Matrix calculation

[gdemichele]

Dear all,

I am trying to run a "2.5" phase method with pyradiance.
This is the workflow in bash:

1. obj/window.rad temp/sky_glow1_r1.rad Zone_1_tot.rad > data/matrix/DSF_win.dmx
2. gendaymtx -O1 wea/wea_location.wea > data/matrix/wea_location.smx
3. rmtxop data/matrix/DSF_60.mtx data/matrix/DSF_win.dmx data/matrix/wea_location.smx | rmtxop -c 0.33 0.33 0.34 - > data/layerAbs_DSF_60.dat
   With DSF_60.mtx including the BSDF of the absorprion for 4 layers.

By now, I was just able to replicat the second step (gendaymtx) with pyradiance. I am having trouble in calculating the daylithing matrix; looking at the description of the function, see image below, I do not know where to input the sender. Sender, receiver, and scene are paths to specific file in my code.

Any suggestion ? Also for the last step using rmtxop.

Best,
Giuseppe

[taoning]

Dear Giuseppe,
If I understand you question correctly, it should look something like this:

mtx = rfluxmtx(receiver="temp/sky_glow1_r1.rad", surface="obj/window.rad", scene=["Zone_1_tot.rad"])

We need to update the doc string.

[gdemichele]

Hi Taoning,

super, it works. I had to change the scene input to point directly to the rad file of scene and material. Initially, within zone_1_tot.rad I was including all the geometries and material with !xform, but apparently there is a problem with path, I did some test changing paths, but didn't work, probably is a matter WINDOW OS path managment, I need to check this better.

Additionally, is multiprocess supported for rfluxmtx in pyradiance? I tried to include in params , but it did not work. Actually, also with other paremters eg "-ab 1", I got an error! See my code below. Without the params option the command works fine.

day_mtx = pr.rfluxmtx(receiver=path_receiver, surface=path_sender, scene=[mat_path, scene_path], params=["-n 6 -ab 1"])

Last, can you also suggest an implementation for this command: rmtxop data/matrix/DSF_60.mtx data/matrix/DSF_win.dmx data/matrix/wea_location.smx | rmtxop -c 0.33 0.33 0.34 - > data/layerAbs_DSF_60.da

thanks for the quick support!

[taoning]

!xform commands in file should work, let me know if you see any error.

Multiprocessing won't work on Windows but should work on Linux and MacOS and WSL.
params should be a list of strings. So in your case this will work.

day_mtx = pr.rfluxmtx(receiver=path_receiver, surface=path_sender, scene=[mat_path, scene_path], params="-n 6 -ab 1".split())

I realize rmtxop is broken in pyradiance. Will let you know when it's fixed.

In the meantime, I recommend you using frads for this type of workflow as the matrix multiplication will be a lot faster if you are doing many operations.

import frads as fr

# in this case, window.rad is just regular polygon without the #@rfluxmtx tags
window_sender = fr.SurfaceSender(surfaces=["obj/window.rad"], basis="kf")
sky_receiver = fr.SkyReceiver(basis="r4")

day_mtx = fr.Matrix(sender=window_sender, receivers=[sky_receiver], surfaces=['Zone_1_tot.rad"])

day_mtx.generate(params=['-ab', '1'], nprocs=4)

sky_mtx = fr.load_matrix("data/matrix/wea_location.smx")
abs_mtx = fr.load_matrix("data/matrix/DSF_60.mtx")
result = fr.matrix_multiply_rgb(abs_mtx, day_mtx,.array sky_mtx, weights=[0.33, 0.33, 0.34])

[gdemichele]

Hi Taoning,

In the end, frads did the job for the simulation workflow I am setting up! Thank you so much! In this workflow I just need to generate the sky matrix and perform the matrix multiplication, the rest remain fixed, except for bsdf which is generate in another process. Nevertheless, I would like to bring a couple of problems in the process that I was able to overcome performing those in Radiance:

1. day_mtx calculation. I am not able to create the sender. window.rad contains the polygon, I tried also with and without glow material inside the file, but I got the same error (see image below). It seems that the function is not able to point at the sender file, I tried different position of the file as well, but didn't work. Addittionally, if I input a wrong file path, I got still the same error....no idea what is going on

path_sender = r"C:\Users\GDeMichele\Documents\ubuntu_shared\Absorption\DSF_abs_frads\obj\window.rad"
window_sender = fr.SurfaceSender(surfaces=[path_sender], basis="kf")

this is the window.rad file content, I also tried a version including the glow material.

window_mat polygon window
0
0
12
1 1 0
1 1 1
0 1 1
0 1 0

2. epw2wea. I wasn't able to find a function for this in frads or pyradiance. Do you have any suggestion to include epw2wea in my python workflow?

best,
giuseppe

[taoning]

1. sorry missed a step there, here is the code that should run:

import frads as fr

# in this case, window.rad is just regular polygon without the #@rfluxmtx tags
window_prims = fr.unpack_primitives("window.rad")
window_sender = fr.SurfaceSender(surfaces=window_prims, basis="kf")
sky_receiver = fr.SkyReceiver(basis="r4")

day_mtx = fr.Matrix(
    sender=window_sender,
    receivers=[sky_receiver],
)

day_mtx.generate(params=['-ab', '1'], nproc=4)

2. We have the following workflow for converting epw to sky matrix. In this case, you can use sky_matrix directly in the example above.

wea_metadata, wea_data = parse_epw_file("test.epw")
wea_header = wea_metadata.wea_header()
wea_str = wea_header + "\n".join(str(d) for d in wea_data)
sky_matrix = pr.gendaymtx(
            wea_str.encode(),
            sun_only=False,
            onesun=False,
            outform="d",
            daylight_hours_only=True,
            mfactor=4,
        )