In terms of the idea of raga personificationm we have definite texts containing raga-iconification with dhyana slokas etc. starting around the fourteenth century. This was used as one of the bases of taxonomy and classification of ragas similar to the raga/ragini system. The imagery association works at two levels one being the melodic intervallic aspects and the other was based on the aesthetic import. The latter is rooted in the classical rasa theory where a raga personification works much like producing a rasa with the dominant sthayibhava and other secondary characteristics — props, other figures etc taking shape in raga mala paintings. The full force of such detailed word-pictures of raga iconography is seen in Somanatha, PandarikaVitthala etc. Bhairava (of that time and that text) was thus a male based on a strong gandhara with with a association with Rudra’s paraphernalia. Bhupali is feminine, a ragini seated on a swing waiting for her lover. Thus in terms of musical history, this was one of the ways in which melodies were classified. It also had its supplemental use in pedagogy for visualization by a student and also fit in well with the idea of a formless nada and a raga with form and name (nama / rupa) in Indian religious thought.
Thus we have explanations of the idea of a deity descending when invoked and the raga being grasped by the person who sings in The remnants of this system continued in some form in Hindustani music. It fell out of use in Karnataka sangita where other mechanisms of classifications and categorizations took over. Today, atleast in South India, it is used more in a non-technical evocatory and spiritual / religious sense devoid of these specific textual associations.