A trap is simply a resonant circuit. You can measure it like any resonant circuit. But since traps are build to handle significant power, they are usually somewhat large, which makes them more sensitive to their surroundings. So a trap connected to the NanoVNA will probably measure a little different than its actual behavior when installed in the antenna.
The important curve of a trap is that of its reactance over the frequency range of interest. Traps are usually parallel resonant circuits, so the reactance will start low and inductive at low frequency, increase with frequency, reach a very high value at its resonant frequency, wrap over to a very large value of capacitive reactance, and this capacitive reactance will diminish with further increasing frequency. The important things are what's the resonant frequency, and how steeply the reactances vary with frequency.
There are basically two ways to use traps in antennas. The simpler one is to make the trap resonant in the middle of the higher band, then install it in a position so that the inside leg of the antenna resonates on that band too. The trap acts like an insulator on this band. On the lower band the trap has an inductive reactance, and depending how high it is (which depends on the L/C ratio of the trap), the outer leg needs to have a specific length to make the antenna resonate on that lower band. The total length of the antenna is always shorter than a trapless antenna for the lower band.
This method is easy to understand and tune, but the resonance on the higher band is pretty narrow, and the trap has to stand up against a very high voltage on that band. So another method can be used, that places lower stress on the trap and also results in a better bandwidth: The trap is made to resonate in between the two bands. On the lower band it will still be inductive, so the total antenna length is still shorter than a trapless antenna for that band, but on the higher band this trap is capacitive, electrically shortening the antenna. So the entire antenna works on both bands, the antenna is of a length between that of of trapless antennas for the lower and the higher band, the trap sits in a location where the voltage is moderate on both bands, so it's less lossy and can be smaller, and the trap works on its reactance slopes on both bands, rather than working on the resonance peak on the higher band. This makes the antenna more broadband than in the other case, on the higher band.
So the second method is clearly the better one, but it's a little harder to calculate and tune properly, because the trap's resonant frequency, its L/C ratio, the total antenna length, and the position of the trap, all interact and all have an effect.
One way to do is is to decide the antenna length first. For example for center frequencies of 3.7 and 7.1MHz, 15 meters would be a good length for a quarter wave element with this kind of trap. Then decide the trap position. I would try to put the trap in the middle of the element, or even closer to the feedpoint. Then calculate or measure what inductance you need there to resonate the antenna on 3.7MHz, and what capacitance you need for a 7.1MHz resonance. Then calculate the inductance and capacitance you need in a resonant circuit to get them. The ratio between them will also determine the resonant frequency of the trap. Build the trap, measure it with the NanoVNA , adjust it, then install it in the antenna and either pray or swear, depending on your preference. Then start the usual game of tweaking the antenna until it works fine on both bands.
To do the actual measurent, you don't need to bother with both ports. It's enough to connect the trap to the main port. I suggest that you connect cable pigtails to the NanoVNA, long enough to comfortably reach the ends of the trap, while the trap isn't too close to the NanoVNA. Calibrate the NanoVNA with these pigtails, arranging them roughly in a loop for the short and load tests, and roughly in the positions they will be when connected to the trap, for the open calibration. Then connect the trap and measure the reactance over the desired frequency range, as far away as possible from other stuff, specially conductive one.
