SOAR triplespec#2139
Conversation
profxj
added
the
New Spectrograph
|
Companion PR: |
kbwestfall
left a comment
kbwestfall
left a comment
There was a problem hiding this comment.
Thanks! I mostly have some clean-up comments, but there were two items that I wanted to draw your attention to:
- Are we using the
darkclassification appropriately, or should the relevant frames really belampoffflat? - It's interesting to me that you had to turn off
auto_pcawhen it's on for NIRES. Any thoughts on this?
| ``frametype`` is automatically assigned to each frame using the values of | ||
| various header keywords (see :meth:`~pypeit.spectrographs.soar_tspec.SOARTSPECSpectrograph.check_frame_type`): | ||
| the dome flats taken with the lamp **on** are typed ``pixelflat,trace``, the | ||
| dome flats taken with the lamp **off** are typed ``dark``, and the science |
There was a problem hiding this comment.
"dome flats taken with the lamp off are typed dark" feels odd to me. Are they actually darks (and used as such) or are they lampoffflats?
| Dome flats are taken in pairs: lamp **on** and lamp **off**. PypeIt types the | ||
| lamp-on flats as ``pixelflat,trace`` and the lamp-off flats as ``dark``; the | ||
| dark is subtracted to remove the thermal/dome background before the field flat | ||
| and order edges are constructed. Bias subtraction and overscan correction are |
There was a problem hiding this comment.
Got it now. But shouldn't these be typed as lampoffflat?
| - :doc:`../tutorials/soar_triplespec_howto` -- a full worked example reduction. | ||
| - :doc:`../calibrations/wave_calib` | ||
| - :doc:`../calibrations/flat` | ||
| - :doc:`../A-B_differencing` |
There was a problem hiding this comment.
Personally, I think it's better if we use :ref: links because it doesn't require us to put in relative paths.
| (order numbers 7 through 3), covering roughly 0.8--2.47 microns. If you're | ||
| having trouble reducing your data, we encourage you to work through this | ||
| tutorial first. See also the instrument notes at | ||
| :doc:`../spectrographs/soar_triplespec`, join our `PypeIt Users Slack |
There was a problem hiding this comment.
Same preference for :ref: here.
|
|
||
| This produces a directory ``soar_tspec_A`` containing the pypeit file | ||
| ``soar_tspec_A.pypeit``. The key part is the :ref:`data_block`, which for this | ||
| example looks like: |
There was a problem hiding this comment.
Probably should note here that this is not the output from pypeit_setup directly, but after it has been edited by hand.
| if meta_key == 'mjd': | ||
| ttime = Time(headarr[0]['DATE-OBS'], format='isot') | ||
| return ttime.mjd | ||
| else: |
| and used to constuct the :class:`~pypeit.metadata.PypeItMetaData` | ||
| object. | ||
| """ | ||
| return []#'dispname'] |
| gain = np.atleast_1d(3.8), | ||
| ronoise = np.atleast_1d(3.5), | ||
| datasec = np.atleast_1d('[:,:]'), | ||
| oscansec = None #np.atleast_1d('[:,:]') |
| # With the native arxiv every order self-reidentifies at high cc, so the | ||
| # default cc_thresh is fine and no override is needed. | ||
| par['calibrations']['wavelengths']['reid_arxiv'] = 'soar_triplespec.fits' | ||
| # par['calibrations']['wavelengths']['ech_fix_format'] = True |
There was a problem hiding this comment.
remove this and other similar commented lines?
| par['calibrations']['slitedges']['det_buffer'] = 10 | ||
| # The PCA is terrible, but fortunately the polynomial fits are good; so | ||
| # just skip the PCA | ||
| par['calibrations']['slitedges']['auto_pca'] = False |
There was a problem hiding this comment.
Interesting that this is so different from Keck/NIRES
bpholden
left a comment
bpholden
left a comment
There was a problem hiding this comment.
I have two questions that are not important and one more case of flatlampoff. So I will approve with the assumption that Kyle's points are the important ones.
| ---------------- | ||
|
|
||
| PypeIt applies pixel-to-pixel and illumination flat-field corrections built from | ||
| the dome flats; the lamp-off dome flats (typed ``dark``) are subtracted first to |
There was a problem hiding this comment.
Another mention of the dome off flats as "darks"
|
|
||
| After the calibrations, PypeIt identifies sources, performs global and local sky | ||
| subtraction, and extracts 1D spectra; see :ref:`object_finding`. Here is the | ||
| object-finding QA for the science target in order 7: |
There was a problem hiding this comment.
This is more a "Brad is confused question" but we the A-B differencing mentioned above, do we need to do global and local sky subtraction? Or would just one be enough?
| """ | ||
| Return the expected spatial position of each echelle order. | ||
| """ | ||
| return np.array([0.3096, 0.4863, 0.6406, 0.7813, 0.9424]) |
There was a problem hiding this comment.
Another "Brad is confused" question, the fixed order_spat_pos I thought was not used when
par['calibrations']['wavelengths']['ech_fix_format'] = False
|
Dev-suite started. |
|
There are some test failures, but I think they must be unrelated. I expect the P200/NGPS failure is because the data in the Google Drive changed. The Keck/DEIMOS/600ZD failure is odd. The log doesn't report any error message, but just the last few lines of what looks like normal logging messages. I expect something might have simply killed the process. The fact that the |
4 participants
Add support for the SOAR Triplespec spectrograph
Well tested on files provided by Kyle Davis
Includes a Tutorial HOWTO which is nearly complete (missing Ginga screen shots)
Originally coded by X, but completed by Claude:
https://github.com/pypeit/PypeIt-development-suite/blob/dev_soar_tspec/claude_prompts/soar_tspec.md