Initial Spitzer/JWST instrument packs

[jdtsmith]

Intended for testing use with PR#222.

[codecov-commenter]

Codecov Report

Merging #226 (168f0aa) into master (7010efa) will increase coverage by 0.06%. The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #226      +/-   ##
==========================================
+ Coverage   43.84%   43.91%   +0.06%     
==========================================
  Files          12       12              
  Lines         821      822       +1     
==========================================
+ Hits          360      361       +1     
  Misses        461      461              

Impacted Files	Coverage Δ
pahfit/errors.py	100.00% <ø> (ø)
pahfit/features/__init__.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 7010efa...168f0aa. Read the comment docs.

[jdtsmith]

Torsten Boeker pointed me to some lovely FITS files for NIRSPEC resolution, which I fitted and included.

Some other changes:

• instrument.resolution() to go with fwhm.
• Use as many polynomial coefficients as needed (jwst.nirspec.prism needing A LOT).
• Remove trailing zero coeffs.

Just waiting for similar MIRI values (@karllark!).

[karllark]

Here is the MIRI info Alvaro Labiano just emailed me. He says ignore the "etalon" column (ground-test info).

[jdtsmith]

Here is the MIRI info Alvaro Labiano just emailed me. He says ignore the "etalon" column (ground-test info).

Awesome, thanks. I'll convert into "plain" a + bx + cx^2 polynomials and stick them in. Do you happen to have the coverage/range for each segment?

[karllark]

See https://jwst-docs.stsci.edu/jwst-mid-infrared-instrument/miri-observing-modes/miri-medium-resolution-spectroscopy

[jdtsmith]

Thanks Karl. I have pushed MIRI data based on these fits, and updated NIRSPEC coverage. I think this is good to go. Could you give it a quick sanity check, look over the naming conventions, etc.?

Testing for a given segment is trivial:

ins='jwst.miri.mrs.ch3.A'
x=np.linspace(*wave_range(ins))
ax.plot(x,resolution(ins,x))

[jdtsmith]

Here's the MRS FWHM plot for reference: jwst_miri_mrs_fwhm.pdf. Pretty big jumps between 3C and 4A.

[karllark]

Can you provide a resolution plot as well? That is what is provide in the docs - makes for easier comparison.

[karllark]

It would be great if we had a documentation page for the different instrument packs that describes them and provide the FWHM/resolution plots. These can be generated on the fly in the docs. Lots of examples of this in the karllark/dust_extinction package.

[karllark]

Also, could you fix the codestyle issues that are causing the tests to fail? These were introduced in a previous pull request.

[jdtsmith]

Also, could you fix the codestyle issues that are causing the tests to fail? These were introduced in a previous pull request.

I could, but it would be better if someone else did so to exercise the instrument pack functionality.

[jdtsmith]

Also, could you fix the codestyle issues that are causing the tests to fail? These were introduced in a previous pull request.

Are we using a particular automated linter?

[karllark]

If you run black on the files you update, it will cleanup most of the codestyle issues. That's what I do so that I don't have to learn and remember all the codestyle rules.

[karllark]

We should separate the automated testing issues from getting people to test the code. They serve different purposes.

[jdtsmith]

We should separate the automated testing issues from getting people to test the code. They serve different purposes.

Violent agreement: for sure we need more hands-on testing, and test spectra under data/. Automated tests can follow.

[jdtsmith]

Hey I got a green check! I also figured out how to run flake8 continuously using the same settings from tox.ini, so hopefully I'll be a net entropy reducer now. This is ready to merge. I'm sure we'll need to tweak coefficients as we test against real spectra, but something is better than nothing. Any objections anyone?

[jdtsmith]

Thanks.