align/match peaks of two `Spectrum` objects

[tnaake]

Besides the possibility to use binning to match spectra, another possibility to align/match peaks of two Spectrum objects could be the following using bipartitte networks:

1. start with a bipartite network with all possible edges between Spetrum 1 and 2
2. remove edges between peaks (1-5, a-f) that fall not within the given ppm range spanned by two peaks of Spectrum 1 and 2. Peaks 1-5 and a-f are ordered according to ascending m/z values.
3. remove edges between peaks (1-5, a-f) that violate the condition of m/z order maintenance (this step reduces the number of combinations in the next step)
4. from the resulting network, obtain all combinations of alignment/matching
5. calculate the similarity score using a specified metric (e.g. NDP) and select the alignment/matching that optimizes the score, calculate average from the different combinations or similar

required input: Spectrum object 1, Spectrum object 2, ppm value (numeric), metric (character or a function)

see also Watrous et al. (2012), Construction of Molecular MS/MS Networks in the Methods section.