Obsobj

[bd-j]

Work to do includes

• [x] Convert to/from old style observation dictionaries
• [x] Put responsibility for Noise Models including outliers in individual Observation instances
• [x] Make predictions even when there is no fittable data (e.g. model spectra when fitting photometry only)
• [x] Store Observation objects in HDF5, FITS, etc as structured arrays with metadata
• [ ] Test multi-spectra calibration
• [ ] Test multi-spectral smoothing
• [ ] Test multi-spectra noise modeling
• [x] Update demo scripts
• [x] Update docs
• [ ] Update notebooks
• [ ] Handle emission line marginalization if spectra overlap
• [x] Structured ndarray for output chains and lnlikehoods

[bd-j]

ok, main thing now is to do some testing of the multi-spec capabilities, then this should be ready to merge into v2.0

[bd-j]

It would be good to add another Observation subclass for integrated emission lines.