Alarm Reported When Building the TensorFlow pip Software Package

Symptom

An alarm is reported when building the TensorFlow pip software package. Some alarm information is as follows:

!!
  check.warn(importable)
/usr/local/lib/python3.9/site-packages/setuptools/command/build_py.py:204: _Warning: Package 'tensorflow.xla_aot_runtime_src.tensorflow.tsl.platform' is absent from the `packages` configuration.
!!

        ********************************************************************************
        ############################
        # Package would be ignored #
        ############################
        Python recognizes 'tensorflow.xla_aot_runtime_src.tensorflow.tsl.platform' as an importable package[^1],
        but it is absent from setuptools' `packages` configuration.

        This leads to an ambiguous overall configuration. If you want to distribute this
        package, please make sure that 'tensorflow.xla_aot_runtime_src.tensorflow.tsl.platform' is explicitly added
        to the `packages` configuration field.

        Alternatively, you can also rely on setuptools' discovery methods
        (for example by using `find_namespace_packages(...)`/`find_namespace:`
        instead of `find_packages(...)`/`find:`).

        You can read more about "package discovery" on setuptools documentation page:

        - https://setuptools.pypa.io/en/latest/userguide/package_discovery.html

        If you don't want 'tensorflow.xla_aot_runtime_src.tensorflow.tsl.platform' to be distributed and are
        already explicitly excluding 'tensorflow.xla_aot_runtime_src.tensorflow.tsl.platform' via
        `find_namespace_packages(...)/find_namespace` or `find_packages(...)/find`,
        you can try to use `exclude_package_data`, or `include-package-data=False` in
        combination with a more fine grained `package-data` configuration.

        You can read more about "package data files" on setuptools documentation page:

        - https://setuptools.pypa.io/en/latest/userguide/datafiles.html


        [^1]: For Python, any directory (with suitable naming) can be imported,
              even if it does not contain any `.py` files.
              On the other hand, currently there is no concept of package data
              directory, all directories are treated like packages.
        ********************************************************************************

Key Process and Cause Analysis

Key error information: setuptools methods change as the version evolves. Replace find_packages(...) or find: with find_namespace_packages(...) or find_namespace:.

1

Alternatively, you can also rely on setuptools' discovery methods(for example by using `find_namespace_packages(...)`/`find_namespace:`instead of `find_packages(...)`/`find:`).

Root cause analysis: The setuptools version is incompatible.

Conclusion and Solution

1. Query the current setuptools version.

   1	pip3 list | grep setuptools

   Version information:

   

2. Downgrade setuptools.

   1 2	pip3 uninstall setuptools pip3 install setuptools==56.0.0

   

   Setuptools 56.0.0 is installed successfully.

   

   It has been verified that alarms can be cleared in setuptools 42.0.2, 50.0.0, 56.0.0 and 57.0.0.

3. Rebuild the package.

   1	./bazel-bin/tensorflow/tools/pip_package/build_pip_package ./output