Adapter batching

[aMahanna]

pip install torch
pip install git+https://github.com/arangoml/pyg-adapter.git@feature/batching

As we work towards cleaning up the ArangoDB to PyG & PyG to ArangoDB transformations, reworking the batching process is crucial to optimizing the adapter's memory consumption.

Branch has been based off of feature/exceptions in order to allow for its review & merge after #17

New changes:

1. ArangoDB to PyG batching via the batch_size parameter of the db.aql.execute() method (python-arango)
2. PyG to ArangoDB batching via a new batch_size parameter in the pyg_to_arangodb method
3. Reworking the fetch_adb_docs method to only fetch the document attributes specified by the metagraph (get_aql_return_value)

adbpyg_adapter.arangodb_to_pyg("imdb", metagraph, batch_size=100) # Fetches & processes up to 100 ArangoDB Documents/Edges at a time

adbpyg_adapter.pyg_to_arangodb("karate", data, metagraph, batch_size=100) # Processes & ingests up to 100 PyG Nodes/Edges at a time

[geenen124]

LGTM - nice work!