bin path caching

[joefutrelle]

• caches bin path in the Bin model, along with the DataDirectory it was found in
• uses the cache unless the cache is explicitly invalidated by setting path to the empty string, in which case the next access will look in the originating data directory first before caching

[joefutrelle]

This logic is now problematic:

ifcbdb/ifcbdb/dashboard/accession.py

Lines 91 to 93 in 8b237fb

	if not created and not self.dataset in b.datasets:
	self.dataset.bins.add(b)
	return

because if a dataset is synced and a bin has been removed from that dataset but its files are still in the relevant data directory, it will be re-added to the dataset. I think it can simply be removed, since the rest of the 4.5.0 branch provides it and other features for managing datasets.

The correct logic is now to simply return if the bin already exists.

[mike-kaimika]

This all looks good. I agree with the comment you left about not re-adding the bin back to the dataset

[joefutrelle]

@mike-kaimika do we want to push this to before the public 4.5?

[mike-kaimika]

@joefutrelle I do not have any objections to doing so. The more we can (safely) get merged in this next release, the less there will be to the already significant release we'd be doing for the other functionality