Package Details: python-librosa 0.10.1-1

Git Clone URL: (read-only, click to copy)
Package Base: python-librosa
Description: Python library for music and audio analysis
Upstream URL:
Licenses: ISC
Submitter: petronny
Maintainer: petronny (AutoUpdateBot)
Last Packager: AutoUpdateBot
Votes: 9
Popularity: 0.129183
First Submitted: 2017-08-20 13:38 (UTC)
Last Updated: 2023-08-17 07:56 (UTC)

Pinned Comments

petronny commented on 2023-02-22 11:01 (UTC)

Prebuilt binaries of this package and its dependencies can be found in the arch4edu repository.

Latest Comments

Auerhuhn commented on 2023-06-02 22:35 (UTC)

@petronny Would you mind adding python-lazy-loader and python-soxr to depends? Those seem to be missing according to namcap.

Auerhuhn commented on 2023-03-02 08:52 (UTC) (edited on 2023-03-02 08:52 (UTC) by Auerhuhn)

PSA: python-librosa is currently failing at runtime. This is due to

The error message is:

SystemError: initialization of _internal failed without raising an exception

Once the issue is fixed in python-numba, reinstalling python-numba should fix the issue here as well.

In the meantime, python-numba-git, which already has the patch applied, should work as a drop-in replacement.

zoe commented on 2023-02-22 11:07 (UTC)

@petronny : Version 0.10.0 does build very well, if python-lazy-loader is installed. I suggest to add python-lazy-loader as a makedepends.

petronny commented on 2023-02-22 11:01 (UTC)

Prebuilt binaries of this package and its dependencies can be found in the arch4edu repository.

petronny commented on 2023-02-22 11:00 (UTC)

@zoe Version 0.10.0 could be built in clean chroot. Please check

zoe commented on 2023-02-22 10:44 (UTC) (edited on 2023-02-22 11:04 (UTC) by zoe)

With version 0.10.0, does not build anymore, python-lazy-loader is needed as makedepends :

      ==> Lancement de build()…
      [… … …]
      ModuleNotFoundError: No module named 'lazy_loader'
      ==> ERREUR : Une erreur s’est produite dans build().
       -> error making: python-librosa

SpotlightKid commented on 2022-05-19 13:08 (UTC) (edited on 2022-05-19 13:10 (UTC) by SpotlightKid)

Here's a diff for PKGBUILD (full file) to update the package to the latest release 0.9.1:

  • Updated pkgver.
  • Used sha256sum check sum.
  • Added python-packaging to depends.
  • Added --skip-build option to python setup install call in package function.
diff --git a/PKGBUILD b/PKGBUILD
index bc0c9d1..f379898 100644
@@ -3,16 +3,20 @@

 pkgdesc="Python library for music and audio analysis"
-depends=('python-numpy' 'python-scipy' 'python-six' 'python-numba' 'python-soundfile' 'python-joblib' 'python-matplotlib' 'python-audioread' 'python-decorator' 'python-pooch' 'python-scikit-learn' 'python-resampy')
+depends=('python-audioread' 'python-decorator' 'python-joblib'
+         'python-matplotlib' 'python-numba' 'python-numpy'
+         'python-packaging' 'python-pooch' 'python-resampy'
+         'python-scikit-learn' 'python-scipy' 'python-six'
+         'python-soundfile')

 build() {
   cd "$srcdir/${_pkgname}-${pkgver}"
@@ -21,8 +25,8 @@ build() {

 package() {
   cd "$srcdir/${_pkgname}-${pkgver}"
-  python install --root="$pkgdir"/ --optimize=1
-  install -Dm644 "${pkgdir}/usr/share/licenses/${pkgname}/LICENSE"
+  python install --root="$pkgdir"/ --skip-build --optimize=1
+  install -Dm644 -t "${pkgdir}/usr/share/licenses/${pkgname}/LICENSE"

 # vim:set ts=2 sw=2 et:

HenryJia commented on 2020-08-27 16:02 (UTC)

You're missing pooch as a dependency

linos commented on 2020-06-18 07:34 (UTC) (edited on 2020-06-18 17:46 (UTC) by linos)

Please patch it and pin the numba version to 0.48. Otherwise you will get the error:

ModuleNotFoundError: No module named 'numba.decorators'

--- src/librosa-0.7.2/  2020-01-13 16:18:58.000000000 +0100
+++ src/librosa-0.7.2/  2020-06-18 19:32:36.900645588 +0200
@@ -52,7 +52,7 @@
         'decorator >= 3.0.0',
         'six >= 1.3',
         'resampy >= 0.2.2',
-        'numba >= 0.43.0',
+        'numba == 0.48.0',
         'soundfile >= 0.9.0',

CyberGhost commented on 2017-12-29 02:35 (UTC)

I believe python-decorator and python-scikit-learn are both depedencies for this module.