Error on Arch Linux: np.mat was removed in NumPy 2.0 in specFunctions.py

[JimKnopfIoT]

When trying to run PySpectrometer2-USB-v1.0.py on Arch Linux, python-numpy 2.2.4-1, i'm getting this error:

./PySpectrometer2-USB-v1.0.py --device 2 --fps 15 [info] W, H, FPS 640.0 480.0 25.0 Loading calibration data... Loading of Calibration data failed (missing caldata.txt or corrupted data! Loading placeholder data... You MUST perform a Calibration to use this software!

Calculating second order polynomial...
2
3.125e-05 x + 0.4375 x + 380
Generating Wavelength Data!

Done! Note that calibration with only 3 wavelengths will not be accurate!
Traceback (most recent call last):
File "/home/user1/PySpectrometer2/src/./PySpectrometer2-USB-v1.0.py", line 263, in
intensity = savitzky_golay(intensity,17,savpoly)
File "/home/user1/PySpectrometer2/src/specFunctions.py", line 138, in savitzky_golay
b = np.mat([[k**i for i in order_range] for k in range(-half_window, half_window+1)])
^^^^^^
File "/usr/lib/python3.13/site-packages/numpy/init.py", line 400, in getattr
raise AttributeError(
...<3 lines>...
)
AttributeError: np.mat was removed in the NumPy 2.0 release. Use np.asmatrix instead.

Swapping np.map with np.matrix leads to other errors.
Any hotfix for that problem available? Running PySpectrometer2-Picam2-v1.0.py on a RPi4 via VNC is quite slow. So would prefer running it on my Laptop.

[eabase]

@JimKnopfIoT

Swapping np.map with np.matrix leads to other errors.

Show all errors

[gomills]

Hi, could you manage to solve it?

[JimKnopfIoT]

@gomills I found a better solution for my needs. It's more expensive (220 Euro) but works realy well. It's a Spectryxs Blue Box. Don't know why it's that expensive. Under the hood is a ESP8266 board and a fiber optic coupler or someting like that. It works with Octave, Python and with a software from Spectryx. I prefer Octave (on my Arch Linux Laptop).

My intention building the PySpectrometer was to inspect some UV and IR Lasers and LEDs. I found UV and IR light too weak for the PySpectrometer. I had better results with the Spectryx Blue Box. I attach some images. I build a Ulbricht Sphere for better results and to prevent burning the sensor. It's inside the 3D printed cube (2 parts, screwed together). Maybe it's interesting for others: I first tried BaSO4, mixed with 1% PVA glue and some demineralised water. It dried poorly, cracked after drying, and did not adhere to the surface. It crumbled. I then used a mixture from "Kremer Pigmente, Germany". It's also BaSO4, they called it Blanc Fix. Instead of PVA glue they used something called Klucel(TM) G (Hydroxypropylcellulose) and also demineralised water. This mixture adhered much better to the surface.

[gomills]

@JimKnopfIoT Very interesting and good idea the Ulbricht sphere, as well as your clean set up. If I can ask, what was the purpose of characterizing the lasers?
Undoubtedly you have a hand for building and crafting things. I will list some questions that your post arised me, hope you can answer and hear your thoughts and ideas:

1.- Do you know which sensor Spectryxs Blue Box is using?
2.- What exactly happened with the PySpectrometer? From what I understand you couldn't set it up correctly due to the Numpy exception. If you managed to test it, what do you mean with light being too weak for the instrument?
3.- I became very interested in this fantastic PySpectrometer2 project from Les because of its apparent results from the videos he uploaded. And I was looking for a spectrometer set up to build a primitive, qualitative-only HPLC, but worried about hardware sensitivity and discussing it, since it's the only major issue I see and you mention something of too weak light.

[JimKnopfIoT]

@gomills 1. I opend the Spectryx Blue Box but couldn't identify the sensor. I decided not to disassemble the parts as long as the device works as intended.
2. As you can see, the PySpectrometer works on the RPi. I bought a display and use it as a standalone device. A 20.000mAh Powerbank is included in my setup (below the Spectrometer). For calibrating i build a Sodium lamp (it's not just a bulb, you need a high voltage transformer and a capacitor). The python stuff written for the PySpectrometer2 is based on older python libraries. I couldn't get it to work on a newer RPi image or on my Arch linux. Since i'm not a programmer i couldn't fix the problems when running it on my Arch system. It only works for me in the setup explained by Les.

Using the definition "light too weak" i mean the ability of the device to detect a light from a low power light source like a UV LED or IR LED. Even IR Laser light is not detected properly. Maybe it's wavelength is out of range for the camera. As you can see in the picture, a 405 nm Laser source (5mW) has a low peak level on the PySpectrometer2 but is detected with a high peak level on the Spectryxs Blue Box.

The purpose of characterizing LED's and Lasers arose from a repair job that i've done. Someone asked my if i can repair two of his Fitness clocks. They were controlled by a IR remote. But the guy forgot to mention that and also forgot to hand over the remotes. I found the IR receiver sensor on the board and didn't know what wavelength it was working on. At that point i wanted to identify the wavelength of an unknown UV or IR LED. I found Les project and couriosity took over my wallet.

For me, the PySpectrometer2 works well as a standalone device in the range of 420 to 800 nm. I have a ComMarker Omni 1 and a xTool F1 Ultra. Both uses wavelengths outside this range (Omni 1 355 nm, xTool 1064 nm). For experiments with this devices and for low power sources i found the Spectryx Blue Box the better choice.

I may upload my 3D prints for the Ulbricht Sphere and the PySpectrometer2 enclosure/fixture.

[gomills]

Thank you for answering me. Glad to hear you could solve your problem. It's a shame PySpectr couldn't make it work the problem.

By what you say, PySpectr won't work well in the UV range regarding sensitivity purposes, which is the range I'm interested in. So sad. About the outdated software, yes, PySpectrometer2 would need an integral refactor and update to modern python frameworks and coding practices. I would commit to that arduous task if I was to use it, but I'm still choosing what would be best for my set up.

[JimKnopfIoT]

FYI i uploaded my 3D files from the enclosure/fixture and Ulbricht sphere https://github.com/JimKnopfIoT/Spectrometer-fixture-and-Ulbricht-Sphere

[gomills]

FYI i uploaded my 3D files from the enclosure/fixture and Ulbricht sphere https://github.com/JimKnopfIoT/Spectrometer-fixture-and-Ulbricht-Sphere

I enjoyed reading it. Maybe BaSO4 didn't adhere because of the material?

[eabase]

@JimKnopfIoT
Great photos and thatnk you so much for sharing your detailed experience! 💯 🥇

@gomills

Great questions.

Regrading UV sensitivity issues:
Make sure you are not using any kind of glass/plastic window as almost all materials apart the following, is not transparent to UV.

• FS : Fused silica glass - Drawback: refractive indices & dispersions
• FP : Fluoride phosphate glass - Drawback: poor chemical stability
• BS : Borosilicate glass - Drawback: None!

From here.

[gomills]

Regrading UV sensitivity issues: Make sure you are not using any kind of glass/plastic window as almost all materials apart the following, is not transparent to UV.

* `FS` : Fused silica glass  - **Drawback**: refractive indices & dispersions

* `FP` : Fluoride phosphate glass - **Drawback**: poor chemical stability

* `BS` : Borosilicate glass - **Drawback**: None!

From here.

Thank you for the help! I definitely will take a look at those. I'm documenting all I can for this.