Installing v0.7.288 on NewWorlds (Toulouse)

[clairem789]

I tried installing v0.7.288 in Toulouse, doing:

conda create --name=ap288 python=3.9
conda deactivate
conda deactivate
conda deactivate
conda deactivate
conda activate ap288
cd apero-drs
git fetch
git branch
git checkout v0.7.288-stable-test
pip install -r requirements_developer.txt
python setup/install.py --name=ap288

I got this error message during install.py:

Settings for SPIROU

Traceback (most recent call last): File “/Users/moutou/apero-drs/setup/install.py”, line 669, in main() File “/Users/moutou/apero-drs/setup/install.py”, line 598, in main allparams, args = install.user_interface(params, args, language) File “/Users/moutou/apero-drs/apero/tools/module/setup/drs_installation.py”, line 470, in user_interface value = getattr(args, DATA_ARGS[path]) AttributeError: ‘Namespace’ object has no attribute ‘lbldir’

Any ideas what I got wrong? Thanks!

[njcuk9999]

Hmm okay weird I did test this but then I may have made some changes afterward, I'll look into this.

[eartigau]

I have the least helpful comment: It worked on my laptop … I have no clue as to why it doesn’t work on your machine! I did have to force install a few packages version with pip as I got an error message at the end of the pip install -r requirements_developer.txt

Le 15 sept. 2023 à 12:22, clairem789 @.***> a écrit :

I tried installing v0.7.288 in Toulouse, doing:

conda create --name=ap288 python=3.9 cd apero-drs git fetch git branch git checkout v0.7.288-stable-test pip install -r requirements_developer.txt python setup/install.py --name=ap288

I got this error message during install.py:

Settings for SPIROU

Traceback (most recent call last): File “/Users/moutou/apero-drs/setup/install.py”, line 669, in main() File “/Users/moutou/apero-drs/setup/install.py”, line 598, in main allparams, args = install.user_interface(params, args, language) File “/Users/moutou/apero-drs/apero/tools/module/setup/drs_installation.py”, line 470, in user_interface value = getattr(args, DATA_ARGS[path]) AttributeError: ‘Namespace’ object has no attribute ‘lbldir’

Any ideas what I got wrong? Thanks!

— Reply to this email directly, view it on GitHub https://github.com/njcuk9999/apero-drs/issues/722, or unsubscribe https://github.com/notifications/unsubscribe-auth/AIDD7EHCIKO3P43IZFB5LNLX2R6C7ANCNFSM6AAAAAA42BMC4Y. You are receiving this because you were assigned.

[njcuk9999]

what was the error at the end of the pip install command? That may need to be another issue (or at least sorted out) unless you did something "undesirable".

[njcuk9999]

Oh saying that @clairem789 you are missing a few commands (after the conda create command)

conda deactivate
conda deactivate
conda deactivate
conda deactivate
conda activate ap288

(before the pip install)

Though this may just be a typo on your part - it would mean you've screwed up your conda base environment (and now have packages installed on the base) which would require a reinstall of miniconda

same for your @eartigau, actually this may be why you got a pip error - you already have stuff on your base env (which should be empty)

[njcuk9999]

I actually get the same error as @clairem789 so I'm not sure how yours worked @eartigau.

[njcuk9999]

Okay, I've fixed it for me, it was a small typo not sure when I introduced that.

@clairem789 let me know if it works for you!

I'm going to update your original message with the missing command, just in case someone finds this and tries to install stuff.

[clairem789]

OK this worked. It's running. Thanks!

[clairem789]

hi all, I ran the v288 on the minidata set. It terminated quietly. There is, however, a problem with the drift file. If I show you the first 3 columns of it: cat drift.rdb |awk '{print $1,$2,$3}' rjd vrad svrad N N N 59030.89990958315 -13.688763157518009 0.03322393844511832 59064.900207214036 -0.9807136002033332 0.017711118724625784 59092.90154091403 4.354826159286845 0.023352164214961932 59129.90622941403 14.984310847383584 0.02847004634416082 59065.90160090403 -0.5337401848550704 0.016514566422963887 59063.89836258404 -2.3229601337810357 0.018665750734734487 59115.90494660403 10.607819193017775 0.02952680055619064 58983.90483704314 -24.74407709082389 0.06184237365135822 59059.903573423144 -1.6345180818702645 0.039357397401199916 59060.89786008402 -1.3427143219147544 0.01811968143410379 59062.90343965402 -2.570887800765612 0.018193743193838 59071.899735274026 0.02596114969609671 0.018606732276776998 59154.90748468404 19.23515729628556 0.03180416273350041

So, it's very small and contains only the calibrations, not the Gl699 simultaneous drift measurements. I didn't see an obvious error popping up. Do you have a similar problem? What can I fix? thanks!

[njcuk9999]

I haven't checked myself.

Could you check what is in the /lbl/science/FP directory? It may be that I only copied the FP_FP files and need to add the simultaneous FP files there too.

[clairem789]

Yes, this is it!

[njcuk9999]

Okay I'll look over my code and make sure it is copying them, if it is already then I'll probably ask you more. Thanks for testing it!

[njcuk9999]

Okay I've found the problem, for DPRTYPE I was just getting OBJ_FP (this works for NIRPS) but for spirou I also need to look for DPRTYPE=POLAR_FP (and all GL699 are POLAR_FP in this case).

I've updated v0.7.288-stable-test so a git pull at your end should do it. You can either run the full mini run again, or first check with a new run.ini similar to the lbl_run.ini (though you may have to update it to match the mini run settings with SCIENCE_TARGETS etc).

[clairem789]

I confirm the test runs great with this new version. This is the content of the final file, obtained with :

cat lbl2_GL699_GL699_drift.rdb |awk '{print $1,$2,$3}'

rjd vrad svrad N N N 58984.05390878429 -110193.21788971453 0.9343694308854495 59030.93447572854 -110194.41180462166 0.9862917202035016 59059.86752607301 -110200.42317944583 0.8300514678547082 59060.8571271227 -110194.98335332949 0.7578102092137656 59062.883282140596 -110198.11619409005 0.7481104646586443 59063.77909281326 -110200.81693833967 1.0866472197376829 59064.896404195344 -110196.53830306236 0.8103029075121354 59065.83420328738 -110198.98893844377 0.7569455412705957 59071.83309346251 -110198.83668472123 0.809213616049847 59092.7195982578 -110198.28948795485 0.9288271950084797 59115.702675125794 -110193.74603541825 0.8347963663131514 59129.76672036934 -110196.10300444231 0.780737870584925 59154.68532708858 -110193.4961615487 0.8222449809959744

Shall we re-reduce all data? or wait for a proper comparison between places?

[njcuk9999]

If you can just drop your GL699_GL699 files in here: https://drive.google.com/drive/folders/1XpgLS_mD0-vZXMFIJCblZPoKBsC0xoC1?usp=sharing

You should be able to drop stuff in there after signing in to gmail (I think)

Maybe all the .rdb files for GL699_GL699

I've put a directory in there called v0.7.288-stable-test and a sub-directory for each data set

If I have time I'll compare but I'll also drop my data in there so others can compare

[clairem789]

OK, thanks; file transfer done on my side. Rali versus NewWorlds:

comparison on lbl2_GL699_GL699_drift.rdb

std dev of RVs NW and RALI (m/s) = 2.5535928784049684 2.551991094348238 RMS of difference (m/s) = 0.05231690143588415

comparison on drift.rdb

std dev NW/RALI (m/s) = 5.021218915648642 5.0218668753135765 RMS of difference (m/s) = 0.0037176083757248157

[njcuk9999]

I actually realized the rali ones weren't the most up-to-date version of v0.7.288 so I've re-run them today (with the other machines) I'll upload them as soon as they are done.

2.5 m/s seems a lot - I imagine it's probably the LBL again - this is something to fix on my return - or maybe bug @eartigau about it in the meantime.

I guess we should also upload our tcorr files + blaze files so someone could run lbl on all versions and/or compare the tcorr themselves - that would be the next step: really understanding where the difference starts!

[njcuk9999]

Lets continue the comparison discussion here: https://github.com/njcuk9999/apero-drs/issues/708