[Python][C++] Support CSV compression in write_dataset

### Describe the usage question you have. Please include as many useful details as possible.

Hi,

I write a dataset to S3 with partitioned on a column as CSV files. But my CSV files should be gzip compressed, I couldn't find anything on docs to do that. Only way it seems, I should write with compressed outputstream however it's not one file.

This is the code I'm using;

```
ds.write_dataset(data, write_path,
format='csv',
partitioning=["p_col"],
partitioning_flavor="hive",
existing_data_behavior='delete_matching')
```

Is there anyway to achieve this to be output gzip files ?

If it's not available, I think it's a good enhancement to consider.

### Component(s)

Format, Python

Start at the Python ds.write_dataset entry point with format='csv' and trace the corresponding C++ dataset-writing path, using the partitioned S3 example as the reproduction case. Examine how compressed output streams are handled, then verify gzip-compressed CSV output for partitioned writes; done means the requested API works and the resulting files are readable.

Written by the indexing model from the issue text.