IntField: add MinMaxValidator

[markus-96]

Description

Before sending anything to the DB, IntField values are checked beforehand now. As far as I could test, the added overhead is not much, lets see what Codspeed says.

Will add documentation later.

Motivation and Context

Using sqlite, you could send values greater than defined in IntField.constraints.get('le') to the DB and nothing was raised. If you for example want to migrate to another DBMS, you have a problem.

Also, have a look at the Issue I raised: #1853

How Has This Been Tested?

Existent tests and new tests

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 #1868 will not alter performance

_{Comparing markus-96:intfield-validator (8f317cf) with develop (960b1c1)}

Summary

✅ 12 untouched benchmarks

[coveralls]

Pull Request Test Coverage Report for Build 13053921297

Details

• 22 of 28 (78.57%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage decreased (-0.09%) to 89.146%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
tortoise/fields/data.py	20	26	76.92%

Totals
Change from base Build 13012319268:	-0.09%
Covered Lines:	6500
Relevant Lines:	7105

---

💛 - Coveralls

[henadzit]

I think we should be adding a validator instead of customizing to_db_values similarly how it is done in CharField

[markus-96]

that's what I did first, but it adds far too much overhead decreasing the benchmarks like 10% (see the edit history of the comment of codspeed..)

I think that's because it will actually enter the for each loop in self.validate instead of doing nothing. Also I think that overriding to_do_value actually increases speed because the interpreter does not have to go that much up the MRO.

[henadzit]

Extending the field configuration with ge and le introduces the second way of setting up validation. I think we should have a single way of doing it to keep things simple - through validators. If some one needs specific validation, they can add a validator. So I suggest to remove these options from the constructor.

[markus-96]

ge and le are currently baked in in the constraints property. Instead of that I wanted them to be customizable, a bit like with max_char in CharField where this is reflected during validation and also by the constraints property.

[henadzit]

I don't think this is correct.

[markus-96]

If you have like ten validators in self.validators, you will always have to perform the check whether the field may be null and if the given value is None/null. Both informations do not change during itering other the validators. So instead of performing the same check over and over again, we can instead directly say that nothing else will happen here and that the value is valid.