I suggest you have a look at this tutorial which describes in detail how to create a merged transcriptome containing both the spliced and unspliced sequences of each annotated transcript. You can then use it with tools like
salmon to quantify your reads against. This is what one does in scRNA-seq to get the input for RNA velocity (in this case with the scRNA-seq-tailored module
alevin from the
salmon software) but the idea should apply equally well to bulk RNA-seq data in terms of getting spliced and unspliced transcript quantifications: combine-lab.github.io/alevin-tutorial/2020/alevin-velocity/
This would give you transcript abundance estimates (like counts basically) for each spliced and/or unspliced transcript.