set pfiCenter and fiberStatus correctly when writing pfsConfig.

Looking at the cobra_match entries, I've noticed that I did some mistakes in making pfsConfig back in november.
I was getting 2394 matches each time which is unlikely, because the pfi_center_(x,y)_mm. are set to the theorical black dot position if there is no match, I don't like that.

• I'll put NaN where there are no measured mm position.
• I will set fiberStatus to BROKEN where FIBER_BROKEN_MASK.
• I will set fiberStatus to BLACKSPOT where pfi.status==OK and where there is no match.

I wonder even on fiberId 3, which is labelled as broken even if it's not broken-broken it's just low throughput eg I have a match with mcs and I have a spectra on sps.