This PR is primarily to get feedback on the approach. The main thing to look at is dataset_strategy.py which defines a hypothesis strategy for creating datasets.
The checks I added in reader.py might only be necessary because of deficiencies in InMemoryDataset, and we might be able to remove them as the implementation improves.
Coverage increased (+13.6%) to 85.337% when pulling cbaab7c12de3250cb3d216972ddaf89f79708cc3 on abingham:reader-tests into 5d1391df615c26f6a8a17baa042c1706a619d7bb on sixty-north:master.
Coverage increased (+13.6%) to 85.337% when pulling b4eb7ecaf8c9b2b85dba70cb73791ec2828ca9f8 on abingham:reader-tests into 5d1391df615c26f6a8a17baa042c1706a619d7bb on sixty-north:master.
This PR is primarily to get feedback on the approach. The main thing to look at is
dataset_strategy.py
which defines a hypothesis strategy for creating datasets.The checks I added in
reader.py
might only be necessary because of deficiencies inInMemoryDataset
, and we might be able to remove them as the implementation improves.