feat: 🎸 add /sizes (#679)
* feat: 🎸 add /sizes
* feat: 🎸 setup worker for /sizes endpoint + e2e
* style: 💄 fix import order
* fix: 🐛 fix error
* feat: 🎸 update docker image
* feat: 🎸 give both the num_bytes of RAM and parquet files
The previous value for the parquet size was erroneous, it was
representing the size of the splits in the memory (pa_table.nbytes).
/size now depends on /parquet-and-dataset-info instead of /dataset-info
in order to get the size of the parquet files.
* feat: 🎸 update docker images
