Adding UUID support to hashing

[szoupanos]

Fixes issue #1771
UUID is not hashed properly under SQLA and this is fixed by this PR.

[sphuber]

Maybe we should add a test? If this was broken but was not caught by the current tests

[codecov-io]

Codecov Report

Merging #1861 into release_v0.12.2 will increase coverage by <.01%.
The diff coverage is 100%.

@@                 Coverage Diff                 @@
##           release_v0.12.2    #1861      +/-   ##
===================================================
+ Coverage            54.48%   54.48%   +<.01%     
===================================================
  Files                  246      246              
  Lines                32438    32441       +3     
===================================================
+ Hits                 17674    17677       +3     
  Misses               14764    14764

Impacted Files	Coverage Δ
aiida/common/hashing.py	84.37% <100%> (+0.5%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 98602cf...cdeb613. Read the comment docs.

[szoupanos]

Tests were added. The problem was a corner case: When the UUID is asked as a result from the QueryBuilder, the UUID is returned as a uuid.UUID type in SQLA.

When this results is passed in another QB query as a filtering argument, the query fails because hashing fails (it doesn't support UUID hashing).

I fixed hashing to work also in the case that a uuid.UUID argument is passed.
I believe that QB should return a unicode UUID as the Node.uuid method does.
(and the behaviour should be uniform in both backends)
But I leave this as a separate issue for @lekah

[lekah]

I'm OK with this as it is. @sphuber if you judge your requested changes enough, go ahead and merge.