Frontend for Segmentation

[chubetho]

Implement intergration for fe

Update 12.09

.env

• env is no longer loaded directly by the backend, but is passed by the frontend.
• you can now create an .env.local near .env to set api keys for acoustid and shazam. This will be read automatically and ignored by git.
• when use env in backend please call get_env("SERVICE_ACOUSTID_API_KEY") when you want to use it. Declare to a global variable may return None

# .env

VITE_APP_NAME=Audio Splitter
VITE_OUTPUT_FILE_NAME_TEMPLATE={{TITLE}}

# .env.local

VITE_APP_NAME=Audio Splitter
VITE_OUTPUT_FILE_NAME_TEMPLATE={{TITLE}}
VITE_SERVICE_ACOUSTID_API_KEY={{SERVICE_ACOUSTID_API_KEY}}
VITE_SERVICE_SHAZAM_API_KEY={{SERVICE_SHAZAM_API_KEY}}

Update 14.09

• build.spec now excludes files in src/backend/tests
• segment now can be saved through backend

[chubetho]

@JosuaE-FHWS can you fix error, when I want to split audio, I have this error from this line

yield librosa.load(
            path, mono=False, sr=samplerate, offset=offset, duration=duration
        ), start_time, duration

raise NoBackendError()
audioread.exceptions.NoBackendError

Before I merge main into this branch. It still works. Look this checkpoint

[JosuaE-FHWS]

@JosuaE-FHWS can you fix error, when I want to split audio, I have this error from this line

yield librosa.load(
            path, mono=False, sr=samplerate, offset=offset, duration=duration
        ), start_time, duration

raise NoBackendError()
audioread.exceptions.NoBackendError

Before I merge main into this branch. It still works. Look this checkpoint

I'm not entirely sure what the error is, however we had plans to remove the librosa.load() call from the yield there, since we currently don't use the result of it anyways. That would most likely resolve the issue.

[chubetho]

@ChrisItisdud I think we can merge it into main because the segmentation for fe worked in commit before the new segmentation logic was merged.

[chubetho]

@ChrisItisdud Wait. I will also implement store with backend here by the way

[twynb]

I was about to say, this looks good to me but storing via frontend still doesn't work :D

[chubetho]

I was about to say, this looks good to me but storing via frontend still doesn't work :D

yeah in linux i could save but audio was like 2x speed :cry:

[4lex0017]

@chubetho Please check backend modules . Seems to me there are a lot of lines that are reverted back to a previous state

[chubetho]

@chubetho Please check backend modules . Seems to me there are a lot of lines that are reverted back to a previous state

Segmentation.py is reverted because only old version is working. Just for demo how fe works. Will be reverted again in the future With other modules I can really tell what is missing

[4lex0017]

@chubetho I wouldnt revert it. The old version is quite unaccurate. Also the new version works if only used in backend. Maybe its a problem with the Filetype you are trying to read? Otherwise just remove the librosa load. also maybe try with package "ffmpeg"

[chubetho]

@chubetho I wouldnt revert it. The old version is quite unaccurate. Also the new version works if only used in backend. Maybe its a problem with the Filetype you are trying to read? Otherwise just remove the librosa load. also maybe try with package "ffmpeg"

FE only pass file path back to BE to process. So Im not sure something is wrong w file type. Furthermore I dont want to touch so much in BE, because I dont understand it much like you all do. Can you try the latest version of this branch? And then switch back to new segmentaion logic to find out why it causing "no backend error"?

[JosuaE-FHWS]

@chubetho I can't reproduce the error on my side, it seems to work just fine. Can you try to merge the remove-load-from-segmentation-generator branch into this and test again? I think this might fix the error.

[chubetho]

@chubetho I can't reproduce the error on my side, it seems to work just fine. Can you try to merge the remove-load-from-segmentation-generator branch into this and test again? I think this might fix the error.

I confirm this works now :+1:

[chubetho]

@ChrisItisdud This branch is ready to be merged