A caveat is that the terms a priori and a posteriori are open to debate, and knowledge and justification pertaining to them is controversial. See A Priori Justification and Knowledge (SEP). However, responding through the lens through mathematical intuitionism (SEP):
You ask:
Is any fictional knowledge purely a priori knowledge?
With the exception of mathematical knowledge that is independent of language (let's say, the recognition of a triangle by a baby before the acquisition of language), no. Knowledge of fictional matters in both the lay sense (fictional narratives, for instance) and the technical sense on par similar to nominalism, aka fictionalism (SEP), is largely a posteriori. This is because both forms of fiction are an expression of the language which is largely a social convention that must be learned with experience to be used. Two examples will serve our needs to illustrate a response framed in both theses from the SEP article above:
The linguistic thesis is, roughly, that already expressed above, according to which utterances of sentences of the discourse are best seen not as efforts to say what is literally true, but as useful fictions of some sort. The ontological thesis, by contrast, is the thesis that the entities characteristic of the discourse do not exist, or have the ontological status of fictional entities.
First, consider Lord of the Rings. There is no doubt it is a work of fiction in the lay sense. Gandalf is not real. Sauron is not real. Middle-Earth is not real. And despite you can read it in dozens of languages, more languages and more readers do not make it real. And the story did not spring into existence by the powers of your mind (though obviously some a priori faculties are required to make sense of language in the general case). This work of fiction must be learned, and it must be learned after the language it is expressed in is learned.
Second, consider mathematical fictionalism (SEP), the technical sense of fiction. If you were to grant this as a respectable philosophy of mathematics (and certainly not everyone does), you would certainly need in a broadly Kantian understanding a priori abilities to make sense of the mathematics. No doubt. However, are some form of Peano's axioms learned? Absolutely. And so is the mathematical notation. And if you wanted to engage in a debate over the proper formulation of the basis of set theory, is it self-evident that NBG is superior to ZFC, or that any of these axiomatic systems are preferable to any other? If it were self-evident, extensive study wouldn't be required to learn the language and the ontology of graduate level mathematics.
Remember what it means to be a priori. From the WP article:
A priori knowledge is independent from any experience. Examples include mathematics,[i] tautologies and deduction from pure reason.[ii] A posteriori knowledge depends on empirical evidence. (emphasis mine)
Any, of course, is a difficult bar to clear, so it might be better to think in tempered terms when dealing with the term a priori
