Add option to write parquet output

[shntnu]

Currently, cyto_utils.output can write out csv, csv.gz, and tsv.

We'd like it to also be able to write out parquet.

The immediate use case we have in mind is converting SQLite backends to Parquet format, which make it a lot easier to work with large single-cell datasets.

This will be kinda redundant with cytominer-database functionality which can directly write out parquet but it will serve our very frequent need of working with existing large SQLite files.

[shntnu]

I wonder if it is as simple as this (needs to be tested)

pyarrow.parquet.write_table(pyarrow.Table.from_pandas(df), output_filename)

[shntnu]

I've tested it and that (pyarrow.parquet.write_table(pyarrow.Table.from_pandas(df), output_filename)) is probably all we need to do here

https://colab.research.google.com/drive/1D0J7FiqmzR49IjSKy2lQVtZ8Gez4cDkD?authuser=1#scrollTo=W3FJmSQXgBOd

[gwaybio]

related to #112

[gwaybio]

addressed in 2327ebc