PulseSequence/VirtualAWG: Improvements

[fr34q]

• Shortening of waveform if readout pulse is at the end should happen in load function of AWG (as it is not relevant without hardware where you want to save memory) @AlexStehli
• PulseSequence(call) adds 0 at the end and front of the pulse, but they are originally not part of the sequence, so this should happen during AWG load and/or in the plot command, where necessary (probably best solution: optional parameter zero_padding = False which can enable it) @fr34q
• Move load() function of VirtualAWG to the real Tabor AWG -> it is device specific and VirtualAWG shouldn't need to have dependencies to all possible hardware devices. Instead make something like tawg.load_pulses(VirtualAWG class) @TimWolz @AlexStehli
• Add PulseSequence.sync() command that marks the position of global phase = 0 and if multiple TdChannels exist, will be the alignment point in the future (otherwise, sequences will be aligned from the start) @fr34q
• Change Pulse.phase to be in radian instead of degree (stay consistent) + adapt in SequenceLibrary @fr34q

Future stuff:

• Make some nice save/load function for VirtualAWG objects to persist them onto the hard disk

You can use branch feature/improve-pulse-sequence linked to PullRequest #75