Merge branch 'main' into improve_ci_speed

* main:
  Remove line with typo from script (#2001)
  Fix simulation error caused by Pydantic version in Ray (#2002)
  Update mxnet examples dependency versions (#1982)
  Update torch version in pyproject.toml files (#1993)
  Update tensorflow examples dependency versions (#1984)
  Fix the baseline template docstrings (#1983)
  Fix Baselines CI PyLint error by using Resampling (#1998)
  Make tensorflow optional when using tensorboard (#1879)
  Update dependencies for MXNet example (#1988)
  Fix `starlette` vulnerability (#1992)
  Fix vulnerability detected by Dependabot (#1990)
  Fix dependabot critical errors (#1989)
  Add Ruff UP check (#1946)
  Update torch examples dependency versions (#1981)
  Remove && from the end of lines in test.sh (#1979)
  Baselines docs updates (#1977)
  Fixes GitHub CI issue because of behaviour change (#1976)

.github/workflows/flower-swift_sync.yml

@@ -11,7 +11,7 @@ jobs:
     steps:
       - uses: actions/checkout@v3
       - name: Pushes src/swift to flower-swift repository
-        uses: cpina/github-action-push-to-another-repository@0a14457
+        uses: cpina/github-action-push-to-another-repository@0a14457bb28b04dfa1652e0ffdfda866d2845c73
         env:
           SSH_DEPLOY_KEY: ${{ secrets.FLOWER_SWIFT_SSH }}
         with:

baselines/baseline_template/baseline_template/__init__.py

@@ -0,0 +1 @@
+"""Template baseline package."""

baselines/baseline_template/baseline_template/client.py

@@ -1,5 +1,5 @@
-##################################################################################
-# Please use this module to define your client class and a function to construct #
-# such clients. Most of the code here is expected to be defined inside a class   #
-# that inherits from `flwr.client.NumPyClient` or `flwr.client.Client`.          #
-##################################################################################
+"""Define your client class and a function to construct such clients.
+
+Please overwrite `flwr.client.NumPyClient` or `flwr.client.Client` and create a function
+to instantiate your client.
+"""

baselines/baseline_template/baseline_template/dataset.py

@@ -1,9 +1,10 @@
-##################################################################################
-# Please use this module to define the functions to create the dataloaders for   #
-# your dataset (for both the clients and the server). If you are using a custom  #
-# dataset class, this module is the place to define it. If your dataset requires #
-# to be downloaded (and this is not done automatically -- e.g. as it is the case #
-# for many dataset in TorchVision) and partitioned, please include all those     #
-# functions and logic in the `dataset_preparation.py` module. You can all those  #
-# functions from functions/methods defined here of course.                       #
-##################################################################################
+"""Handle basic dataset creation.
+
+In case of PyTorch it should return dataloaders for your dataset (for both the clients
+and the server). If you are using a custom dataset class, this module is the place to
+define it. If your dataset requires to be downloaded (and this is not done
+automatically -- e.g. as it is the case for many dataset in TorchVision) and
+partitioned, please include all those functions and logic in the
+`dataset_preparation.py` module. You can use all those functions from functions/methods
+defined here of course.
+"""

baselines/baseline_template/baseline_template/dataset_preparation.py

@@ -1,15 +1,11 @@
-####################################################################################
-# So you have a dataset that needs to be partitioned? Or does your dataset require #
-# a non-standard way of obtaining it (i.e. because you are not limiting your base- #
-# line to MNIST, CIFAR-10 et al)? Then this is the module for you! Please add here #
-# all the necessary logic to either download, uncompress, pre/post-process your    #
-# dataset (or all of the above). If the desired way of running your baseline is    #
-# to first download the dataset and partition it and then run the experiments,     #
-# please uncomment the lines below and tell us in the README.md (see the "Running  #
-# the Experiment" block) that this file should be executed first.                  #
-####################################################################################
-
-
+"""Handle the dataset partitioning and (optionally) complex downloads.
+
+Please add here all the necessary logic to either download, uncompress, pre/post-process
+your dataset (or all of the above). If the desired way of running your baseline is to
+first download the dataset and partition it and then run the experiments, please
+uncomment the lines below and tell us in the README.md (see the "Running the Experiment"
+block) that this file should be executed first.
+"""
 # import hydra
 # from hydra.core.hydra_config import HydraConfig
 # from hydra.utils import call, instantiate

baselines/baseline_template/baseline_template/main.py

@@ -1,42 +1,37 @@
-###########################################################################
-# This is the main file, it creates and connects all the building         #
-# blocks of your experiment (processes the dataset, defines the clients,  #
-# the strategy, and how the global model is going to be evaluated, etc)   #
-# and it starts the simulation. At the end, this script saves the results.#
-###########################################################################
-
+"""Create and connect the building blocks for your experiments; start the simulation.
 
+It includes processioning the dataset, instantiate strategy, specify how the global
+model is going to be evaluated, etc. At the end, this script saves the results.
+"""
 # these are the basic packages you'll need here
 # feel free to remove some if aren't needed
-import flwr as fl
 import hydra
-from hydra.core.hydra_config import HydraConfig
-from hydra.utils import call, instantiate
 from omegaconf import DictConfig, OmegaConf
 
 
 @hydra.main(config_path="conf", config_name="base", version_base=None)
 def main(cfg: DictConfig) -> None:
-    """Main function to run this baseline.
+    """Run the baseline.
 
     Parameters
     ----------
     cfg : DictConfig
         An omegaconf object that stores the hydra config.
     """
-
-    ## 1. print parsed config
+    ## 1. Print parsed config
     print(OmegaConf.to_yaml(cfg))
 
-    ## 2. prepare your dataset
+    ## 2. Prepare your dataset
     # here you should call a function in datasets.py that returns whatever is needed to:
-    # (1) ensure the server can access the dataset used to evaluate your model after aggregation
-    # (2) tell each client what dataset partitions they should use (e.g. a this could be a location
-    # in the file system, a list of dataloader, a list of ids to extract from a dataset, it's up to you)
+    # (1) ensure the server can access the dataset used to evaluate your model after
+    # aggregation
+    # (2) tell each client what dataset partitions they should use (e.g. a this could
+    # be a location in the file system, a list of dataloader, a list of ids to extract
+    # from a dataset, it's up to you)
 
     ## 3. Define your clients
-    # Define a function that returns another function that will be used during simulation
-    # to instantiate each individual client
+    # Define a function that returns another function that will be used during
+    # simulation to instantiate each individual client
     # client_fn = client.<my_function_that_returns_a_function>()
 
     ## 4. Define your strategy
@@ -55,4 +50,4 @@ def main(cfg: DictConfig) -> None:
     # data = strategy.get_my_custom_data() -- assuming you have such method defined.
     # Hydra will generate for you a directory each time you run the code. You
     # can retrieve the path to that directory with this:
-    # save_path = HydraConfig.get().runtime.output_dir
+    # save_path = HydraConfig.get().runtime.output_dir

baselines/baseline_template/baseline_template/models.py

@@ -1,7 +1,7 @@
-##################################################################################
-# Please use this module to define your models, and training and eval            #
-# functions. If your model is 100% off-the-shelf (e.g. directly from torchvision #
-# without requiring modifications) you might be better off instantiating your    #
-# model directly from the Hydra config. In this way, swapping your model for     #
-# another one can be done without changing the python code at all                #
-##################################################################################
+"""Define our models, and training and eval functions.
+
+If your model is 100% off-the-shelf (e.g. directly from torchvision without requiring
+modifications) you might be better off instantiating your  model directly from the Hydra
+config. In this way, swapping your model for  another one can be done without changing
+the python code at all
+"""

baselines/baseline_template/baseline_template/server.py

@@ -1,6 +1,5 @@
-##############################################################################
-# Please use this module to define the function called by your strategy in   #
-# order to evaluate the state of the global model. You might use this module #
-# also to define a new Server class (please note this is not needed in most  #
-# settings).                                                                 #
-##############################################################################
+"""Create global evaluation function.
+
+Optionally, also define a new Server class (please note this is not needed in most
+settings).
+"""

baselines/baseline_template/baseline_template/strategy.py

@@ -1,5 +1,5 @@
-#############################################################################
-# Only use this module if you need to define a custom strategy whether      #
-# because it is not yet implemented in Flower or because you want to extend #
-# or modify the functionality of an existing strategy.                      #
-#############################################################################
+"""Optionally define a custom strategy.
+
+Needed only when the strategy is not yet implemented in Flower or because you want to
+extend or modify the functionality of an existing strategy.
+"""

baselines/baseline_template/baseline_template/utils.py

@@ -1,5 +1,6 @@
-#################################################################################
-# Please define here any utility function that's not directly relevant to       #
-# the other (more FL specific) python modules. For example, you may define      #
-# here things like: loading a model from a checkpoint, saving results, plotting #
-#################################################################################
+"""Define any utility function.
+
+They are not directly relevant to  the other (more FL specific) python modules. For
+example, you may define here things like: loading a model from a checkpoint, saving
+results, plotting.
+"""

baselines/flwr_baselines/dev/setup-defaults.sh

@@ -31,7 +31,6 @@ else
     # If pyenv is already installed, check for a newer version
     read -p 'Pyenv already installed, do you want to updating it y/[n]? ' update
     update="${update:-"n"}"
-    update="${update,,}"
 
     if [ $update == "y" ]
     then

baselines/flwr_baselines/flwr_baselines/publications/leaf/femnist/dataset/nist_preprocessor.py

@@ -188,7 +188,7 @@ def _preprocess_images(self) -> pd.DataFrame:
             file_path = row["path_by_writer"]
             img = Image.open(file_path)
             gray = img.convert("L")
-            gray.thumbnail(resized_size, Image.LANCZOS)
+            gray.thumbnail(resized_size, Image.Resampling.LANCZOS)
             writer_id = row["writer_id"]
             character = row["character"]
             if writer_id not in writer_to_character_to_count:

dev/setup-defaults.sh

@@ -31,7 +31,6 @@ else
     # If pyenv is already installed, check for a newer version
     read -p 'Pyenv already installed, do you want to updating it y/[n]? ' update
     update="${update:-"n"}"
-    update="${update,,}"
 
     if [ $update == "y" ]
     then

dev/test.sh

@@ -6,52 +6,52 @@ echo "=== test.sh ==="
 
 echo "- Start Python checks"
 
-echo "- clang-format:  start" &&
-clang-format --Werror --dry-run src/proto/flwr/proto/* &&
-echo "- clang-format:  done" &&
+echo "- clang-format:  start"
+clang-format --Werror --dry-run src/proto/flwr/proto/*
+echo "- clang-format:  done"
 
-echo "- isort: start" &&
-python -m isort --check-only --skip src/py/flwr/proto src/py/flwr &&
-echo "- isort: done" &&
+echo "- isort: start"
+python -m isort --check-only --skip src/py/flwr/proto src/py/flwr
+echo "- isort: done"
 
-echo "- black: start" &&
-python -m black --exclude "src\/py\/flwr\/proto" --check src/py/flwr &&
-echo "- black: done" &&
+echo "- black: start"
+python -m black --exclude "src\/py\/flwr\/proto" --check src/py/flwr
+echo "- black: done"
 
-echo "- init_py_check: start" &&
-python -m flwr_tool.init_py_check src/py/flwr src/py/flwr_tool &&
-echo "- init_py_check: done" &&
+echo "- init_py_check: start"
+python -m flwr_tool.init_py_check src/py/flwr src/py/flwr_tool
+echo "- init_py_check: done"
 
-echo "- docformatter: start" &&
-python -m docformatter -c -r src/py/flwr -e src/py/flwr/proto &&
-echo "- docformatter:  done" &&
+echo "- docformatter: start"
+python -m docformatter -c -r src/py/flwr -e src/py/flwr/proto
+echo "- docformatter:  done"
 
-echo "- ruff: start" &&
-python -m ruff check src/py/flwr &&
-echo "- ruff: done" &&
+echo "- ruff: start"
+python -m ruff check src/py/flwr
+echo "- ruff: done"
 
-echo "- mypy: start" &&
-python -m mypy src/py &&
-echo "- mypy: done" &&
+echo "- mypy: start"
+python -m mypy src/py
+echo "- mypy: done"
 
-echo "- pylint: start" &&
-python -m pylint --ignore=src/py/flwr/proto src/py/flwr &&
-echo "- pylint: done" &&
+echo "- pylint: start"
+python -m pylint --ignore=src/py/flwr/proto src/py/flwr
+echo "- pylint: done"
 
-echo "- flake8: start" &&
-python -m flake8 src/py/flwr &&
-echo "- flake8: done" &&
+echo "- flake8: start"
+python -m flake8 src/py/flwr
+echo "- flake8: done"
 
-echo "- pytest: start" &&
-python -m pytest --cov=src/py/flwr &&
-echo "- pytest: done" &&
+echo "- pytest: start"
+python -m pytest --cov=src/py/flwr
+echo "- pytest: done"
 
 echo "- All Python checks passed"
 
 echo "- Start Markdown checks"
 
-echo "- mdformat: start" &&
-python -m mdformat --check --number doc/source/tutorial examples &&
-echo "- mdformat: done" &&
+echo "- mdformat: start"
+python -m mdformat --check --number doc/source/tutorial examples
+echo "- mdformat: done"
 
 echo "- All Markdown checks passed"

doc/source/conf.py

@@ -124,7 +124,6 @@
     #     "color-brand-content": "#292F36",
     #     "color-admonition-background": "#F2B705",
     # },
-    "announcement": 'Flower Summit 2023 <a href="https://flower.dev/conf/flower-summit-2023/">register now</a>',
 }
 
 # Add any paths that contain custom static files (such as style sheets) here,

doc/source/contributing-baselines.rst

@@ -48,6 +48,6 @@ We are adopting `Hydra <https://hydra.cc/>`_ as the default mechanism to manage
 Usability
 ---------
 
-Flower is known and loved for its usability. Therefore, make sure that your baseline or experiment can be executed with a single command such as :code:`conda run -m <your-baseline>.main` or :code:`python main.py` (when sourced into your environment). We provide you with a `template-baseline <https://github.com/adap/flower/tree/sor_template/baselines/flwr_baselines>`_ to use as guidance when contributing your baseline. Having all baselines follow a homogenous structure helps users to tryout many baselines without the overheads of having to understand each individual codebase. Similarly, by using Hydra throughout, users will immediately know how to parameterise your experiments directly from the command line.
+Flower is known and loved for its usability. Therefore, make sure that your baseline or experiment can be executed with a single command such as :code:`conda run -m <your-baseline>.main` or :code:`python main.py` (when sourced into your environment). We provide you with a `template-baseline <https://github.com/adap/flower/tree/main/baselines/baseline_template>`_ to use as guidance when contributing your baseline. Having all baselines follow a homogenous structure helps users to tryout many baselines without the overheads of having to understand each individual codebase. Similarly, by using Hydra throughout, users will immediately know how to parameterise your experiments directly from the command line.
 
 We look forward to your contribution!

examples/advanced_pytorch/pyproject.toml

@@ -12,8 +12,8 @@ authors = [
 ]
 
 [tool.poetry.dependencies]
-python = "^3.7"
-flwr = "^1.0.0"
-torch = "^1.12.0"
-torchvision = "^0.13.0"
+python = ">=3.8,<3.11"
+flwr = ">=1.0,<2.0"
+torch = "1.13.1"
+torchvision = "0.13.0"
 validators = "0.18.2"

examples/advanced_pytorch/requirements.txt

@@ -1,3 +1,4 @@
-flwr~=1.4.0
-torch~=2.0.1
-torchvision~=0.15.2
+flwr>=1.0, <2.0
+torch==1.13.1
+torchvision==0.13.0
+

examples/advanced_tensorflow/pyproject.toml

@@ -10,7 +10,6 @@ authors = ["The Flower Authors <hello@flower.dev>"]
 
 [tool.poetry.dependencies]
 python = ">=3.8,<3.11"
-flwr = "^1.0.0"
-tensorflow-cpu = {version = "^2.9.1, !=2.11.1", markers="platform_machine == 'x86_64'"}
-tensorflow-macos = {version = "^2.9.1, !=2.11.1", markers="sys_platform == 'darwin' and platform_machine == 'arm64'"}
-
+flwr = ">=1.0,<2.0"
+tensorflow-cpu = {version = ">=2.9.1,<2.11.1 || >2.11.1", markers = "platform_machine == \"x86_64\""}
+tensorflow-macos = {version = ">=2.9.1,<2.11.1 || >2.11.1", markers = "sys_platform == \"darwin\" and platform_machine == \"arm64\""}

examples/advanced_tensorflow/requirements.txt

@@ -1,2 +1,3 @@
-flwr~=1.4.0
-tensorflow~=2.12.0
+flwr>=1.0, <2.0
+tensorflow-macos>=2.9.1, != 2.11.1 ; sys_platform == "darwin" and platform_machine == "arm64"
+tensorflow-cpu>=2.9.1, != 2.11.1 ; platform_machine == "x86_64"

examples/android/pyproject.toml

@@ -10,7 +10,7 @@ authors = ["The Flower Authors <hello@flower.dev>"]
 
 [tool.poetry.dependencies]
 python = ">=3.8,<3.11"
-flwr = "^1.0.0"
 # flwr = { path = "../../", develop = true }  # Development
-tensorflow-cpu = {version = "^2.9.1, !=2.11.1", markers="platform_machine == 'x86_64'"}
-tensorflow-macos = {version = "^2.9.1, !=2.11.1", markers="sys_platform == 'darwin' and platform_machine == 'arm64'"}
+flwr = ">=1.0,<2.0"
+tensorflow-cpu = {version = ">=2.9.1,<2.11.1 || >2.11.1", markers = "platform_machine == \"x86_64\""}
+tensorflow-macos = {version = ">=2.9.1,<2.11.1 || >2.11.1", markers = "sys_platform == \"darwin\" and platform_machine == \"arm64\""}

examples/android/requirements.txt

@@ -1 +1,3 @@
-numpy~=1.21.1
+flwr>=1.0, <2.0
+tensorflow-macos>=2.9.1, != 2.11.1 ; sys_platform == "darwin" and platform_machine == "arm64"
+tensorflow-cpu>=2.9.1, != 2.11.1 ; platform_machine == "x86_64"

examples/dp-sgd-mnist/pyproject.toml

@@ -12,9 +12,9 @@ authors = [
 ]
 
 [tool.poetry.dependencies]
-python = "^3.8"
-flwr = "^1.0.0"
+python = ">=3.8,<3.11"
 # flwr = { path = "../../", develop = true }  # Development
-tensorflow-cpu = {version = "^2.9.1, !=2.11.1", markers="platform_machine == 'x86_64'"}
-tensorflow-macos = {version = "^2.9.1, !=2.11.1", markers="sys_platform == 'darwin' and platform_machine == 'arm64'"}
-tensorflow-privacy = {git = "https://github.com/tensorflow/privacy", rev = "aaf4c25"}
+flwr = ">=1.0,<2.0"
+tensorflow-cpu = {version = ">=2.9.1,<2.11.1 || >2.11.1", markers = "platform_machine == \"x86_64\""}
+tensorflow-macos = {version = ">=2.9.1,<2.11.1 || >2.11.1", markers = "sys_platform == \"darwin\" and platform_machine == \"arm64\""}
+tensorflow-privacy = "0.8.10"

examples/dp-sgd-mnist/requirements.txt

@@ -1,4 +1,4 @@
-flwr~=1.4.0
-numpy~=1.21.1
-tensorflow~=2.12.0
-tensorflow_privacy~=0.8.9
+flwr>=1.0, <2.0
+tensorflow-macos>=2.9.1, != 2.11.1 ; sys_platform == "darwin" and platform_machine == "arm64"
+tensorflow-cpu>=2.9.1, != 2.11.1 ; platform_machine == "x86_64"
+tensorflow-privacy==0.8.10

examples/mt-pytorch/pyproject.toml

@@ -9,8 +9,8 @@ description = "Multi-Tenant Federated Learning with Flower and PyTorch"
 authors = ["The Flower Authors <hello@flower.dev>"]
 
 [tool.poetry.dependencies]
-python = "^3.7"
-flwr = { path = "../../", develop = true, extras = ["simulation", "rest"] }
-torch = "^1.12.0"
-torchvision = "^0.13.0"
-tqdm = "^4.63.0"
+python = ">=3.8,<3.11"
+flwr-nightly = {extras = ["rest", "simulation"], version = "^1.5.0.dev20230629"}
+torch = "1.13.1"
+torchvision = "0.13.0"
+tqdm = "4.65.0"