Update internal use of decimal to float (#10333)

# Description
We made the decision that our floating point type should be referred to
as `float` over `decimal`.
Commands were updated by #9979 and #10320

Now make the internal codebase consistent in referring to this data type
as `float`.

Work for #10332

# User-Facing Changes

`decimal` has been removed as a type name/symbol. 

Instead of 
```nushell
def foo [bar: decimal] decimal -> decimal {}
```
use 
```nushell
def foo [bar: float] float -> float {}
```

Potential effect of `SyntaxShape`'s `Display` implementation now also
referring to `float` instead of `decimal`

# Details
- Rename `SyntaxShape::Decimal` to `Float`
- Update `Display for SyntaxShape` to `float`
- Update error message + fn name in dataframe code
- Fix docs in command examples
- Rename tests that are float specific
- Update doccomment on `SyntaxShape`
- Update comment in script

# Tests + Formatting
Updates the names of some tests

crates/nu-std/std/dt.nu

@@ -149,7 +149,7 @@ export def datetime-diff [from: datetime, to: datetime] {
         $result = (borrow-second $from_expanded $result)
     }
 
-    $result.millisecond = ($result.nanosecond / 1_000_000 | into int) # don't want a decimal
+    $result.millisecond = ($result.nanosecond / 1_000_000 | into int) # don't want a float
     $result.microsecond = (($result.nanosecond mod 1_000_000) / 1_000 | into int)
     $result.nanosecond = ($result.nanosecond mod 1_000 | into int)