Remove unneeded tensorflow-privacy dependencies.

Please note:
1. I have manually tested the TFP release for python 3.10.
2. TFP does not define the set of python version that it is compatible with.

This change should be followed up with the following changes:
* Define the set of python version that TFP is compatible with.
* Setup a Python package building test.
* Release TFP so that projects depending on TFP can support Python 3.11

PiperOrigin-RevId: 570172856

README.md

@@ -31,14 +31,6 @@ better performance, it is also recommended to install TensorFlow with GPU
 support (detailed instructions on how to do this are available in the TensorFlow
 installation documentation).
 
-In addition to TensorFlow and its dependencies, other prerequisites are:
-
-*   `scipy` >= 0.17
-
-*   `mpmath` (for testing)
-
-*   `tensorflow_datasets` (for the RNN tutorial `lm_dpsgd_tutorial.py` only)
-
 ### Installing TensorFlow Privacy
 
 If you only want to use TensorFlow Privacy as a library, you can simply execute

requirements.txt

@@ -26,7 +26,6 @@ dm-tree==0.1.8
 dp-accounting==0.4.3
 immutabledict~=2.2
 matplotlib~=3.3
-mpmath~=1.2
 numpy~=1.21
 packaging~=22.0
 pandas~=1.4

setup.py

@@ -13,15 +13,14 @@
 # limitations under the License.
 """TensorFlow Privacy library setup file for pip."""
 
-from setuptools import find_packages
-from setuptools import setup
+import setuptools
 
 with open('tensorflow_privacy/version.py') as file:
   globals_dict = {}
   exec(file.read(), globals_dict)  # pylint: disable=exec-used
   VERSION = globals_dict['__version__']
 
-setup(
+setuptools.setup(
     name='tensorflow_privacy',
     version=VERSION,
     url='https://github.com/tensorflow/privacy',
@@ -36,15 +35,13 @@
         'numpy~=1.21',
         'packaging~=22.0',
         'pandas~=1.4',
-        'parameterized~=0.8',
         'scikit-learn>=1.0,==1.*',
         'scipy~=1.9',
         'statsmodels~=0.13',
-        'tensorflow-datasets~=4.5',
         'tensorflow-estimator~=2.4',
         'tensorflow-probability~=0.22.0',
         'tensorflow~=2.4',
         'tf-models-official~=2.13',
     ],
-    packages=find_packages(),
+    packages=setuptools.find_packages(),
 )

tools/test_pip_package.sh

@@ -28,7 +28,7 @@ main() {
   pip install --upgrade pip
   pip --version
 
-  # Publish the pip package.
+  # Test the pip package.
   package="$(ls "dist/"*".whl" | head -n1)"
   pip install --upgrade "${package}"
   pip freeze