Overall improvement of the storefront 2.0 performances

[NyanKiyoshi]

This fixes #3631.

Changes

Optimization of Single Nodes

This pull enables optimizations for single nodes. Which optimizes a lot of requests. See tfoxy/graphene-django-optimizer#18 for more information about this.

Overriding the Retrieval of a Single Node

Following the tfoxy/graphene-django-optimizer#18 changes, getting a single node whether optimized or not is now simpler. You should always call maybe_optimize(info, qs, pk) and not directly call get_optimized_node(info, qs, pk) when overriding get_node(info, pk) as it would result to issues when the node is not supposed to actually get optimized.

This sums up of doing this:

During The Payment Capture

• Prefetch order lines that then get proceeded;
• Prefetch the user that will later get used to get the email address.

Comparison

Page	Before	After
Product Listing	29 queries	11 queries
Product Details	40 queries	14 queries
Checkout Shipping Method Update	82 queries	38 queries
Placing an Order	135 queries	64 queries
...	...	...

Pull Request Checklist

1. Privileged views and APIs are guarded by proper permission checks.
2. All visible strings are translated with proper context.
3. All data-formatting is locale-aware (dates, numbers, and so on).
4. Database queries are optimized and the number of queries is constant.
5. Database migration files are up to date.
6. The changes are tested.
7. The code is documented (docstrings, project documentation).
8. GraphQL schema and type definitions are up to date.
9. Changes are mentioned in the changelog.

[NyanKiyoshi]

There are bugs/issues around the optimizer library, so this PR will sadly (brings a lot of great optimizations) not be ready until the root of this issue is figured out. Hopefully I can dig more tomorrow around the troubles (relates to saleor/saleor-storefront#295).

[NyanKiyoshi]

@auvipy please don't review drafts, that generating useless notifications on my side (thus making me lose precious time). Just leave a little 👍 reaction on the PR if you like it, that's enough to say you approve it!

[NyanKiyoshi]

I added from_global_id_strict_type that has the aim to get easily a node from a given ID while using a custom queryset instead of using a simple .get(pk) that doesn't prefetch anything.

Let me know what you think, or if you see a better approach to this matter.

[codecov]

Codecov Report

Merging #3968 into master will decrease coverage by 0.05%.
The diff coverage is 91.3%.

@@            Coverage Diff             @@
##           master    #3968      +/-   ##
==========================================
- Coverage   91.37%   91.32%   -0.06%     
==========================================
  Files         274      274              
  Lines       14947    14950       +3     
  Branches     1451     1452       +1     
==========================================
- Hits        13658    13653       -5     
- Misses        898      904       +6     
- Partials      391      393       +2

Impacted Files	Coverage Δ
saleor/graphql/payment/mutations.py	97.7% <100%> (+0.02%)	⬆️
saleor/graphql/checkout/mutations.py	93.47% <100%> (+0.04%)	⬆️
saleor/graphql/product/types/products.py	93.1% <100%> (-2.19%)	⬇️
saleor/dashboard/order/views.py	79.25% <100%> (ø)	⬆️
saleor/graphql/core/connection.py	100% <100%> (ø)	⬆️
saleor/graphql/core/utils.py	91.3% <71.42%> (-8.7%)	⬇️

---

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 b83a0eb...70e7aa5. Read the comment docs.

[maarcingebala]

The PR looks good, but I'm only wondering about the from_global_id_strict_type function. I'm not sure if this should be a part of BaseMutation since it's only used in two places and is meant rather for specific use cases. I guess it should be a utility function somewhere in "saleor/graphql/core".