Adding EXPRES interface

[Kamuish]

• S2D spectrum
• sub-Instruments
• KWs of interest
• Observatory properties for telfit
• Load telluric mask
• Avoid multiple telluric removal applications

[Kamuish]

Missing KWs:

• BERV (achromatic)
• water content + ambient temperature (for telfit)

No need of implementing telfit for now. We can live with SELENITE, instead opening new issue to enhance the EXPRESS implementation

[Kamuish]

The SELENITE telluric model has negative values , which leads to numerical issues (near zero) when trying to divide \frac{spectrum}{model}

[Kamuish]

The SELENITE model is for removal, not correction... Not a good idea to remove different wavelength regions in each observation, selecting worst night of all to use as "reference" (similarly to current SBART approaches to tellurics). Needs to allow bypassing the creation of transmittance spectrum in the two templates

[Kamuish]

First skeleton implementation is complete.
To be tested:

• Data loading of multiple observations + CCF access checks
• Test creation of Telluric Template
• Apply RV extractions

Important note 1: Due to the spectrum being normalized we will, most likely, run into numerical issues.
Important note 2: Due to the spectrum being normalized the Stellar Template creation will break

[Kamuish]

From the docs, we might be able to retrieve original counts:

blaze gives the counts of the median flat exposure that the extractions were done with respect to. Multiplying the spectrum by this blaze will return the original counts for each pixel. These counts, which are presumed Poisson distributed, provide the best measure of SNR for the exposure. Note that because this is simply a median flat exposure, it may contain some detector defects that would otherwise be flat-fielded out in the extracted exposures through the flat-relative extraction. Users are encouraged to fit a smooth function to this blaze before multiplying by the extracted spectrum to recover true counts.