k2sc example fails to run

[gjtucson]

Hi folks,

I am trying to use the k2sc code for Kepler K2 data. I've run into a couple of issues for which I hope you can direct me back to the right path.

To get started, I am trying to run the example notebook at:

https://github.com/OxES/k2sc/blob/master/notebooks/lightkurve.ipynb

I've had to make a change for the current version of lightkurve, in that KeplerTargetPixelFile.from.archive is not available, so I switched to search_targetpixelfile, and I specify (cadence='long', mission='K2').

Then I can get a tpf returned and the image looks exactly like the one in the example.

However, the lc.flux plot looks a bit different, which may be ok, but suspicious. Interestingly, when I look at the FITS header in the example, it shows as campaign 6 (correct). The output for the command:

lc.k2sc()

though, prints "Using default splits [2344] for campaign 5. Where did campaign 5 come from, and is that an issue?

Then k2sc iterates for 103 times and eventually completes. When I run the code, lc.k2sc() uses splits for campaign 6 (which seems right), iterates 37 times, and subsequently dies after:

"Starting final outlier detection" in "detrend".

Any suggestions would be great!

George Jacoby
[email protected]

[benjaminpope]

Hi George,

Try cloning the latest version from GitHub - and can you post your exact notebook commands and outputs?

All the best,

Ben

[gjtucson]

Hi Ben, I've been trying all sorts of approaches to doing what you suggest and I'm running into walls. I admit that I am very new to Python, but I've used about a dozen other languages in the past. Anyhow, my problem is in trying to install the cloned code. I can install from the archive just fine ("pip install k2sc"), but not from my local directory ("pip install -e local_dir"). If I do a "pip list", the k2sc code shows up from that directory, but I cannot "import k2sc", even after restarting the kernel. This is unfortunate because the error appears to be corrected in the local code - the error is that sigma_clip is returning 3 values and only 2 are expected according to the traceback. I see that a third variable has been added in detrend to fix that. Part of the problem may be related to environments, which I haven't figured out yet.  I cannot install from the host (Mac OSX) command line, as shown in the README file; I have to install from within Anaconda using pip. Any suggestions would be great! Thanks, George

…

On 5/4/20 12:57 PM, Benjamin Pope wrote: Hi George, Try cloning the latest version from GitHub - and can you post your exact notebook commands and outputs? All the best, Ben — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#26 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/APNLY7FCD3NQEXJFGBKJU53RP4M25ANCNFSM4MZBDQEQ>.

[benjaminpope]

Hi George,

Either

pip install k2sc

again, as I have updated the PyPI distribution now, or if doing it from the GitHub, use

python setup.py install

All the best,

Ben

[gjtucson]

Hi Ben, You are fantastic! Thanks! That was the problem. Now I can look at the data it sure cleans up nicely after k2sc. I really appreciate the quick fix. I owe you a beer (or coffee or whatever). George

…

On 5/5/20 4:28 PM, Benjamin Pope wrote: Hi George, Either |pip install k2sc| again, as I have updated the PyPI distribution now, or if doing it from the GitHub, use |python setup.py install| All the best, Ben — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#26 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/APNLY7G3OITM4HDU7XRMVULRQCOJ5ANCNFSM4MZBDQEQ>.