Feasibility: once the code is installed how difficult would it be to extract comparisons of all normalized and band-passed Early Reflection slopes?

[digbyphonic]

Please forgive my naivety: I use R as best I can, and will be attempting to install a clone of the Spinorama Repository in the near future.
I intend to use WSL2 and VS Code under Windows 11 Pro.
My first attempt throws up errors of missing modules, so I will clear the existing WSL2 and start afresh.

I would like to compare normalized and band-passed Early Reflection slopes for all 'high' quality and 'default' loudspeaker measurements in the database. (Amongst other tasks).
Should I be looking at importing all existing ER data into dataframes long-hand, and working from there?
Or should I be able to figure out how to use the existing code kindly provided within the site to more easily combine results?

Again, dear reader, please forgive my lack of coding expertise.
I'm highly skilled in audio engineering, but brand new to Python.
I'm determined to achieve my goal in the most efficient way possible, but realise I may have to stumble around and go the long way.

Thank you for making the data so accessible in this excellent and valuable resource.

[pierreaubert]

Thanks for trying the library. Spinorama is complicated to install and only well tested on linux. You may have better chances with a Linux VM instead of WSL2.

The easiest way is to dump the data in a dataframe that R can read.
After you run

generate_graphs.py

you will have all the data in the cache directory. There is cache_load function that will load all in memory (or you can filter if you want to). An example of how to use it is in generate_meta.py. The code is defined

spinorama/generate_common.py

Line 264 in 5d19fb4

def cache_load_seq(filters, smoke_test):

if that does not work out of the box (likely), please ping me here. I can have a look this weekend.

[pierreaubert]

See https://github.com/pierreaubert/spinorama/blob/develop/experimental/discussion-339.ipynb for an example.

[digbyphonic]

Thank you very much for your assistance.
I think I need to remain on WSL2 as I'm a fair way through studies and need to keep my main machine working as-is.

I've created a venv, and am accessing it using JupyterLab in my browser.
There's an error in one of the logger code blocks "build/debug_optim.log" does not exist.
I created this file.

Next error is block [8]:

---

KeyError Traceback (most recent call last)
Cell In[8], line 1
----> 1 datas['DIYSG HT-10']['Misc']['misc-mtg90'].keys()

KeyError: 'DIYSG HT-10'

---

I next tried %run generate_graphs.py and got the following error:

---

Error Traceback (most recent call last)
File ~/.venv/spinorama/generate_graphs.py:72
70 from datas import metadata, Symmetry, Parameters
71 from datas.helpers import measurement2distance
---> 72 from spinorama.load import parse_graphs_speaker, parse_eq_speaker
73 from spinorama.speaker import print_graphs
74 from spinorama.plot import plot_params_default

File ~/.venv/spinorama/src/spinorama/load.py:50
40 from spinorama.compute_estimates import compute_sensitivity, compute_sensitivity_details
42 from spinorama.misc import (
43 graph_melt,
44 graph_unmelt,
(...) 48 measurements_complete_spl,
49 )
---> 50 from spinorama.load_klippel import parse_graphs_speaker_klippel
51 from spinorama.load_princeton import parse_graphs_speaker_princeton
52 from spinorama.load_rew_text_dump import parse_graphs_speaker_rew_text_dump

File ~/.venv/spinorama/src/spinorama/load_klippel.py:28
25 from spinorama.ltype import StatusOr
26 from spinorama.misc import sort_angles
---> 28 locale.setlocale(locale.LC_ALL, "en_US.UTF-8")
30 removequote = str.maketrans({'"': None, "\n": ""})
33 def parse_graph_freq_klippel(filename: str) -> StatusOr[tuple[str, pd.DataFrame]]:

File /usr/lib/python3.12/locale.py:615, in setlocale(category, locale)
612 if locale and not isinstance(locale, _builtin_str):
613 # convert to string
614 locale = normalize(_build_localename(locale))
--> 615 return _setlocale(category, locale)

Error: unsupported locale setting

---

Are these your previously stated WSL2 limitations?
If I'm too far away from being helped with these errors in a reasonable time, can you perhaps suggest another method of importing the data into R, please?
At this stage my main interest is high quality and default measurements only, and specifically Direct and Early Reflection frequency / level data from a manageable selection of loudspeakers.
However, a more comprehensive dataset would be ideal.

Again, thanks for your interest and help.

May I have permission to cite your valuable resource as part of my research, please?
If so, what is your preferred citation?

Thank you