Wrong metadata produced for tfds-nightly

Short description
When trying to download the C4 dataset with apache-beam on google cloud dataflow, this exception gets thrown:

Traceback (most recent call last):
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/apache_beam/utils/processes.py", line 91, in check_output
    out = subprocess.check_output(*args, **kwargs)
  File "/usr/lib/python3.6/subprocess.py", line 356, in check_output
    **kwargs).stdout
  File "/usr/lib/python3.6/subprocess.py", line 438, in run
    output=stdout, stderr=stderr)
subprocess.CalledProcessError: Command '['/mnt/pccfs/backed_up/crytting/persuasion/createc4/bin/python3', '-m', 'pip', 'download', '--dest', '/tmp/dataflow-requirements-cache', '-r', '/tmp/beam_requirements.txt', '--exists-action', 'i', '--no-binary', ':all:']' returned non-zero exit status 1.

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/bin/tfds", line 8, in <module>
    sys.exit(launch_cli())
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/tensorflow_datasets/scripts/cli/main.py", line 86, in launch_cli
    app.run(main, flags_parser=_parse_flags)
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/absl/app.py", line 300, in run
    _run_main(main, args)
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/absl/app.py", line 251, in _run_main
    sys.exit(main(argv))
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/tensorflow_datasets/scripts/cli/main.py", line 81, in main
    args.subparser_fn(args)
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/tensorflow_datasets/scripts/cli/build.py", line 198, in _build_datasets
    _download_and_prepare(args, builder)
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/tensorflow_datasets/scripts/cli/build.py", line 353, in _download_and_prepare
    download_config=dl_config,
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/tensorflow_datasets/core/dataset_builder.py", line 434, in download_and_prepare
    download_config=download_config,
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/tensorflow_datasets/core/dataset_builder.py", line 1163, in _download_and_prepare
    in utils.tqdm(split_generators.items(), unit=" splits", leave=False)
  File "/usr/lib/python3.6/contextlib.py", line 88, in __exit__
    next(self.gen)
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/tensorflow_datasets/core/split_builder.py", line 177, in maybe_beam_pipeline
    self._beam_pipeline.__exit__(None, None, None)
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/apache_beam/pipeline.py", line 568, in __exit__
    self.result = self.run()
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/apache_beam/pipeline.py", line 518, in run
    allow_proto_holders=True).run(False)
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/apache_beam/pipeline.py", line 547, in run
    return self.runner.run_pipeline(self, self._options)
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/apache_beam/runners/dataflow/dataflow_runner.py", line 493, in run_pipeline
    artifacts=environments.python_sdk_dependencies(options)))
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/apache_beam/transforms/environments.py", line 626, in python_sdk_dependencies
    skip_prestaged_dependencies=skip_prestaged_dependencies))
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/apache_beam/runners/portability/stager.py", line 180, in create_job_resources
    setup_options.requirements_file, requirements_cache_path)
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/apache_beam/utils/retry.py", line 236, in wrapper
    return fun(*args, **kwargs)
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/apache_beam/runners/portability/stager.py", line 571, in _populate_requirements_cache
    processes.check_output(cmd_args, stderr=processes.STDOUT)
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/apache_beam/utils/processes.py", line 99, in check_output
    .format(traceback.format_exc(), args[0][6], error.output))
RuntimeError: Full traceback: Traceback (most recent call last):
  File "/mnt/pccfs/backed_up/crytting/persuasion/createc4/lib/python3.6/site-packages/apache_beam/utils/processes.py", line 91, in check_output
    out = subprocess.check_output(*args, **kwargs)
  File "/usr/lib/python3.6/subprocess.py", line 356, in check_output
    **kwargs).stdout
  File "/usr/lib/python3.6/subprocess.py", line 438, in run
    output=stdout, stderr=stderr)
subprocess.CalledProcessError: Command '['/mnt/pccfs/backed_up/crytting/persuasion/createc4/bin/python3', '-m', 'pip', 'download', '--dest', '/tmp/dataflow-requirements-cache', '-r', '/tmp/beam_requirements.txt', '--exists-action', 'i', '--no-binary', ':all:']' returned non-zero exit status 1.

 Pip install failed for package: -r
 Output from execution of subprocess: b"Collecting tfds-nightly[c4]\n  Using cached tfds-nightly-4.1.0.dev202012030107.tar.gz (3.1 MB)\n  WARNING: Generating metadata for package tfds-nightly produced metadata for project name tensorflow-datasets. Fix your #egg=tfds-nightly fragments.\nERROR: Requested tensorflow-datasets from https://files.pythonhosted.org/packages/c3/da/b28fdd59fb35feedcee53bbffc00d6afdcd10a4aa52daccff50919c0b093/tfds-nightly-4.1.0.dev202012030107.tar.gz#sha256=fb453c00f65e8033153d5038999857b801e2cbad0a3d62796aaee35a3523d640 (from -r /tmp/beam_requirements.txt (line 1)) has different name in metadata: 'tensorflow-datasets'\n"

Environment information

• Operating System: Ubuntu 18.04

• Python version: 3.6.9

• tensorflow-datasets/tfds-nightly version: tfds-nightly==4.1.0.dev202012030107

• tensorflow/tf-nightly version: tensorflow==2.3.1

• Does the issue still exists with the last tfds-nightly package (pip install --upgrade tfds-nightly) ?
  Yes

Reproduction instructions

python3 -m venv createc4
source createc4/bin/activate
pip install --upgrade pip
pip install tensorflow
pip install 'tfds-nightly[c4]'
pip install 'apache-beam[gcp]'

export GOOGLE_APPLICATION_CREDENTIALS="/mnt/pccfs/backed_up/crytting/persuasion/CreateC4-credential.json"

MY_PROJECT="createc4"
MY_BUCKET="create-c4"
MY_REGION="us-west1"

python3 -m tensorflow_datasets.scripts.download_and_prepare \
#tfds build \
  --datasets=c4/en \
  --data_dir=gs://$MY_BUCKET/tensorflow_datasets \
  --beam_pipeline_options="project=$MY_PROJECT,job_name=c4,staging_location=gs://$MY_BUCKET/binaries,temp_location=gs://$MY_BUCKET/temp,runner=DataflowRunner,requirements_file=/tmp/beam_requirements.txt,experiments=shuffle_mode=service,region=$MY_REGION,max_num_workers=448"

Expected behavior
My google cloud dataflow job to spin up a bunch of workers!

Additional context
I'm trying to follow the instructions here to process and download C4. Every time I try this again (by deleting my virtualenv and following the instructions from scratch), I run into a different problem with every up-to-date version of pip, tensorflow, tfds-nightly, and apache-beam.

Start by reproducing the tfds-nightly installation and Dataflow dependency-staging failure from the commands and versions in the issue. Trace how the tfds-nightly distribution metadata is generated and compare its declared project name with the package name requested by pip; done means the dependency download completes without the metadata-name error.