Updates in FreesoundExtractor #1510
Open
ffont wants to merge 22 commits into
Open
Conversation
Removed check for completely silent audio in FreesoundExtractor. There are files in Freesound which will be complete silence or almost silence, so it is fine to provide results for them as well. Also, this happens before descriptors have been computed, so it should not make things fail? I have to test that, for now making this PR and see how tests behave...
Before it was hardcoded to use python3.6, which is no longer installed in our manylinux images.
We don't need to use FFmpeg 2.8 as since #1494 we support FFmpeg 5+
Eigen 5.x (current Homebrew formula) requires C++14 features (std::enable_if_t, std::integer_sequence, etc.), causing build failures with the previous C++11 default.
Update cibuildwheel-tensorflow.toml arm64 override: MACOSX_DEPLOYMENT_TARGET 15.2 → 15.4 (matching the current libtensorflow 2.20.0 bottle's minimum)
Let's use the same pipeline in both workflows so that we don't have to fix deps individually everytime we change something.
This avoids running checkout@v4, which depends on Node 20 on Centos7 as it is not supported. Build doc inside docker to mimic the cibuildwheel behaviour.
mac-13 --> mac-15-intel. mac-13 is deprecated. https://github.com/actions/runner-images/blob/main/images/macos/macos-15-Readme.md
- 15.0 for the normal wheels as required by libcrypto.3.dylib and libssl.3.dylib - 15.2 for the tensorflow wheels as required by libtensorflow
Let's take the opportunity to bump our tooling: - ubuntu22.04 --> ubuntu24.04 - checkout v4 --> v6 - cibuildwheel v2.23 -> v3.2 - upload-artifact v4 -> v5
- cp38: project does not target Python 3.8.
- *t-*: free-threaded (no-GIL) builds. The Essentia C bindings rely on
legacy PyTypeObject definitions that fail to instantiate under
cp313t/cp314t ("Type does not define the tp_name field"). Re-enable
once the bindings are audited for free-threading compatibility.
- *-musllinux*, *i686: not supported by our manylinux2014 build images.
Since Node.js 20 actions are deprecated
Co-authored-by: Copilot <copilot@github.com>
ffont
changed the title
ffont
changed the title
Member
Author
|
I think the implementation is probably fine, but I'm a bit concerned about the size of the hard-coded literals in Also note that this includes Pablo's changes to make builds work. Maybe better to merge these first to master so the diff here makes more sense. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR includes some updates in the
FreesoundExtractor:fs.xnamespace). This is used in Freesound when indexing the descriptors. We used to do that post-processing in Freesound code, but doing it here it is useful as people that uses theFreesoundExtractorto analyze sounds locally, gets descriptors which are directly compatible with those indexed in Freesound.