GH-39779: [Python] Expose force_virtual_addressing in PyArrow

[teihenn]

Rationale for this change / What changes are included in this PR?

To expose force_virtual_addressing in PyArrow.

Are these changes tested?

Existing unit tests are not broken, and a new test case have been added.

Are there any user-facing changes?

pyarrow.fs.S3FileSystem: it becomes possible to specify the argument 'force_virtual_addressing'.

• Closes: [Python] Expose force_virtual_addressing in PyArrow #39779

[github-actions]

⚠️ GitHub issue #39779 has been automatically assigned in GitHub to PR creator.

[AlenkaF]

Thank you for you contribution @yo1956 !

From a quick look I think that the Cython layer will have to be updated to include the force_virtual_addressing

arrow/python/pyarrow/includes/libarrow_fs.pxd

Lines 161 to 185 in 21ffd82

	cdef cppclass CS3Options "arrow::fs::S3Options":
	c_string region
	double connect_timeout
	double request_timeout
	c_string endpoint_override
	c_string scheme
	c_bool background_writes
	c_bool allow_bucket_creation
	c_bool allow_bucket_deletion
	shared_ptr[const CKeyValueMetadata] default_metadata
	c_string role_arn
	c_string session_name
	c_string external_id
	int load_frequency
	CS3ProxyOptions proxy_options
	CS3CredentialsKind credentials_kind
	shared_ptr[CS3RetryStrategy] retry_strategy
	void ConfigureDefaultCredentials()
	void ConfigureAccessKey(const c_string& access_key,
	const c_string& secret_key,
	const c_string& session_token)
	c_string GetAccessKey()
	c_string GetSecretKey()
	c_string GetSessionToken()
	c_bool Equals(const CS3Options& other)

[teihenn]

@AlenkaF
Thank you for reviewing!
I'll fix it.

[teihenn]

@AlenkaF
I've made the modifications, but it seems there are errors with CI that doesn't appear to be related to my changes.
I'm not entirely sure, though. I would be grateful if you could provide some advice when you have time.

[AlenkaF]

The failures are not related, there is a fix already open: #39827. Similar for the Sphinx build.

We can wait for them to get merged and rebase then.

[AlenkaF]

LGTM, thank you for the updates!

I do not think there are any additions needed in the tests, but would like to have one extra approve before merging.

[teihenn]

LGTM, thank you for the updates!

I do not think there are any additions needed in the tests, but would like to have one extra approve before merging.

@AlenkaF
Thank you!!

[teihenn]

@jorisvandenbossche
Could you please review this when you have time?

[jorisvandenbossche]

Looks good, thanks!

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 6 benchmarking runs that have been run so far on merge-commit 3d45ac9.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details. It also includes information about 2 possible false positives for unstable benchmarks that are known to sometimes produce them.