In Old English, thou was used for addressing one person and ye for more than one, both as clause subject. Thee and you were used as object.
During the Middle English period, ye/you came to be used as a polite singular form alongside thou/thee.
During Early Modern English, the distinction between subject and object uses of ye and you gradually disappeared. Ye continued in use, but by the end of the 16th century it was restricted to archaic, religioous, or literary contexts. By 1700, the thou forms were also largely restricted in this way.
(Adapted from ‘The Cambridge Encyclopedia of the English Language’ by David Crystal.)
Wert is more likely to be subjunctive
In the King James Bible, as well as in some other texts, wert appears only in "subjunctive" contexts. Using the online KJV bible search engines at King James Bible Online and the University of Michigan Bible: King James Version -- Simple Searches shows that wert always occurs in a construction that licenses the subjunctive mood such as an if clause.
However, there is not always a difference: wert can also be found as an indicative form. Be in its history shows a multitude of forms that aren’t all strictly used for separate functions.
The forms wast and wert are both exceptions to the general pattern for the occurrence of the second-person singular suffix -st, which usually occurred only in the indicative mood of the present tense (for all verbs), and in the indicative mood of the past tense (for weak verbs: verbs with a dental suffix in the past tense). The subjunctive mood is regularly formed without -st in the second person singular, just as it is formed without -s or -th in the third person singular.
Therefore, the expected second-person singular past subjunctive form of be is thou were; and because were is not a weak verb, the Old English second-person singular past indicative did not have -st either.
The forms wast and wert come from irregular addition of the -(s)t suffix after the Old English period.
The Oxford English Dictionary says
As is usual for strong verbs, the 2nd singular past indicative in Old English has the same form as the singular past subjunctive (wǣre , Anglian wēre ; compare Old Saxon 2nd singular past indicative wāri ). By late Middle English, the form has become identical with the plural past indicative and subjunctive as well, except in varieties where there is levelling from the 3rd singular to the 2nd singular past indicative (chiefly in the north and the north midlands). Attempts to mark the form more clearly by attaching distinctive 2nd singular endings (either –st , the ending of the present indicative and the (weak) past indicative of regular verbs, or –t , the ending of preterite-present verbs (e.g. shalt ) and of present indicative art ) result (respectively) in Middle English wast , early modern English werst , and Middle English wert , all of which are continued in English regional varieties. Due to the general obsolescence of the distinctly marked 2nd singular in standard English, no canonical form for the category was established, with both wast and wert surviving in archaic and poetic use.
The form wert showed overlap both with indicative wast and with subjunctive were (and similarly, in the present, the form beest showed overlap with both indicative art and subjunctive be).
Goold Brown's "Grammar of English Grammars" from 1882, provides the following observations about the usage of wert, noting that some grammarians prescribed it as a subjunctive form in opposition to indicative wast, but that usage did not necessarily conform to that prescription:
(page 375):
OBS. 2.--Respecting the verb wert, it is not easy to determine whether it is most properly of the indicative mood only, or of the subjunctive mood only, or of both, or of neither. The regular and analogical form for the indicative, is "Thou wast;" and for the subjunctive, "If thou were." Brightland exhibits, "I was or were, Thou wast or wert, He was or were," without distinction of mood, for the three persons singular; and, for the plural, were only. Dr. Johnson gives us, for the indicative, "Thou wast, or wert;" with the remark, "Wert is properly of the conjunctive mood, and ought not to be used in the indicative."--Johnson's Gram., p. 8. In his conjunctive (or subjunctive) mood, he has, "Thou beest," and "Thou wert." So Milton wrote, "If thou beest he."--P. Lost, B. i, l. 84. Likewise Shakspeare: "If thou beest Stephano."--Tempest. This inflection of be is obsolete: all now say, "If thou be." But wert is still in use, to some extent, for both moods; being generally placed by the grammarians in the subjunctive only, but much oftener written for the indicative: as, "Whate'er thou art or wert."--Byron's Harold, Canto iv, st. 115. "O thou that wert so happy!"--Ib., st. 109. "Vainly wert thou wed."--Ib., st. 169.
OBS. 3.--Dr. Lowth gave to this verb, BE, that form of the subjunctive mood, which it now has in most of our grammars; appending to it the following examples and questions: "'Before the sun, Before the Heavens, thou wert.'--Milton. 'Remember what thou wert.'--Dryden. 'I knew thou wert not slow to hear.'--Addison. 'Thou who of old wert sent to Israel's court.'--Prior. 'All this thou wert.'--Pope. 'Thou, Stella, wert no longer young.'--Swift. Shall we, in deference to these great authorities," asks the Doctor, "allow wert to be the same with wast, and common to the indicative and [the] subjunctive mood? or rather abide by the practice of our best ancient writers; the propriety of the language, which requires, as far as may be, distinct forms, for different moods; and the analogy of formation in each mood; I was, thou wast; I were, thou wert? all which conspire to make wert peculiar to the subjunctive mood."--Lowth's Gram., p. 37; Churchill's, p. 251. I have before shown, that several of the "best ancient writers" did not inflect the verb were, but wrote "thou were;" and, surely, "the analogy of formation," requires that the subjunctive be not inflected. Hence "the propriety which requires distinct forms," requires not wert, in either mood. Why then should we make this contraction of the old indicative form werest, a solitary exception, by fixing it in the subjunctive only, and that in opposition to the best authorities that ever used it? It is worthier to take rank with its kindred beest, and be called an archaism.
OBS. 4.--The chief characteristical difference between the indicative and the subjunctive mood, is, that in the latter the verb is not inflected at all, in the different persons: IND. "Thou magnifiest his work." SUBJ. "Remember that thou magnify his work."--Job, xxxvi, 24. IND. "He cuts off, shuts up, and gathers together." SUBJ. "If he cut off, and shut up, or gather together, then who can hinder him?"--Job, xl, 10. There is also a difference of meaning. The Indicative, "If he was," admits the fact; the Subjunctive, "If he were," supposes that he was not. These moods may therefore be distinguished by the sense, even when their forms are alike: as, "Though it thundered, it did not rain."--"Though it thundered, he would not hear it." The indicative assumption here is, "Though it did thunder," or, "Though there was thunder;" the subjunctive, "Though it should thunder," or, "Though there were thunder." These senses are clearly different. Writers however are continually confounding these moods; some in one way, some in an other. [...]
Best Answer
Yes, "thou (...) who art" or "thee (...) who art" are correct.
I wasn't sure from the title whether you were asking about relative pronouns or interrogative pronouns, so I will discuss both in my post.
In the hymn that you quote, the relative pronoun "which" takes second-person singular agreement because its antecedent is the second-person singular pronoun "thee". This is a special thing that happened/happens in old-fashioned or formal English' related questions about this topic are What rules make “Remember me, who am your friend” grammatical? and "Me who is" or "me who am"?
In terms of interrogatives, "Who art..." would be possible in a sentence with "thou": "Who art thou?" In modern English, sentences of this type (e.g. "Who am I?") are best analyzed as having "who" as the (fronted) predicate rather than as the subject: a piece of evidence that "I" and not "who" is the subject of "Who am I?" is that we can't say *"Who am me", even though in predicate position "me" is usually possible (we can say “It was me,” regardless of whether it’s considered “incorrect” from a prescriptive point of view). In older varieties of English, I'm not sure whether there is any clear way of establishing which word is the subject in questions like this.