Fix direct chapter links for searches containing text highlights

[cblanken]

With the current link creation, some generated links are invalid because they surround a <span> tag or parts of it get included in the URL.

For example

Uses this link https://wanderinginn.com/2017/06/21/2-31/#:~:text=%E2%80%9CMagnolia%2C%20please%2C%20Erin.%20%3Cspan which doesn't properly link to the text fragment since it should only include the text from the source page.

[cblanken]

It should be safe to pass the TextRef text through a BeautifulSoup parser and grab the text.

If the parse fails then the the text can be iteratively reduced until a valid chunk is found for the link or simply default to the chapter link without the fragment cruft if a valid parse can't be found.