Fix: handle tinyint(1) BOOLEAN inference with modifiers like NOT NULL

[WittgensteinBeetle]

Problem

The getFivetranDataType function originally only returned BOOLEAN if the MySQL type was exactly "tinyint(1)". This excluded common real-world variants like:

• tinyint(1) NOT NULL
• tinyint(1) unsigned
• tinyint(1) zerofill
• TINYINT(1) (case insensitivity)

These types should be safely treated as booleans when treatTinyIntAsBoolean is enabled — which matches MySQL connector behavior and expectations from most CDC systems.

---

Fix

The logic has been updated to:

• Normalize mysqlType to lowercase
• Split it by whitespace (using strings.Fields)
• Only evaluate the base type (e.g., "tinyint(1)") for the boolean match

This avoids false negatives while ensuring that tinyint(2+) are not misclassified as booleans.

---

Test Coverage

Added unit tests for:

• tinyint(1) -- Pass
• tinyint(1) NOT NULL -- Pass
• tinyint(1) unsigned -- Pass
• tinyint(2) -- Fail
• tinyint -- Fail
• TINYINT(1) -- Pass

Added an additional unit test to include these parameters as the original did not include these edge cases

[maxenglander]

this looks good, will approve this once tests pass and my comments are addressed

[maxenglander]

can remove this comment and the next line

[maxenglander]

is there any reason these test cases can't fit as cases into the existing TestSchema_CanPickRightFivetranType test rather than defining a new test?

[maxenglander]

nice, can we add a test case for this?