Refactor image formatting, and allow supporting new image formats in a backward compatible way

[MabezDev]

cc: #701

This is quite a complex task, and we'll likely need to iterate a bit with the design here. Currently image formatting, is tightly coupled to the targets, which is then coupled to the flasher module. We need to find a way for elf inputs to be turned into the image format in a more self contained way.

In the meeting, we discussed some kind of API/Abstraction to take CodeSegments as an input, and then depending on the selected image format create an appropriate output. This was just a rough outline, and there is a lot of investigation and iteration needed to come up with a final solution here.

[jessebraham]

It's still a bit rough, but I have pushed a branch making the changes discussed between myself, @MabezDev, and @SergioGasquez:
https://github.com/jessebraham/espflash/tree/feature/image-format

There are a few areas which could probably use some improvement, but the biggest concern right now for me is the Esp32Params struct. This struct contains a number of ESP-IDF-specific fields, and we should decide how to move forward with this. We could potentially keep the fields and just zero/ignore them for other formats, otherwise we will need to add these for each chip within the esp_idf module or something.