In the standard there is the difference between a complete modulation (V.ii7 or ii.V7) and a temporary change of the tonic (V/ii ii). In my opinion, that is a very good differentiation. However in some passages of temporary changes sometimes there are secondary dominants which can not be annotated in a right way.
Here is a example from Mendelssohn String Quartet op 12 No 1 E flat major:
Here the viio7/VII and V65/VII is not really related to VII. I like to suggest for this cases an annotation in this way: viio7/V/iii V65/V/iii
In the standard there is the difference between a complete modulation (V.ii7 or ii.V7) and a temporary change of the tonic (V/ii ii). In my opinion, that is a very good differentiation. However in some passages of temporary changes sometimes there are secondary dominants which can not be annotated in a right way.
Here is a example from Mendelssohn String Quartet op 12 No 1 E flat major:
Here the viio7/VII and V65/VII is not really related to VII. I like to suggest for this cases an annotation in this way: viio7/V/iii V65/V/iii