Tokenizer doesn't allow no ignore characters

Problem/Motivation

The Tokenizer processor has an input for Ignored characters which may be adjusted to suit a site's particular content. If the input is empty then instead of no characters being ignored, the default ignore characters ._- are removed. Although consistent with the behaviour of the Whitespace characters it is unexpected behaviour based on the help text.
For my particular use case I don't want to use any of the default ignore characters, so that the filenames interesting-document.pdf and interesting_document.pdf are still searchable (ie, they don't both become interestingdocumentpdf). See also related issue 2965995 for another user's similar request.

Steps to reproduce

1. Add file to site called interesting-document.pdf.
2. Default install of search_api.
3. Add index on File datasource.
4. Add the Tokenizer Processor.
5. Delete all content from the Ignored characters input box.
6. Index content
7. Search for 'interesting'. The file interesting-document.pdf is not in the results.

Proposed resolution

Allow users to have no ignore characters by updating the prepare function in web/modules/contrib/search_api/src/Plugin/search_api/processor/Tokenizer.php (the else branch below):

    if (!isset($this->ignored)) {
      if ($this->configuration['ignored'] !== '') {
        $this->ignored = str_replace('/', '\/', $this->configuration['ignored']);
      }
      else {
        $this->ignored = '._-';
      }
    }