Trouble with Part 2.0 part.map.grid_rhi_sweeps(radar).

[wxwolff]

I am getting an error when I try to use the new grid_rhi_sweep method: "AttributeError: 'Xradar' object has no attribute 'extract_sweeps'"

See Jupyter screenshot below:

[zssherman]

@wxwolff Thanks for raising this issue! Yeah there might be a missing extract sweeps function for xradar. @mgrover1 how did we want to go about this? Do we need to add the extract sweeps method to xradar, or is there another way of approaching this?

[zssherman]

I discussed we @mgrover1. I'll work on a fix soon

[zssherman]

@wxwolff I added in the extract sweep method which should fix the issue now.

[wxwolff]

pyart_install.txt

[zssherman]

@wxwolff Interesting, is this a fresh pyart repo as well? I'm wondering if there is leftover or outdated files from the old rsl code we removed?

[wxwolff]

Yes. I cloned from https://github.com/ARM-DOE/pyart.git [signature_2656591273]<https://www.nasa.gov/> David B. Wolff Assistant Lab Chief, Code 612 GPM Deputy Project Scientist GPM GV System Manager Goddard Space Flight Center Wallops Flight Facility Code 612 BLDG N159, ROOM E220 M: 443-262-1146 O: 757-825-1515 ***@***.*** From: Zach Sherman ***@***.***> Date: Wednesday, February 26, 2025 at 13:50 To: ARM-DOE/pyart ***@***.***> Cc: Wolff, David B. (WFF-6120) ***@***.***>, Mention ***@***.***> Subject: [EXTERNAL] [BULK] Re: [ARM-DOE/pyart] Trouble with Part 2.0 part.map.grid_rhi_sweeps(radar). (Discussion #1734) CAUTION: This email originated from outside of NASA. Please take care when clicking links or opening attachments. Use the "Report Message" button to report suspicious messages to the NASA SOC. @wxwolff<https://github.com/wxwolff> Interesting, is this a fresh pyart repo as well? I'm wondering if there is leftover or outdated files from the old rsl code we removed? — Reply to this email directly, view it on GitHub<#1734 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ADJRUYWSPO6VSPILBCP4IYD2RYEINAVCNFSM6AAAAABXY6WJ62VHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTEMZTGAYDENA>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[zssherman]

Gotcha I'll see if I can duplicate the issue

[zssherman]

I sadly haven't been able to reproduce the issue yet, what cython version do you have for this environment?

[wxwolff]

Hi Zack, Thanks for trying, I am running conda 25.1.1 and cython 3.12.2. I installed cython via “conda install -c conda-forge cython”. D [signature_310786882]<https://www.nasa.gov/> David B. Wolff Assistant Lab Chief, Code 612 GPM Deputy Project Scientist GPM GV System Manager Goddard Space Flight Center Wallops Flight Facility Code 612 BLDG N159, ROOM E220 M: 443-262-1146 O: 757-825-1515 ***@***.*** From: Zach Sherman ***@***.***> Date: Thursday, February 27, 2025 at 10:12 To: ARM-DOE/pyart ***@***.***> Cc: Wolff, David B. (WFF-6120) ***@***.***>, Mention ***@***.***> Subject: [EXTERNAL] [BULK] Re: [ARM-DOE/pyart] Trouble with Part 2.0 part.map.grid_rhi_sweeps(radar). (Discussion #1734) CAUTION: This email originated from outside of NASA. Please take care when clicking links or opening attachments. Use the "Report Message" button to report suspicious messages to the NASA SOC. I sadly haven't been able to reproduce the issue yet, what cython version do you have for this environment? — Reply to this email directly, view it on GitHub<#1734 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ADJRUYRNH3YMUKE7ZHENYN32R4TMPAVCNFSM6AAAAABXY6WJ62VHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTEMZUGA2DMMY>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[zssherman]

Gotcha, hmm yeah I'm trying to think of what else to check... did you have pyart previously conda installed when you were trying to use the gridding code at the top of this discussion? I can do a small release on conda with the extract sweeps fix until I figure out why it wont install from git

[wxwolff]

I tried the git install on a fresh anaconda install. The only packages that I had pre-loaded besides the anaconda default were cython and pyproj. If you could do a quick update with Conda that would be great. I have been wanting to have this functionality for so long, thanks! S [signature_1030964688]<https://www.nasa.gov/> David B. Wolff Assistant Lab Chief, Code 612 GPM Deputy Project Scientist GPM GV System Manager Goddard Space Flight Center Wallops Flight Facility Code 612 BLDG N159, ROOM E220 M: 443-262-1146 O: 757-825-1515 ***@***.*** From: Zach Sherman ***@***.***> Date: Thursday, February 27, 2025 at 10:18 To: ARM-DOE/pyart ***@***.***> Cc: Wolff, David B. (WFF-6120) ***@***.***>, Mention ***@***.***> Subject: [EXTERNAL] [BULK] Re: [ARM-DOE/pyart] Trouble with Part 2.0 part.map.grid_rhi_sweeps(radar). (Discussion #1734) CAUTION: This email originated from outside of NASA. Please take care when clicking links or opening attachments. Use the "Report Message" button to report suspicious messages to the NASA SOC. Gotcha, hmm yeah I'm trying to think of what else to check... did you have pyart previously conda installed when you were trying to use the gridding code at the top of this discussion? I can do a small release on conda with the extract sweeps fix until I figure out why it wont install from git — Reply to this email directly, view it on GitHub<#1734 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ADJRUYQX77TM5YVBOF4TI732R4UDHAVCNFSM6AAAAABXY6WJ62VHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTEMZUGA2TOMI>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[zssherman]

@wxwolff Sounds good! I have a conda release going now, might take a bit to load into their system. For the rsl issue, so i was digging further, i think its trying to pull from /usr/local/trmm/include/. Do you have a copy of trmm_rsl in that directory? I'm wondering if its pulling from there and having issues. I had a colleague install a fresh pyart and conda installed trmm_rsl and set the export RSL_PATH= to the environment directory of conda and it worked. So I'm curious if there is a mismatch happening or something

[zssherman]

And I think the reason its not failing for conda is its not compiling as trmm_rsl is by default excluded from the dependencies. But that's the theory so far

[wxwolff]

I didn’t think that I had installed RSL on this machine as it is new and I haven’t used RSL in years, but /usr/local/trmm is on my system. It’s probably not a full install so that might be the issue. I can try to remove it and see what happens. [signature_704042332]<https://www.nasa.gov/> David B. Wolff Assistant Lab Chief, Code 612 GPM Deputy Project Scientist GPM GV System Manager Goddard Space Flight Center Wallops Flight Facility Code 612 BLDG N159, ROOM E220 M: 443-262-1146 O: 757-825-1515 ***@***.*** From: Zach Sherman ***@***.***> Date: Thursday, February 27, 2025 at 11:40 To: ARM-DOE/pyart ***@***.***> Cc: Wolff, David B. (WFF-6120) ***@***.***>, Mention ***@***.***> Subject: [EXTERNAL] [BULK] Re: [ARM-DOE/pyart] Trouble with Part 2.0 part.map.grid_rhi_sweeps(radar). (Discussion #1734) CAUTION: This email originated from outside of NASA. Please take care when clicking links or opening attachments. Use the "Report Message" button to report suspicious messages to the NASA SOC. @wxwolff<https://github.com/wxwolff> Sounds good! I have a conda release going now, might take a bit to load into their system. For the rsl issue, so i was digging further, i think its trying to pull from /usr/local/trmm/include/. Do you have a copy of trmm_rsl in that directory? I'm wondering if its pulling from there and having issues. I had a colleague install a fresh pyart and conda installed trmm_rsl and set the export RSL_PATH= to the environment directory of conda and it worked. So I'm curious if there is a mismatch happening or something — Reply to this email directly, view it on GitHub<#1734 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ADJRUYS5ZJ6Y4IFDUW6V4ED2R45ZBAVCNFSM6AAAAABXY6WJ62VHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTEMZUGE3DCMA>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[wxwolff]

Thanks! Almost there...

Ok, so I can now "git" updates after removing a stale /usr/local/trmm from my mac. I can also execute the grid_rhi_sweeps method, but it does not produce any valid "y" values (they are all 0's). I have attached a notebook (added a .txt suffix) and a plot of the file using the traditional pyart display method which shows there is valid data.

You can download the CF file here: https://wallops-prf.gsfc.nasa.gov/Data/WOLFF/NPOL1_2019_1014_000206_rhi.cf

Here is the line I use to grid the RHI.

radar_ds = pyart.map.grid_rhi_sweeps(radar, target_sweeps=None, grid_size=(801,801),
                                     grid_limits='auto', max_z=10000., 
                                     az_rounding_frac=0.25)

Grid_pyart2.ipynb.txt

[zssherman]

@wxwolff More so, to pinpoint the issue, does the gridding work correctly if you read with pyart.io.read instead and not use xradar? Seeing if this is still and xradar object issue or a gridding issue

[wxwolff]

Hi, Same result using either radar = tree.pyart.to_radar(f) or radar = pyart.io.read(f) D [signature_3609435193]<https://www.nasa.gov/> David B. Wolff Assistant Lab Chief, Code 612 GPM Deputy Project Scientist GPM GV System Manager Goddard Space Flight Center Wallops Flight Facility Code 612 BLDG N159, ROOM E220 M: 443-262-1146 O: 757-825-1515 ***@***.*** From: Zach Sherman ***@***.***> Date: Friday, February 28, 2025 at 10:39 To: ARM-DOE/pyart ***@***.***> Cc: Wolff, David B. (WFF-6120) ***@***.***>, Mention ***@***.***> Subject: [EXTERNAL] [BULK] Re: [ARM-DOE/pyart] Trouble with Part 2.0 part.map.grid_rhi_sweeps(radar). (Discussion #1734) CAUTION: This email originated from outside of NASA. Please take care when clicking links or opening attachments. Use the "Report Message" button to report suspicious messages to the NASA SOC. @wxwolff<https://github.com/wxwolff> More so, to pinpoint the issue, does the gridding work correctly if you read with pyart.io.read instead and not use xradar? Seeing if this is still and xradar object issue or a gridding issue — Reply to this email directly, view it on GitHub<#1734 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ADJRUYW6TLXOQCDKX5LF2UD2SB7L7AVCNFSM6AAAAABXY6WJ62VHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTEMZVGMYTQNI>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[zssherman]

@wxwolff Gotcha.... I'll dig into the gridding code to see what I can find. Thanks for confirming that!

[wxwolff]

Not that it really matters, but the read.io radar has six more attrs than the Xradar. Otherwise they are identical.

{
'n_gates_vary': False,
'field_names': [ZZ, CZ, DR, RH, PH, KD, SW, VR, RR, RP, RC, NW, FH, DM, HC],
'volume_number': 0,
'platform_type': 'fixed',
'instrument_type': 'radar',
'primary_axis': 'axis_z'
}

[wxwolff]

Oops... the data_var attrs are also slightly different. The pyart version contain additional:

{
'_FillValue:':, -32767,
'coordinates' : 'time range'
}

[zssherman]

@wxwolff We are trying to pinpoint the issue still, but looks to be either the georeferencing in extract_sweeps or grid_rhi_sweeps, @mgrover1 is going to work with me Monday to see if we can find the issue.

[zssherman]

@wxwolff Can you set your own grid_limits instead of 'auto' in the meantime, I believe it has to do with how grid_limits is being calculated with negative values in the auto function

[wxwolff]

Ok, but what is the format of the grid limits for an RHI? I would like to have it go to 10 km vertically and 100 km radially. [signature_3101352163]<https://www.nasa.gov/> David B. Wolff Assistant Lab Chief, Code 612 GPM Deputy Project Scientist GPM GV System Manager Goddard Space Flight Center Wallops Flight Facility Code 612 BLDG N159, ROOM E220 M: 443-262-1146 O: 757-825-1515 ***@***.*** From: Zach Sherman ***@***.***> Date: Monday, March 24, 2025 at 12:30 To: ARM-DOE/pyart ***@***.***> Cc: Wolff, David B. (WFF-6120) ***@***.***>, Mention ***@***.***> Subject: [EXTERNAL] [BULK] Re: [ARM-DOE/pyart] Trouble with Part 2.0 part.map.grid_rhi_sweeps(radar). (Discussion #1734) CAUTION: This email originated from outside of NASA. Please take care when clicking links or opening attachments. Use the "Report Message" button to report suspicious messages to the NASA SOC. @wxwolff<https://github.com/wxwolff> Can you set your own grid_limits instead of 'auto' in the meantime, I believe it has to do with how grid_limits is being calculated with negative values in the auto function — Reply to this email directly, view it on GitHub<#1734 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ADJRUYQOR4CZ52JBF7UI2GT2WAXJ3AVCNFSM6AAAAABXY6WJ62VHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTENRQGQ3TONY>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[zssherman]

@wxwolff I think for this case:
grid_limits=(
(0.0, 10_000),
(-100000.0, 0.0),
(-100000.0, 0.0),)

[zssherman]

@wxwolff We figured out a fix in #1740