fix(utils): Enhance the dependencies check to include pip distribution

[aiqwe]

With the 0.4.0 release, Datatrove transitioned to numpy 2.0, and the pyproject.toml was updated to ensure that the version of fasttext used is compatible with numpy 2.x.

However, if users do not start with a totally new virtual environment and just continue using fasttext-wheels compatible with numpy 1.x, we check only verifies the module name fasttext, which can cause an error to be raised during execution.

To solve this issue, I modfied below:

• add _is_distribution_available in _import_utils to strictly manage pip distributions.
• modified the check_required_dependencies function to perform above.
• also modified lid, language_filter, fasttext_filter to check pip distribution fasttext-numpy2-wheel

During pytest, an error related to the Tibetan language occurred, but since it is not relevant to my region, I did not make any additional changes.

[guipenedo]

Thanks a lot! Can you check the "Allow edits from maintainers" checkbox so that I may fix the tests?
https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/allowing-changes-to-a-pull-request-branch-created-from-a-fork

[aiqwe]

@guipenedo thx! 😀

However, the checkbox was already checked when I submitted the PR.

I think the feature could not be working properly.
If necessary, I’m fine with submitting the PR again after the test code is updated.

[guipenedo]

my bad, seems to have worked now