Fix index creation in Tortoise.generate_schemas() for MySQL and Postgres

[henadzit]

Description

• Replace CREATE INDEX IF NOT EXISTS queries with indexes defined inside the table definition for MySQL
• Fix partial index creation for Postgres - CREATE INDEX "idx_index_partial_c5be6a" ON "index" USING ("partial") WHERE id = 1; isn't correct syntax because USING should be followed by the index type.
• Replace the logic for index creation from Index and its subclasses with schema_generator, hence making schema_generator a single place to have schema generation logic. This also also makes the Index and schema_generator relation one directional.
• Fix the issue with using Model.describe when Index is in Model.Meta.indexes

Motivation and Context

This should fix #1836.

How Has This Been Tested?

make ci

I also tested the generated SQLs against corresponding databases.

Checklist:

• My code follows the code style of this project.
• My change requires a change to the documentation.
• I have updated the documentation accordingly.
• I have added the changelog accordingly.
• I have read the CONTRIBUTING document.
• I have added tests to cover my changes.
• All new and existing tests passed.

[codspeed-hq]

CodSpeed Performance Report

Merging #1847 will not alter performance

_{Comparing henadzit:fix/mysql-and-postgres-idx (d4f7c11) with develop (e34c2c0)}

Summary

✅ 8 untouched benchmarks

[coveralls]

Pull Request Test Coverage Report for Build 12751398752

Details

• 42 of 48 (87.5%) changed or added relevant lines in 8 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.006%) to 90.258%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
tortoise/backends/oracle/schema_generator.py	2	3	66.67%
tortoise/indexes.py	5	6	83.33%
tortoise/models.py	5	6	83.33%
tortoise/backends/base/schema_generator.py	14	17	82.35%

Totals
Change from base Build 12714096970:	0.006%
Covered Lines:	6441
Relevant Lines:	7019

---

💛 - Coveralls

[abondar]

@waketzheng what do you think about this approach?

Even though it moves more code - I think end result seems a bit clearer, as we don't introduce new entities and improve existing code by moving out dialect-specific schema generation code to it's intended place

[waketzheng]

Looks good.

@henadzit Cloud you add more test cases? For example, add class Meta: Index(fields=['name']) to the Node class in tests/testmodels.py

[henadzit]

Looks good.

@henadzit Cloud you add more test cases? For example, add class Meta: Index(fields=['name']) to the Node class in tests/testmodels.py

Will do!

[henadzit]

I added tests and fixed the issue with using Model.describe when an Index is in Model.Meta.indexes.