Attempts at doing something to extract individual snippets from JesseWills FA extracted portion

jesse_wills_extract.xq

xquery version "3.1";

let $xml := fn:doc("https://raw.githubusercontent.com/HeardLibrary/finding-aids/master/ASpace/Working%20EADs/WillsJesseEly_MSS_0001_working_pieces.xml")

let $date := fn:data($xml//unitdate)
let $title := fn:data($xml//unittitle)

(: This was just something I was playing around with, trying to figure things out.:)
(: let $new-title := fn:string-join($title, $date) :)

let $string-title := fn:string-join($title, ' ')
let $new-title := fn:tokenize($string-title, ';')

return $string-title

(: What I want the code to do:
Take from here:
....<unitdate>1958</unitdate>
<unittitle>30 April to Allen Tate; 9 June to Allen Tate; 9 June to . "Bill"; (William Yandell Elliott); F. 1-1. 
17 December to &#34;Lib&#34; (Mrs. R.D. Crabtree); F. 3-4.</unittitle>....

Give me something like below by creating a new tagged piece every time a ';' is encountered.
<new_unittitle>30 April to Allen Tate</new_unittitle>
<new_unittitle>9 June to Allen Tate<new_unittitle>
etc. etc.

Preferably, I could eventually get it to make this:
<c02><did>
<unittitle>To Allen Tate</unittitle>
<unitdate>1958 April 30</unitdate>
</did></c02>

<c02><did>
<unittitle>To Allen Tate</unittitle>
<unitdate>1958 June 9</unitdate>
</did></c02>
etc. etc.

FYI, 'F. #1-#2' means that everything preceding it should actually be Box #1, Folder #2. This info should 
be in a container tag for each <c02>, but I may just add by hand if it's not consistent enough in the original
<unittitles> to be extracted easily.
:)