search

Project-OSmOSE / datarmor-toolkit

This repo gathers all our analytical codes to process big (mostly audio) data (eg related to machine learning, ambient noise analysis…)
0 stars 6 forks source link

refacto previsualization #62

Closed mathieudpnt closed 1 day ago

mathieudpnt commented 3 weeks ago

I fixed the adjust_spectro located in src/utils_datarmor, it should run smoothly provided the jupyter hub session runs with enough ressources

By default, we recommend to use Datarmor - Jupyterlab - 8 cores, 32 GB RAM, 2 hours

Gautzilla commented 3 days ago

When trying to call generate_spectro(), I have the following error:

AssertionError: "'datetime_begin' must be either 'None' or a datetime, None not a valid value"

I didn't dig into what you wrote, but I guess it comes from the removing of the validate_datetime function?

Gautzilla commented 1 day ago

Shouldn't the temporary folder that holdsthe adjustment spectrograms be deleted after the display rather than before displaying new spectrograms?

After plotting some adjustment spectros, I'm left with this structure:

image

Gautzilla commented 1 day ago

When trying to plot from specific files, I get the following error:

ValueError: "['2023_04_05_15_09_06.wav', '2023_04_05_15_19_06.wav'] not found in file_metadata file."

Since the audio is reshaped, there is no file_metadata.csv file, shouldn't the info be searched in timestamp.csv?

EDIT: I just saw in the code that the audio were searched only in the original folder. If that's the intended behaviour of the feature, that's ok

mathieudpnt commented 1 day ago

Shouldn't the temporary folder that holdsthe adjustment spectrograms be deleted after the display rather than before displaying new spectrograms?

After plotting some adjustment spectros, I'm left with this structure:

image

@Gautzilla The temporary folder will be deleted when you compute the actual spectrograms of the dataset so this is normal behavior to me 🥸