client.recv()
34
+ return pickle.loads(message)
35
+
36
+ class ConnectionClient:
37
+ def __init__(self, connection_type, uri):
38
+ self.connection_type = connection_type
39
+ self.uri = uri
40
+ self.connection = None
41
+
42
+ async def connect(self):
43
+ if self.connection_type == 'websocket':
44
+ self.connection = await websockets.connect(self.uri)
45
+ else:
46
+ raise NotImplementedError(f"Connection type {self.connection_type} not supported")
47
+
48
+ async def send(self, message):
49
+ serialized_message = pickle.dumps(message)
50
+ await self.connection.send(serialized_message)
51
+
52
+ async def receive(self):
53
+ message = await self.connection.recv()
54
+ return pickle.loads(message)
@@ -0,0 +1,24 @@
1
+ # federated_learning_framework/decorators.py
2
+ import functools
3
+
4
+ def federated_learning_decorator(uri):
5
+ def decorator(func):
6
+ @functools.wraps(func)
7
+ async def wrapper(*args, **kwargs):
8
+ central_server = kwargs.get('central_server')
9
+ client = kwargs.get('client')
10
+ await central_server.run_server()
11
+ await client.connect_to_central_server(uri)
12
+ return await func(*args, **kwargs)
13
+ return wrapper
14
+ return decorator
15
+
16
+ def encryption_decorator(context):
17
+ def decorator(func):
18
+ @functools.wraps(func)
19
+ async def wrapper(*args, **kwargs):
20
+ client = kwargs.get('client')
21
+ client.context = context
22
+ return await func(*args, **kwargs)
23
+ return wrapper
24
+ return decorator
@@ -0,0 +1,22 @@
1
+ import tenseal as ts
2
+
3
+ def create_context():
4
+ context = ts.context(ts.SCHEME_TYPE.CKKS, poly_modulus_degree=32768, coeff_mod_bit_sizes=[60, 40, 40, 60])
5
+ context.generate_galois_keys()
6
+ context.global_scale = 2**40
7
+ return context
8
+
9
+ def encrypt_weights(context, model_weights):
10
+ encrypted_weights = []
11
+ for weight in model_weights:
12
+ weight_array = weight.flatten().tolist()
13
+ encrypted_vector = ts.ckks_vector(context, weight_array)
14
+ encrypted_weights.append(encrypted_vector.serialize())
15
+ return encrypted_weights
16
+
17
+ def decrypt_weights(context, encrypted_weights):
18
+ decrypted_weights = []
19
+ for enc_weight in encrypted_weights:
20
+ enc_vector = ts.ckks_vector_from(context, enc_weight)
21
+ decrypted_weights.append(enc_vector.decrypt())
22
+ return decrypted_weights
File without changes
@@ -0,0 +1,18 @@
1
+ from abc import ABC, abstractmethod
2
+
3
+ class AbstractModel(ABC):
4
+ @abstractmethod
5
+ def get_weights(self):
6
+ pass
7
+
8
+ @abstractmethod
9
+ def set_weights(self, weights):
10
+ pass
11
+
12
+ @abstractmethod
13
+ def train(self, x_train, y_train, epochs=1):
14
+ pass
15
+
16
+ @abstractmethod
17
+ def predict(self, data):
18
+ pass
@@ -0,0 +1,28 @@
1
+ import torch
2
+ import torch.nn as nn
3
+ from federated_learning_framework.models.abstract_model import AbstractModel
4
+
5
+ class PyTorchModel(AbstractModel):
6
+ def __init__(self, model):
7
+ self.model = model
8
+
9
+ def get_weights(self):
10
+ return [param.data.numpy() for param in self.model.parameters()]
11
+
12
+ def set_weights(self, weights):
13
+ for param, weight in zip(self.model.parameters(), weights):
14
+ param.data = torch.tensor(weight, dtype=param.data.dtype)
15
+
16
+ def train(self, x_train, y_train, epochs=1):
17
+ criterion = nn.CrossEntropyLoss()
18
+ optimizer = torch.optim.SGD(self.model.parameters(), lr=0.01)
19
+ for _ in range(epochs):
20
+ optimizer.zero_grad()
21
+ outputs = self.model(x_train)
22
+ loss = criterion(outputs, y_train)
23
+ loss.backward()
24
+ optimizer.step()
25
+
26
+ def predict(self, data):
27
+ with torch.no_grad():
28
+ return self.model(data).numpy()
@@ -0,0 +1,18 @@
1
+ import tensorflow as tf
2
+ from federated_learning_framework.models.abstract_model import AbstractModel
3
+
4
+ class TensorFlowModel(AbstractModel):
5
+ def __init__(self, model):
6
+ self.model = model
7
+
8
+ def get_weights(self):
9
+ return self.model.get_weights()
10
+
11
+ def set_weights(self, weights):
12
+ self.model.set_weights(weights)
13
+
14
+ def train(self, x_train, y_train, epochs=1):
15
+ self.model.fit(x_train, y_train, epochs=epochs, verbose=0)
16
+
17
+ def predict(self, data):
18
+ return self.model.predict(data)
@@ -0,0 +1,4 @@
1
+ import logging
2
+
3
+ def setup_logging(logfile='federated_learning.log'):
4
+ logging.basicConfig(level=logging.INFO, format='%(asctime)s %(levelname)s:%(message)s', handlers=[logging.FileHandler(logfile), logging.StreamHandler()])
@@ -0,0 +1,37 @@
1
+ # Custom License for Federated Learning Framework
2
+
3
+ This license governs the use of the Federated Learning Framework. By using this framework, you agree to the following terms and conditions:
4
+
5
+ ## Academic Use
6
+
7
+ Academic use of this framework is permitted free of charge, provided that proper attribution is given. When using this framework in academic works, please include the following citation:
8
+
9
+ ```
10
+ @framework{
11
+ author = {Mehrdad Javadi},
12
+ author_email = {mehrdaddjavadi@gamil.com}
13
+ title = {Federated Learning Framework},
14
+ year = {2024},
15
+ url = {https://github.com/mehrdaddjavadi/federated_learning_framework}
16
+ }
17
+ ```
18
+
19
+ ## Commercial, Governmental, and Other Non-Academic Use
20
+
21
+ For commercial, governmental, and other non-academic uses, please contact the author to obtain a usage license. All such uses are subject to approval and may require a licensing fee.
22
+
23
+ Contact: mehrdaddjavadi@gmail.com or mehrdadjavadi64@gmail.com
24
+
25
+
26
+ ## Rights
27
+
28
+ All rights to this framework, including its use and distribution, belong to the author. Unauthorized use, distribution, or modification of this framework is strictly prohibited.
29
+
30
+ ## Disclaimer
31
+
32
+ This framework is provided "as is", without warranty of any kind, express or implied, including but not limited to the warranties of merchantability, fitness for a particular purpose, and no infringement. In no event shall the author be liable for any claim, damages, or other liability, whether in an action of contract, tort, or otherwise, arising from, out of, or in connection with the framework or the use or other dealings in the framework.
33
+
34
+ ---
35
+
36
+ Mehrdad javadi
37
+ 2024
@@ -0,0 +1,333 @@
1
+ Metadata-Version: 2.1
2
+ Name: federated-learning-framework
3
+ Version: 0.0.61
4
+ Summary: A modular and extensible framework for federated learning applications.
5
+ Home-page: https://github.com/mehrdaddjavadi/federated_learning_framework
6
+ Author: Mehrdad Javadi
7
+ Author-email: mehrdaddjavadi@gmail.com
8
+ Keywords: federated learning,machine learning,deep learning,active learning,encryption,homomorphic encryption
9
+ Classifier: Development Status :: 4 - Beta
10
+ Classifier: Intended Audience :: Developers
11
+ Classifier: Intended Audience :: Education
12
+ Classifier: Intended Audience :: Science/Research
13
+ Classifier: Programming Language :: Python :: 3
14
+ Classifier: Programming Language :: Python :: 3.6
15
+ Classifier: Programming Language :: Python :: 3.7
16
+ Classifier: Programming Language :: Python :: 3.8
17
+ Classifier: Programming Language :: Python :: 3.9
18
+ Classifier: Programming Language :: Python :: 3.10
19
+ Classifier: Topic :: Software Development :: Libraries :: Python Modules
20
+ Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
21
+ Requires-Python: >=3.6
22
+ Description-Content-Type: text/markdown
23
+ License-File: LICENSE
24
+ Requires-Dist: numpy
25
+ Requires-Dist: tensorflow
26
+ Requires-Dist: websockets
27
+ Requires-Dist: pytest
28
+ Requires-Dist: tenseal
29
+
30
+ # Federated Learning Framework
31
+
32
+ ## Overview
33
+
34
+ Welcome to the Federated Learning Framework, a modular and extensible solution for implementing federated learning across various applications. Harness the power of collective intelligence, ensure data privacy with homomorphic encryption, and apply it to domains like NLP, autonomous vehicles, drones, and more.
35
+
36
+ ## Features
37
+
38
+ - **Modular and Extensible**: Easily customizable for different machine learning and deep learning applications.
39
+ - **Secure**: Utilizes homomorphic encryption to ensure data privacy.
40
+ - **Active Learning**: Incorporates active learning strategies to improve model performance.
41
+ - **Flexible Communication**: Supports various connection methods including socket programming.
42
+ - **Customizable**: Users can edit and control every part of the framework with various functions.
43
+
44
+ ## Potential Applications
45
+
46
+ ### Healthcare
47
+
48
+ Federated learning can be used to train models on patient data from multiple hospitals without sharing sensitive information. This approach can improve medical diagnostics and treatment recommendations while preserving patient privacy.
49
+
50
+ ### Autonomous Vehicles
51
+
52
+ By collecting and learning from data across multiple autonomous vehicles, the framework can help improve the safety and performance of self-driving cars without exposing individual vehicle data.
53
+
54
+ ### Drones
55
+
56
+ Drones can use federated learning to share and learn from data collected during their operations, enhancing their navigation, object detection, and other capabilities while ensuring data security.
57
+
58
+ ### Natural Language Processing (NLP)
59
+
60
+ Federated learning can be applied to train NLP models on data from multiple sources, such as user devices, to improve language understanding and generation without compromising user privacy.
61
+
62
+ ### Finance
63
+
64
+ Financial institutions can use federated learning to develop fraud detection and risk management models by leveraging data from multiple sources while keeping customer data secure.
65
+
66
+ ### Smart Homes and IoT Devices
67
+
68
+ IoT devices in smart homes can collaboratively learn from user interactions to optimize performance and provide better services without sharing raw data.
69
+
70
+ ## Detailed Component Description
71
+
72
+ ### Central Server
73
+
74
+ **File:** `central_server.py`
75
+
76
+ The central server orchestrates the federated learning process by coordinating the communication and aggregation of model weights from various client devices.
77
+
78
+ **Key Functions:**
79
+
80
+ - `run_server`: Starts the server to handle client connections.
81
+ - `handle_client`: Manages incoming messages from clients.
82
+ - `transmit_weights`: Broadcasts the aggregated weights to clients.
83
+ - `send_data_to_client`: Sends specific data to a client.
84
+ - `get_data_from_client`: Requests and receives data from a client.
85
+ - `query_active_learning`: Implements active learning strategies to select data for labeling.
86
+
87
+ ### Client Device
88
+
89
+ **File:** `client_device.py`
90
+
91
+ Client devices perform local training on their datasets and communicate with the central server.
92
+
93
+ **Key Functions:**
94
+
95
+ - `connect_to_central_server`: Connects to the central server.
96
+ - `federated_learning`: Coordinates local training and communication with the server.
97
+ - `receive_weights`: Receives model weights from the central server.
98
+ - `send_weights`: Sends model weights to the central server.
99
+ - `receive_data`: Receives data from the central server.
100
+
101
+ ### Encryption
102
+
103
+ **File:** `encryption.py`
104
+
105
+ Provides functions for creating encryption contexts and encrypting/decrypting model weights.
106
+
107
+ **Key Functions:**
108
+
109
+ - `create_context`: Sets up the encryption context using TenSEAL.
110
+ - `encrypt_weights`: Encrypts model weights.
111
+ - `decrypt_weights`: Decrypts encrypted model weights.
112
+
113
+ ### Active Learning
114
+
115
+ **File:** `active_learning.py`
116
+
117
+ Implements active learning strategies to enhance the training process by selectively querying informative data points.
118
+
119
+ **Key Functions:**
120
+
121
+ - `select_informative_samples`: Selects samples for labeling based on uncertainty.
122
+
123
+ ### Connection
124
+
125
+ **File:** `connection.py`
126
+
127
+ Manages the connection types and protocols (e.g., WebSocket) for communication between the central server and client devices.
128
+
129
+ **Key Functions:**
130
+
131
+ - `run_server`: Starts a WebSocket server.
132
+ - `connect_to_server`: Establishes a WebSocket connection to the server.
133
+
134
+ ### Decorators
135
+
136
+ **File:** `decorators.py`
137
+
138
+ Provides decorators for adding federated learning and encryption functionalities to functions.
139
+
140
+ **Key Functions:**
141
+
142
+ - `federated_learning_decorator`: Wraps a function to enable federated learning.
143
+ - `encryption_decorator`: Wraps a function to enable homomorphic encryption.
144
+
145
+ ### Utilities
146
+
147
+ **File:** `utils.py`
148
+
149
+ Includes utility functions used throughout the framework.
150
+
151
+ ## Installation
152
+
153
+ Clone the repository:
154
+
155
+ ```sh
156
+ git clone https://github.com/mehrdaddjavadi/federated_learning_framework.git
157
+ ```
158
+
159
+ Navigate to the directory:
160
+
161
+ ```sh
162
+ cd federated_learning_framework
163
+ ```
164
+
165
+ Install the dependencies:
166
+
167
+ ```sh
168
+ pip install -r requirements.txt
169
+ ```
170
+
171
+ ## Usage
172
+
173
+ ### Setting Up the Central Server
174
+
175
+ ```python
176
+ import asyncio
177
+ from federated_learning_framework.central_server import CentralServer
178
+
179
+ async def main():
180
+ server = CentralServer()
181
+ await server.run_server()
182
+
183
+ asyncio.run(main())
184
+ ```
185
+
186
+ ### Setting Up the Central Server On Interactive Environment Like Jupyter Notebook
187
+
188
+ ```python
189
+ import nest_asyncio
190
+ import asyncio
191
+ from federated_learning_framework.central_server import CentralServer
192
+
193
+ nest_asyncio.apply()
194
+
195
+ async def main():
196
+ server = CentralServer()
197
+ await server.run_server()
198
+
199
+ # If running in an environment with an existing event loop
200
+ if __name__ == "__main__":
201
+ asyncio.run(main())
202
+ ```
203
+
204
+ ### Setting Up a Client Device
205
+
206
+ ```python
207
+ import asyncio
208
+ import tensorflow as tf
209
+ from federated_learning_framework.client_device import ClientDevice
210
+ from federated_learning_framework.encryption import create_context
211
+
212
+ # Define your model
213
+ model = tf.keras.Sequential([
214
+ tf.keras.layers.Dense(4, activation='relu', input_shape=(3072,)),
215
+ tf.keras.layers.Dense(10, activation='softmax')
216
+ ])
217
+
218
+ # Create context for encryption
219
+ context = create_context()
220
+
221
+ # Initialize the client device
222
+ client = ClientDevice(client_id=1, model=model, context=context)
223
+
224
+ async def main():
225
+ uri = "ws://localhost:8089"
226
+ await client.connect_to_central_server(uri)
227
+ x_train, y_train = ... # Load your training data
228
+ await client.federated_learning(uri, x_train, y_train)
229
+ # Optionally receive data from central server
230
+ data = await client.receive_data()
231
+ print(f"Received data: {data}")
232
+
233
+ asyncio.run(main())
234
+ ```
235
+
236
+ ### Sample Execution Script Using Decorators For Interactive Environments Like Colab And Jupyter Notebook
237
+
238
+ ```python
239
+ import asyncio
240
+ import tensorflow as tf
241
+ import numpy as np
242
+ from federated_learning_framework.client_device import ClientDevice
243
+ from federated_learning_framework.central_server import CentralServer
244
+ from federated_learning_framework.encryption import create_context
245
+ from federated_learning_framework.models.tensorflow_model import TensorFlowModel
246
+
247
+ # Setup logging
248
+ import logging
249
+ logging.basicConfig(level=logging.INFO)
250
+
251
+ # Define a simple TensorFlow model
252
+ model = tf.keras.Sequential([
253
+ tf.keras.layers.Dense(4, activation='relu', input_shape=(3072,)),
254
+ tf.keras.layers.Dense(10, activation='softmax')
255
+ ])
256
+ wrapped_model = TensorFlowModel(model)
257
+
258
+ # Create encryption context
259
+ context = create_context()
260
+
261
+ # Initialize server and clients
262
+ central_server = CentralServer(context=context)
263
+ client1 = ClientDevice(client_id=1, model=wrapped_model, context=context)
264
+ client2 = ClientDevice(client_id=2, model=wrapped_model, context=context)
265
+
266
+ # Dummy training data
267
+ x_train = np.random.rand(10, 3072)
268
+ y_train = np.random.randint(0, 10, 10)
269
+
270
+ async def main():
271
+ await asyncio.gather(
272
+ central_server.run_server(),
273
+ client1.connect_to_central_server("ws://localhost:8089"),
274
+ client2.connect_to_central_server("ws://localhost:8089"),
275
+ client1.federated_learning(x_train, y_train),
276
+ client2.federated_learning(x_train, y_train)
277
+ )
278
+
279
+ asyncio.run(main())
280
+
281
+ ```
282
+
283
+ ### Using Decorators
284
+
285
+ ```python
286
+ import asyncio
287
+ import tensorflow as tf
288
+ from federated_learning_framework.decorators import federated_learning_decorator, encryption_decorator
289
+ from federated_learning_framework.client_device import ClientDevice
290
+ from federated_learning_framework.encryption import create_context
291
+
292
+ # Create context for encryption
293
+ context = create_context()
294
+
295
+ # Define your model
296
+ model = tf.keras.Sequential([
297
+ tf.keras.layers.Dense(4, activation='relu', input_shape=(3072,)),
298
+ tf.keras.layers.Dense(10, activation='softmax')
299
+ ])
300
+
301
+ @federated_learning_decorator(uri="ws://localhost:8089")
302
+ @encryption_decorator(context=context)
303
+ async def main():
304
+ client = ClientDevice(client_id=1, model=model, context=context)
305
+ await client.connect_to_central_server('ws://localhost:8089')
306
+ x_train, y_train = ... # Load your training data
307
+ await client.federated_learning('ws://localhost:8089', x_train, y_train)
308
+
309
+ asyncio.run(main())
310
+ ```
311
+
312
+ ## Running Tests
313
+
314
+ To run the tests, execute the following command in the root directory:
315
+
316
+ ```sh
317
+ python -m unittest discover -s tests
318
+ ```
319
+
320
+ ## License
321
+
322
+ The usage of this library is free for academic work with proper referencing. For business, governmental, and any other types of usage, please contact me directly. All rights are reserved.
323
+
324
+ **Contact:** mehrdaddjavadi@gmail.com
325
+
326
+ ## Contributing
327
+
328
+ Feel free to contribute by submitting a pull request or opening an issue.
329
+
330
+ ```
331
+
332
+ Copy and paste this into your README.md file. This format provides a clear, organized structure and includes all necessary details and instructions for potential users and contributors.
333
+ ```
@@ -0,0 +1,24 @@
1
+ federated_learning_framework/__init__.py,sha256=HvZKThSQYzF5ofyU4UIgaaInB8mR9Y5jeZEd43l1_t0,482
2
+ federated_learning_framework/active_learning.py,sha256=jv6xzbRsHdzieMDwEtTpEn8YL_0M4a3DOawul0X5EnU,269
3
+ federated_learning_framework/central_server.py,sha256=eexr9XoZt0I5hGj1iG3HwX7CmnTX2KAjXfZGMkpRbB8,2620
4
+ federated_learning_framework/client_device.py,sha256=8b5FPXuWJ-PukR4CYzpY9JSVbn-kxtJRq-AhXXAy_rM,2659
5
+ federated_learning_framework/connection.py,sha256=D7zkKkgcFBhUEMcHUOLWZWJEO3jjxxFUrAishBDBT2w,1958
6
+ federated_learning_framework/decorators.py,sha256=gihFUEeIla0osy3319u8sMx0YxAdZqmAfkdRKmJsHXA,825
7
+ federated_learning_framework/encryption.py,sha256=bNarMUqYuyJG52yceGdSyW5jIDyVky2KnX7gqzFdKtY,840
8
+ federated_learning_framework/utils.py,sha256=t4dCUuAqvDach0qZObVoTfaQjXh06v5cpVfLxs_cRQ0,231
9
+ federated_learning_framework/models/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
10
+ federated_learning_framework/models/abstract_model.py,sha256=oNn4auUpt7XqOnfF5l0ebCWn9olPqb5ADwfibIcajQA,357
11
+ federated_learning_framework/models/pytorch_model.py,sha256=kibGM_dzM4uZIcxB3ZCdbv8KPkiKzFT9Vqnu4CLh9fg,988
12
+ federated_learning_framework/models/tensorflow_model.py,sha256=a-ZZbzoYjOL8Pza2KRn4q-TYBm-7OBuYeeOBzTFXuUs,548
13
+ tests/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
14
+ tests/test_active_learning.py,sha256=qrnaCkYwdWTu_Jvz17jz0QElogPmKtayKy_qG7P_U08,535
15
+ tests/test_central_server.py,sha256=OOGvSOlKMyMHbrXwDOgdA-QYJMOP7uLCiLpzLBLoM9M,604
16
+ tests/test_client_device.py,sha256=hRLZWUUPCrF-BHK46QtpJBdGjO4WZm5tcoZGQtefKyk,932
17
+ tests/test_connection.py,sha256=E5lFwzvB8FCpjfuePwfc8eQq-CGDJBLW6GmtFB0FldA,921
18
+ tests/test_encryption.py,sha256=PHSRxqS4ltepe7ztOhNQcrfHOYhMMRKqKQJU0Dsa8-o,374
19
+ tests/test_utils.py,sha256=H7Cp9nOf4IOcLeNZKtAUZt-YY3x0c1WiUHAmQKROHWo,272
20
+ federated_learning_framework-0.0.61.dist-info/LICENSE,sha256=Igduf49yy8CINu7sRZg4TvMGy7jOi0xNgmL3e4PRoe0,1646
21
+ federated_learning_framework-0.0.61.dist-info/METADATA,sha256=K-JIgZKigswGbyx97Z4RBXcNDg9A8NrElpG9MAf9b3o,11188
22
+ federated_learning_framework-0.0.61.dist-info/WHEEL,sha256=R0nc6qTxuoLk7ShA2_Y-UWkN8ZdfDBG2B6Eqpz2WXbs,91
23
+ federated_learning_framework-0.0.61.dist-info/top_level.txt,sha256=bnGlrzFER9O24oPGb7hrGhFKBtcrx321yE2CJuwiRMY,35
24
+ federated_learning_framework-0.0.61.dist-info/RECORD,,
@@ -0,0 +1,5 @@
1
+ Wheel-Version: 1.0
2
+ Generator: setuptools (72.1.0)
3
+ Root-Is-Purelib: true
4
+ Tag: py3-none-any
5
+
@@ -0,0 +1,2 @@
1
+ federated_learning_framework
2
+ tests
tests/__init__.py ADDED
File without changes
@@ -0,0 +1,12 @@
1
+ import numpy as np
2
+ import tensorflow as tf
3
+ from federated_learning_framework.active_learning import query_active_learning
4
+
5
+ def test_active_learning():
6
+ model = tf.keras.Sequential([
7
+ tf.keras.layers.Dense(10, input_shape=(3072,), activation='softmax')
8
+ ])
9
+ model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy'])
10
+ unlabeled_data = np.random.rand(100, 3072)
11
+ selected_indices = query_active_learning(model, unlabeled_data, 5)
12
+ assert len(selected_indices) == 5
@@ -0,0 +1,18 @@
1
+ import asyncio
2
+ import pytest
3
+ from federated_learning_framework.central_server import CentralServer
4
+ from federated_learning_framework.encryption import create_context
5
+
6
+ @pytest.mark.asyncio
7
+ async def test_central_server():
8
+ context = create_context()
9
+ server = CentralServer(context=context)
10
+ server_task = asyncio.create_task(server.run_server())
11
+
12
+ await asyncio.sleep(1) # Give the server some time to start
13
+
14
+ # Simulate client connections and other test scenarios here
15
+
16
+ server_task.cancel()
17
+ with pytest.raises(asyncio.CancelledError):
18
+ await server_task
@@ -0,0 +1,25 @@
1
+ import asyncio
2
+ import pytest
3
+ import tensorflow as tf
4
+ from federated_learning_framework.client_device import ClientDevice
5
+ from federated_learning_framework.encryption import create_context
6
+ from federated_learning_framework.models.tensorflow_model import TensorFlowModel
7
+
8
+ @pytest.mark.asyncio
9
+ async def test_client_device():
10
+ context = create_context()
11
+ model = tf.keras.Sequential([
12
+ tf.keras.layers.Dense(10, input_shape=(3072,), activation='softmax')
13
+ ])
14
+ model = TensorFlowModel(model)
15
+ client = ClientDevice(client_id=1, model=model, context=context)
16
+
17
+ connect_task = asyncio.create_task(client.connect_to_central_server('ws://localhost:8089'))
18
+
19
+ await asyncio.sleep(1) # Give the client some time to connect
20
+
21
+ # Simulate communication and other test scenarios here
22
+
23
+ connect_task.cancel()
24
+ with pytest.raises(asyncio.CancelledError):
25
+ await connect_task
@@ -0,0 +1,28 @@
1
+ import asyncio
2
+ import pytest
3
+ from federated_learning_framework.connection import ConnectionServer, ConnectionClient
4
+
5
+ @pytest.mark.asyncio
6
+ async def test_connection():
7
+ async def handle_client(connection, client_id):
8
+ message = await connection.receive()
9
+ await connection.send(message)
10
+
11
+ server = ConnectionServer('websocket', 'localhost', 8089, handle_client)
12
+ server_task = asyncio.create_task(server.start())
13
+
14
+ await asyncio.sleep(1) # Give the server some time to start
15
+
16
+ client = ConnectionClient('websocket', 'ws://localhost:8089')
17
+ await client.connect()
18
+
19
+ test_message = "test_message"
20
+ await client.send(test_message)
21
+ received_message = await client.receive()
22
+
23
+ assert received_message == test_message
24
+
25
+ await client.connection.close()
26
+ server_task.cancel()
27
+ with pytest.raises(asyncio.CancelledError):
28
+ await server_task
@@ -0,0 +1,9 @@
1
+ import numpy as np
2
+ from federated_learning_framework.encryption import create_context, encrypt_weights, decrypt_weights
3
+
4
+ def test_encryption():
5
+ context = create_context()
6
+ weights = [np.random.rand(10, 10)]
7
+ encrypted = encrypt_weights(context, weights)
8
+ decrypted = decrypt_weights(context, encrypted)
9
+ assert np.allclose(weights[0], decrypted[0])
tests/test_utils.py ADDED
@@ -0,0 +1,10 @@
1
+ import logging
2
+ from federated_learning_framework.utils import setup_logging
3
+
4
+ def test_setup_logging():
5
+ logger = setup_logging()
6
+
7
+ assert isinstance(logger, logging.Logger)
8
+ assert logger.level == logging.INFO
9
+
10
+ logger.info("Test logging setup")