Added GuacaMol Dataset to torchdrug

[adiv5]

This request involves use of GuacaMol's Benchmarking dataset(link) and I have implemented it into torchdrug. I have used guacamol's pip package to download and process the smiles to CSV . There is an alternative available too, where this repo provides a csv version of guacamol directly , let me know if we can use that instead.

PS: This is My First Open Source Contribution, so I am open to feedback and help