add max_precision to onnx parser

[jmitrevs]

Description

This is analogous to PR #1112 but for for ONNX. For ONNX, auto was the default, but max_precision was not set. This copies what we have for the Keras parser.

Type of change

• [x] New feature (non-breaking change which adds functionality)

Tests

We should make sure that this doesn't break anything, but truthfully, I am not sure if we have tests to test the max_precision behavior for any frontend, including Keras. Should we add such tests?

Checklist

• [x] I have read the guidelines for contributing.
• [x] I have commented my code, particularly in hard-to-understand areas.
• [ ] I have made corresponding changes to the documentation. (But some exists in PR #1100)
• [x] My changes generate no new warnings.
• [x] I have installed and run pre-commit on the files I edited or added.
• [ ] I have added tests that prove my fix is effective or that my feature works.