Implement separator validation

[MichaelS239]

Add methods for separator validation in .csv tables

[vs9h]

Let's use path from already existing csv configs
And we don't need constructor

Suggested change

	struct TestSeparatorValidationParam {
	std::filesystem::path csv_path;
	char separator;
	char expected;
	TestSeparatorValidationParam(std::filesystem::path const& path, char sep, char exp)
	: csv_path(path), separator(sep), expected(exp) {}
	};
	struct TestSeparatorValidationParam {
	CSVConfig csv_config;
	char test_separator;
	std::optopnal<char> expected_separator;
	};

[vs9h]

I think, that we also should add some complex cases to derive the separator. Something like:

;,;,;
;,;,;

In this case valid separators are ; and ,.
Also check validation for csvs with quotes:

'a..b',c,d
e,d,'f..g'

In this case, only ',' is a valid separator

[vs9h]

and so we can write here

Suggested change

	EXPECT_EQ(actual.value_or('\0'), p.expected);
	EXPECT_EQ(actual, p.expected_separator);

[vs9h]

In this case we will also take into account those characters that we do not want to take into account - the chars in quotes. See second example in the last comment in the file src/tests/test_util.cpp

[vs9h]

We need to change the function signature. We shouldn't just write a warning to the console, we need to explicitly return a message with error information and actual separator, if found.
We can return something like: std::pair<std::optional<char>, std::string>.
This is necessary because, first of all, this function will be needed for the python server. Let's say we downloaded a user's dataset and want to find out whether he sent the correct separator. And the main goal is to provide readable information about the error.
And accordingly, you need to figure out how to bind this function to python bindings

[vs9h]

Also, I'm not sure that we always want to validate the separator when creating a table. The main problem is that we have to read the table completely before running the algorithm itself. This causes the file system caches and disk cache to fill up. Therefore, if someone tries to conduct experiments and does not know that validation always takes place first, then the results of the experiments will be incorrect (since all caches must be reset before running the test).