change quant_factor to match MATLAB implementation

[schuenke]

Change quantization factor in _compressshape.py from 1e-8 to the MATLAB default value 1e-7. This solves issue https://github.com/imr-framework/pypulseq/issues/42.