fix: handle compare_vals turn into str when parse IgnoreColumnType

[phakawatfong]

when I parse ignoreColumnType=True when call the function, it cast all values to be 'str' which causing the comparison between val1 and val2 to be failed.
for example

val1 = 1505.76189560
val2= 1505.761896
assertDataFrameEqual(source_df, databricks_df, ignoreColumnType=True,)

I will convert the original data into string with this clause first

if ignoreColumnType:
    actual = cast_columns_to_string(actual)
    expected = cast_columns_to_string(expected)
.
.
compare_vals(val1, val2)

it will become string comparison instead of compare by using the tolerance equation
'1505.76189560' != '1505.761896'

Hence assertError, even if the atol and rtol already assigned

What changes were proposed in this pull request?

add more elif case to convert str to float and parse into tolerant calculations

Why are the changes needed?

I think it might not be the only way to fix it, but just to be your ideas Thank you !

Does this PR introduce any user-facing change?

No

How was this patch tested?

Test on my local data source, comparing 2 dataframes
by patch the code into my local env.

Was this patch authored or co-authored using generative AI tooling?

No