refactor: QueryContext and QueryObject by decoupling and making it testable

[ofekisr]

A preparation task for #16991.

decoupling QueryContext and QueryObject classes from superset so they can be used in unit tests and future usages.

What have done:

1. separating the concerns: holding query parameters, objects creations (using factories), and query processing ( data retrieving) into separate modules and Classes.
2. move utils method to utils module
3. code cleanup

It relies on the current test cases since no logic was added

follows up:

1. query context and query context processor naming refactoring
2. keep refactoring query context processor: add factory, create processor for specific query context and not process them as procedure paradigm and remove the "ping-pong" calls with query_actions.
3. add caching parameters to query context
4. remove duplicates code

[codecov]

Codecov Report

Merging #17344 (4fedf18) into master (ee87b01) will decrease coverage by 0.18%.
The diff coverage is 90.26%.

@@            Coverage Diff             @@
##           master   #17344      +/-   ##
==========================================
- Coverage   77.14%   76.95%   -0.19%     
==========================================
  Files        1036     1040       +4     
  Lines       55759    55892     +133     
  Branches     7628     7628              
==========================================
- Hits        43013    43012       -1     
- Misses      12490    12624     +134     
  Partials      256      256              

Flag	Coverage Δ
hive	?
mysql	81.94% <90.26%> (+0.03%)	⬆️
postgres	81.95% <90.26%> (+0.03%)	⬆️
presto	?
python	82.04% <90.26%> (-0.38%)	⬇️
sqlite	81.63% <90.06%> (+0.03%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
superset/viz.py	57.89% <14.28%> (-0.11%)	⬇️
superset/views/api.py	70.68% <42.85%> (-0.74%)	⬇️
superset/common/query_context.py	82.05% <58.33%> (-8.86%)	⬇️
superset/tasks/async_queries.py	95.77% <87.50%> (-1.20%)	⬇️
superset/charts/query_context_cache_loader.py	90.00% <90.00%> (ø)
superset/common/query_object_factory.py	90.62% <90.62%> (ø)
superset/common/query_context_processor.py	91.19% <91.19%> (ø)
superset/charts/api.py	86.58% <91.35%> (+0.75%)	⬆️
superset/common/query_factory.py	94.59% <94.59%> (ø)
superset/common/query_object.py	92.85% <97.50%> (ø)
... and 17 more

---

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 ee87b01...4fedf18. Read the comment docs.

[betodealmeida]

It relies on the current test cases since no logic was added

Note that this is super risky, because we don't have 100% unit test coverage.

[amitmiran137]

It relies on the current test cases since no logic was added

Note that this is super risky because we don't have 100% unit test coverage.

I agree , but on the other hand if we want to move towards better coverage we need to start refactoring and do separation of concerns

[suddjian]

I think if it's possible to split this change into smaller PRs, that will make it much more reviewable. Also, as the purpose of this PR is to make the code testable, I think it would be appropriate and necessary to add tests as part of the refactor.

And if there are any additional tests that cover parts of this, would you mind linking to them? As a reviewer I find tests very helpful for understanding the code.